1 INFO-VAX	Tue, 09 Apr 2002	Volume 2002 : Issue 196       Contents: Re: An alpha" ANN: HGFTP V3.0-1 is now available& Re: ANN: HGFTP V3.0-1 is now available& Re: ANN: HGFTP V3.0-1 is now available& Re: ANN: HGFTP V3.0-1 is now available4 ANNOUNCE: VERITAS NBU OpenVMS 3.4.1 GA now available RE: Anonymous ftp  Re: Anonymous ftp  Re: Anonymous ftp  Re: Anonymous ftp * RE: Anonymous ftp under tcpip5.1 on AXP7.30 Re: another hint of post-merger direction, maybe Re: API for FTP # BACKUP causes crash on VMS V5.5-2H4  Re: behavior of working set  Re: behavior of working set  Re: behavior of working set  Re: behavior of working set  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  RE: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures = Boost Your Windows Reliability!!                         4884  Re: Daylight savings Re: Daylight savings" DCL and batch procedure conversion& Re: DCL and batch procedure conversion Exporting SYSUAF.LIS Re: Exporting SYSUAF.LIS Re: Exporting SYSUAF.LIS Re: Exporting SYSUAF.LIS Re: FTP captive account?* Re: Help! VT100 emulators - some problems.+ Re: How to know if EDS hires for VMS Eng. ? F Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxF Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxF Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinux Re: IA64 is not the VAX  Re: IA64 is not the VAX  Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles' Re: Just another linux nightmare day #2 ' Re: Just another linux nightmare day #2 ' Re: Just another linux nightmare day #2 ' Re: Just another linux nightmare day #2 ' Re: Just another linux nightmare day #2 ( Re: Just another linux nightmare day ... Re: Macro-64 and Vax VMS Re: Memory Corruption  Re: Memory Corruption  Merge Two CMS libraries? Re: More positive OpenVMS news Re: More positive OpenVMS news Re: More positive OpenVMS news Re: More positive OpenVMS news Re: More positive OpenVMS news Re: More positive OpenVMS news Re: More positive OpenVMS news Re: Motif startup problem. Re: Motif startup problem. Re: Motif startup problem. Re: Motif startup problem. Re: Motif startup problem.) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it ) Re: Predictions - just for the hell of it  Processes in state RWSCS Re: Processes in state RWSCS Re: Processes in state RWSCS Re: Processes in state RWSCS Re: Processes in state RWSCS Re: Processes in state RWSCS Re: Processes in state RWSCS Re: Processes in state RWSCS> Re: Relative invulnerability of VMS to buffer-overflow attacks remote mailbox Re: Server Side Includes Re: Server Side Includes4 Re: Spring Forward, Fall Back - unfortunately not :-5 Re: VMS boottime - how to calculate # of days running  Re: VMS doesn't suck Re: VMS doesn't suck Re: VMS doesn't suck VMS License PAK Problems  Re: VMS To Tru64 Migration PlansB Re: What OS Was www.openvms.digital.com Running On n 22-Sep-1998 ?# Re: Who has (ever) used DECmigrate? # Re: Who has (ever) used DECmigrate? # Re: Who has (ever) used DECmigrate? # Re: Who has (ever) used DECmigrate? ( Re: Why VMS is "unhackable" lesson 1 ...4 Re: Why VMS is "unhackable" lesson 3 ... (follow-up)4 Re: Why VMS is "unhackable" lesson 3 ... (follow-up)4 Re: Why VMS is "unhackable" lesson 3 ... (follow-up) Will IBM buy Sun?  Re: Will IBM buy Sun?  Re: X $ [Change topic] Re: Andrew's Literacy( Re: [Change topic] Re: Andrew's Literacy  F ----------------------------------------------------------------------  % Date: Tue, 09 Apr 2002 16:44:42 +0200 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>  Subject: Re: An alpha ( Message-ID: <3CB2FE5A.AFBC87B@gtech.com>   labadie wrote:M > If you try very quickly a lot of passwords, you will disable that username,  > nothing more.    ????  $ LGI_BRK_DISUSER is not set default !   Arne   ------------------------------  # Date: Tue, 09 Apr 2002 15:06:26 GMT - From: goathunter@goatley.com (Hunter Goatley) + Subject: ANN: HGFTP V3.0-1 is now available 1 Message-ID: <3cb30231.229009768@news.process.com>   G HGFTP V3.0-1 is now available for download.  HGFTP is an FTP client and C server for OpenVMS that runs on top of all of the VMS TCP/IP stacks D (MultiNet, TCPware, and the others).  Either the client, the server,B or both can be run without disrupting your existing FTP setup.  Of@ special note to TCP/IP Services users is the fact that the HGFTPA client and server support STRU O VMS mode for VMS file transfers. F And the server supports UNIX-emulation for listing output, allowing it3 to work with most of the PC web browsers out there.   6 HGFTP V3.0-1 includes the following changes over V3.0:  =           o  The HGFTP server has been modified once again to >              try to support the web browsers of the world. The@              support for rooted directories has been improved to>              work with recent versions of IE and Netscape when,              UNIX-emulation mode is enabled.  A           o  The HGFTP client now supports the /PASV command-line @              qualifier to make passive mode the default for file              transfers.   <           o  Coincidentally, a month after /PASV support was:              added, a DCL ECO from Compaq adds /PASSIVE to>              COPY/FTP and DIR/FTP. HGFTP now supports that new-              qualifier on those commands too.   8           o  The HGFTP server will now ignore UNIX-style>              qualifiers on the LIST/NLST commands. Many stupid?              GUI FTP clients assume the remote server is a UNIX ;              system and will incorrectly send commands such >              as ``LIST -aL''. HGFTP now ignores any characters!              prefixed with ``-''.   <           o  A subtle bug in the encoding of alias passwords@              that occurred when the password was longer than the?              host name has been fixed. Stored aliases that meet >              this criteria will most likely need to be REMOVEd>              and ADDed again to get the proper encoding of the              password.  9           o  The commands ON ERROR CONTINUE and ON SEVERE @              CONTINUE have been added to the HGFTP client. These?              commands were documented in previous releases, but $              were, in fact, missing.  ;           o  Several issues that prevented successful local >              compilation from the sources without modification!              have been corrected.   , You can find HGFTP using the following URLs:   http://www.process.com/openvms/   5 ftp://ftp.process.com/vms-freeware/fileserv/hgftp.zip : http://vms.process.com/ftp/vms-freeware/fileserv/hgftp.zip1 ftp://ftp.tmk.com/vms-freeware/fileserv/hgftp.zip 6 http://www.tmk.com/ftp/vms-freeware/fileserv/hgftp.zip  2 And on the other mirrors within the next 24 hours.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  + Date: Tue, 09 Apr 2002 17:29:36 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> / Subject: Re: ANN: HGFTP V3.0-1 is now available ; Message-ID: <01KGD8ZUDH3G9EEGPR@sysdev.deutsche-boerse.com>   I > HGFTP V3.0-1 is now available for download.  HGFTP is an FTP client and E > server for OpenVMS that runs on top of all of the VMS TCP/IP stacks F > (MultiNet, TCPware, and the others).  Either the client, the server,D > or both can be run without disrupting your existing FTP setup.  OfB > special note to TCP/IP Services users is the fact that the HGFTPC > client and server support STRU O VMS mode for VMS file transfers. H > And the server supports UNIX-emulation for listing output, allowing it5 > to work with most of the PC web browsers out there.   E On my home system, I generally run out-of-the-box DEC/Digital/Compaq  H stuff, using (quite a bit of) freeware stuff (especially Madgoat) where ; the corresponding functionality is completely missing with  H DEC/Digital/Compaq stuff (i.e. ZIP, LaTeX etc---an exception is the OSU C server, but I started that before CSWS and don't see any reason to  C switch).  Thus, HGFTP is something I have no experience with.  For  G various reasons, I don't want to move to another IP stack, so the fact  9 that HGFTP runs on top of what I have now is interesting.   H What are the main reasons for using it as opposed to the out-of-the-box  DEC/Digital/Compaq solution?   ------------------------------  # Date: Tue, 09 Apr 2002 17:02:56 GMT - From: goathunter@goatley.com (Hunter Goatley) / Subject: Re: ANN: HGFTP V3.0-1 is now available . Message-ID: <3cb31c25.818767@news.process.com>  8 On Tue, 09 Apr 2002 17:29:36 +0100 (MET), Phillip Helbig+ <HELBPHI@sysdev.deutsche-boerse.com> wrote:   J >> HGFTP V3.0-1 is now available for download.  HGFTP is an FTP client andF >> server for OpenVMS that runs on top of all of the VMS TCP/IP stacksG >> (MultiNet, TCPware, and the others).  Either the client, the server, E >> or both can be run without disrupting your existing FTP setup.  Of C >> special note to TCP/IP Services users is the fact that the HGFTP D >> client and server support STRU O VMS mode for VMS file transfers.I >> And the server supports UNIX-emulation for listing output, allowing it 6 >> to work with most of the PC web browsers out there. > F >On my home system, I generally run out-of-the-box DEC/Digital/Compaq I >stuff, using (quite a bit of) freeware stuff (especially Madgoat) where  < >the corresponding functionality is completely missing with I >DEC/Digital/Compaq stuff (i.e. ZIP, LaTeX etc---an exception is the OSU  D >server, but I started that before CSWS and don't see any reason to D >switch).  Thus, HGFTP is something I have no experience with.  For H >various reasons, I don't want to move to another IP stack, so the fact : >that HGFTP runs on top of what I have now is interesting. > I >What are the main reasons for using it as opposed to the out-of-the-box   >DEC/Digital/Compaq solution?   9 Some of the things I can think of off the top of my head:   E   - Client and server support STRU O VMS (also supported by MultiNet, F     TCPware, CMU-IP, and formerly Pathway---everybody but UCX) so thatG     VMS <-> VMS file transfers maintain the VMS file attributes without #     the use of a separate FDL file.   H   - The server supports any number of anonymous accounts, with differentG     access rules for each account (or the same for all, if you prefer).   I   - The server can run in UNIX-emulation mode and has special code to try H     to work with all the various PC GUI clients and browsers that assume     the world runs UNIX.  :   - A client that supports ftp URLs from the command line:  G       $ ftp/url "ftp://ftp.process.com/vms-freeware/fileserv/hgftp.zip"   I   - A client that supports /ANONYMOUS on the command line and on the user B     command (very handy back in the days before COPY/FTP existed).  G   - A client that displays file transfer stats that are meaningful when C     you press CTRL-T or CTRL-A (so you can see how fast the file is E     transferring, that it is in fact still transferring and not hung, :     what data rate you're seeing, and estimated time left:  8     [sending file dka600:[hunter]file.avi;1 to file.avi]O     1988608 bytes (3884/266732 blocks), 1%, in 00:00:02.10 = 946956 cps, IO=971 .     Estimated time remaining:    0 00:02:22.00  )     I couldn't work without this feature.   D   - The client features a VMS CLI$ interface with support for normal@     UNIX-style commands (SET DEFAULT and CD).  Unlike any of the>     other FTP clients for VMS, HGFTP commands provide the caseA     preservation of a UNIX-style interface.  For example, instead ,     of typing this for a UNIX remote server:         FTP> cd "pub/VMS/alpha"        you can just do:         FTP> cd pub/VMS/alpha   D     A minor thing, but I've used it for so long that I forget I have1     to quote such strings with the other clients.   I  - The HGFTP client and server implement the REIN command, which lets you D    log out of an account and log in to another without disconnecting    from the remote server.  F  - You can easily run the server on a non-standard port, and use /PORTJ    in the client to talk to it, allowing you to run the HGFTP server while0    still running your stack's native FTP server.  I  - The MGET and MPUT commands support /RECURSIVE for copying entire trees 7    back and forth while maintaining the tree structure.   I There are lots of other features, but those are the main ones I can think H of that should make it especially interesting to a TCP/IP Services site.     Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  # Date: Tue, 09 Apr 2002 17:22:54 GMT - From: goathunter@goatley.com (Hunter Goatley) / Subject: Re: ANN: HGFTP V3.0-1 is now available / Message-ID: <3cb322b7.2500866@news.process.com>   P On Tue, 09 Apr 2002 15:06:26 GMT, goathunter@goatley.com (Hunter Goatley) wrote:  , >HGFTP V3.0-1 is now available for download.  7 A minor problem was reported, and I updated the kits at 5 9-APR-2002 13:20 EDT.  I had forgotten to add the new 5 /PASV qualifier to the FTP.CLD that gets installed if 5 you answer YES to that installation question.  If you 6 downloaded the kit before 13:20 EDT, you should either1 pick it up again, or you can use a foreign symbol  to invoke FTP instead:      $ ftp :== $hg_exe:ftp.exe  = That gets around the problem with /PASV not being recognized. ' My apologies for this slipping through.   , You can find HGFTP using the following URLs:   http://www.process.com/openvms/   5 ftp://ftp.process.com/vms-freeware/fileserv/hgftp.zip : http://vms.process.com/ftp/vms-freeware/fileserv/hgftp.zip1 ftp://ftp.tmk.com/vms-freeware/fileserv/hgftp.zip 6 http://www.tmk.com/ftp/vms-freeware/fileserv/hgftp.zip  2 And on the other mirrors within the next 24 hours.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  % Date: Tue, 09 Apr 2002 07:02:46 +0000 % From: Alan Fay <alan.fay@veritas.com> = Subject: ANNOUNCE: VERITAS NBU OpenVMS 3.4.1 GA now available B Message-ID: <1018335802.28212.0.nnrp-14.9e98bc8a@news.demon.co.uk>  E The 3.4.1 GA release of the NBU OpenVMS client is now available as a  E downloadable patch upgrade. This patch will upgrade any existing 3.2  , or 3.4 VMS client installation to 3.4.1GA.    H This release has been fully GA tested with 3.4, 3.4.1 and 4.5 GA master G servers. It has been specifically tested to operate correctly with the  I 4.5 Java interface. This release *is* required when running with the new  & 4.5 GA DataCenters and BusinesServers.  > This release includes the following new features and changes:   /   o  Support for OpenVMS V7.3 (VAX and Alpha).    G   o  Full support for the case sensitive and extended character set of         OpenVMS ODS-5 file systems.  H   o  Full support for logical name search lists. Users can now backup a G      single system wide logical name search list of user defined local  C      devices. This method can and should be used when implementing  1      ALL_LOCAL_DRIVES backups on OpenVMS systems.   G   o  The files-file structure has been modified to include a top level  I      root '/'  directory (which does not exist on OpenVMS) and any other  K      required intervening directory steps to the VMS file structure. Users  G      can now browse VMS client backups from the root '/'  directory to  H      display as folders all the backed up device names. It is no longer H      neccessary to know or specify a specific device name when browsing       VMS client backups.  G   o  User backups from the NBU VMS client utility are now synchronized mL      with the master server. This synchronization should ensure that client H      initiated user backups or restores will never hang and will always -      report any error status from the server.s  N   o  The VMS client host properties are now reported to the server correctly. M      This includes the NBU client version, operating system name and OpenVMS eJ      version. Because the VMS client behaves as if it were a UNIX client,       the machine type is UNIX.  J   o  Server VMS client backup catalogues now verify successfully with the       latest 4.5 bpverify.e  H   o  EMC Symmetrix Disk Arrays which do not support extended $QIO's are ;      now detected and non-extended $QIO's are used instead.e  E   o  Only the file-headers of directories are now backed up. OpenVMS  G      directories are backed up as files so they can be referenced with  "      the *.DIR file specification.  I   o  Start and end notify scripts are now run with a P1 parameter within mH      the buffer limits of LIB$SPAWN. Notify scripts are now only run on #      a backup and not on a restore.g  J   o  Correctly pass fatal NBU exit status codes to DCL command procedures.G      If an NBU backup or restore is anything other than successful, or xJ      partially successful, an error status is returned to the DCL command       procedure.l  N   o  Email notification in the VMS client log files is no longer the default. H      To enable this behaviour specify the USEMAIL=username parameter in       NBU$CONFIG.  J   o  UNIX file specifications must now be specified on a server initiated K      alternate path restores. Client initiated alternate path restores can u4      continue to use native VMS file specifications.  I   o  Support has been added to allow the restoration of NBU UNIX backups eM      (which have no VMS file attributes) to a VMS client. Files are restored s+      as fixed length 512 byte record files.   sL   o  The SHOW CLIENT command will now display the hardware name and OpenVMS 
      version.e  I   o  The user guide has been updated to explain 4.5 server configuration  #      procedures for the VMS client.   -   o  The user guide is now fully hyperlinked.t  1 This upgrade release is publicaly available from:h  I ftp://ftp.emea.support.veritas.com/pub/support/Products/NetBackup_OpenVMSt     nbu_v3_4_1_GA.bcke  G If you are using a WWW browser then press the shift key and left click lF (at the same time) on the file name to download, then save file as... @ or right click and Save Target As...  Please install as per the D 'Downloadable Patches' section of the 'user_guide.pdf' available at  the same location. e     Alan Fay VERITAS Software Corporation s Roseville Engineeringe   ------------------------------   Date: 9 Apr 2002 08:11:40 -0500e- From: koehler@encompasserve.org (Bob Koehler)b Subject: RE: Anonymous ftp3 Message-ID: <HyKeXtSvMn4Q@eisner.encompasserve.org>   T In article <pY35oCqQuqOh@eisner.encompasserve.org>, briggs@encompasserve.org writes:  F > The FTP server provided with TCP/IP services version 5.1 is standardE > compliant.  It responds to the LIST command with a list of files inaH > a system-specific format which is not supposed to be machine readable.  F    There are a hell of a lot of FTP GUIs which ignore this in the RFC:  B       "Since the information on a file may vary widely from systemC        to system, this information may be hard to use automaticallys>        in a program, but may be quite useful to a human user."  G    Fortunately most of them have some clue as how to recognize a lot oflD    different systems and the output they produce.  Some of them evenD    have a setting which forces them to use the machine readable NLST'    command instead of the LIST command.-  C    Unfortunately this is a violation of the RFC, there's nothing to D    prevent the next release of any FTP server from having a slightly6    different format which the client won't understand.  F    But at least it's not as bad as those FTP clients and servers which!    believe "binary always works".h   ------------------------------   Date: 9 Apr 2002 08:15:40 -0500 - From: koehler@encompasserve.org (Bob Koehler)  Subject: Re: Anonymous ftp3 Message-ID: <qD0irUoniPph@eisner.encompasserve.org>A  i In article <20020408191810.B21751@eisenschmidt.org>, John Eisenschmidt <jweisen@eisenschmidt.org> writes:  > N > This drives me nuts with HTML. I strive to only write W3C Compliant HTML, s=N > o it will work on Netscape, Mozilla, IE, Konqueror, Opera, etc. Very few pe=N > ople do this - they think making it IE compliant is enough. I am forever se=N > nding emails to webmasters that say "Your website sucks. Stop developing in=N >  a vacuum". Our developers have to test anything for Netscape and IE on the== >  Mac and PC before it can be considered for productions.=20   G    I generally test with lynx.  Since I hand code most of my HTML, I'vee    had very few problems.  V  J    And since I now have to deal with federal regulations requiring better E    access to web pages that made my life easier (didn't have to fix as
    thing).        ------------------------------  # Date: Tue, 09 Apr 2002 14:58:07 GMTo- From: goathunter@goatley.com (Hunter Goatley)t Subject: Re: Anonymous ftp1 Message-ID: <3cb300c8.228648058@news.process.com>m  G On Mon, 8 Apr 2002 11:48:44 -0700, "Tom Linden" <tom@kednos.com> wrote:e  C >> You say that you've tried MadGoat and had it fail.  What versionrB >> with what configuration settings?  Failing how?  Do you want to( >> bitch at Compaq or fix your problems? >.B >Use Mozilla 9.9 on VMS and point it to ftp://ftp.process.com  QED >tC I've not yet been able to figure why Mozilla fails to work with the D UNIX-emulation output from the VMS servers.  Netscape V6.2 under W2KD actually gets an access violation while trying to parse the returned= output.  I don't know why.  But Netscape V4.7 and IE work OK.    Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/-8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------  % Date: Tue, 09 Apr 2002 16:49:44 +0200m= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>M Subject: Re: Anonymous ftp) Message-ID: <3CB2FF88.583526F6@gtech.com>    Tom Linden wrote: ! > Running tcpip 5.1 under AXp 7.3R > E > When accessing the system from a command line interface works fine,P? > but when using browser with address ftp://freja.kednos.com it H > display the ftp root directory but when you try to click on any of theB > entries. it responds "The page cannot be display"  This with IE.! > with Opera doesn't get that fari >  > This waht IE displaysr > G > 01/01/1980 12:00PM              0 Directory SYS$SYSDEVICE:[ANONYMOUS]s@ > 01/01/1980 12:00PM              0 INPUT.DIR;1              1/4: > 23-AUG-2001 17:28:00  [SYSTEM]               (RWE,RWE,,)@ > 01/01/1980 12:00PM              0 LOGIN.COM;1              1/4< > 23-AUG-2001 17:27:54  [ANONY,ANONYMOUS]      (RE,RE,RE,RE)@ > 01/01/1980 12:00PM              0 PUB.DIR;1                1/4> > 28-AUG-2001 07:40:54  [SYSTEM]               (RWE,RWE,RE,RE)A > 01/01/1980 12:00PM              0 Total of 3 files, 3/12 blocksm > < > I have a similar set up on Tru64 4.0d and that works fine. > Is this a permission issue?0  C Crappy PC Web-browsers that do not understand VMS filename-syntax !-   Arne   ------------------------------  $ Date: Tue, 9 Apr 2002 07:04:59 -0700# From: "Tom Linden" <tom@kednos.com>c3 Subject: RE: Anonymous ftp under tcpip5.1 on AXP7.3 9 Message-ID: <CIEJLCMNHNNDLLOOGNJICEJKELAA.tom@kednos.com>s   > -----Original Message-----2 > From: sms@antinode.org [mailto:sms@antinode.org]& > Sent: Monday, April 08, 2002 8:50 PM > To: Info-VAX@Mvb.Saic.Com 5 > Subject: Re: Anonymous ftp under tcpip5.1 on AXP7.3S >i > 5 > From:	SMTP%"tom@kednos.com"  8-APR-2002 22:47:40.03u@ > > > > > > > I note in our logs many attempts over the last six > months to loga> > > > > > > > in by a user who uses the ident Xgpuser@home.com > where X is any$ > > > > > > > capital letter.  [...] > > > [...]t? > > >    Actually, no.  I hate to impede your ranting, but theyr > were somethingH > > > else which was also too lame to cope with a VMS DIRECTORY listing.A > > > "[A-Z]gpuser@home.com" is different from "IEUser@", believep > it or don't. > >vJ > > I just assumed that they were other browsers, is not that how you also > > interpret it?t >gH >    That's right.  It is not.  I first saw these in action on a SolarisI > system (WU-FTP server) at work, where it got further (for a while).  ItoE > appears to be a script which, given a write-enabled "/pub/incoming" = > directory, installs some marker files, creates a ("hidden")-J > ".<date-time-code>p" directory (such as ".020214225837p" for [20]02, FebG > 14, 22:58:37, UTC, perhaps), sends some test files of different sizes I > (1KB, 1MB, presumably to test your bandwidth), and, if you're suitable,rI > a collection of various data files which you probably didn't wish to be. > offering on your FTP server.  K I do not have write privleges enabled for anonymous ftp under tcpip5.1, butMI I saw what you are describing about a year ago on the ftp server on Tru64a 4.0dJ , there they created .tmp directory and under it 1MB file called 'tagged', IIRCK That all stopped when I tightened the permissions, but I left .tmp there to + punish them, they could look but not touch.r >pH >     At work, there's a "cron" job which periodically checks the uploadD > directory, and sends e-mail when something arrives.  Our server isJ > willing to accept uploads, but refuses to send those files out again, soD > it was not actually useful to the script kids.  For a while I justH > deleted the stuff soon after it arrived, but eventually I got tired ofG > that and found that not having an incoming directory named "incoming"c > was enough to defeat it. >sA >    These scripts are different varieties of FTP client, but not I > legitimate browsers.  The "IEUser@" and "Squid@" idents are more likelyo: > to be from legitimate clients (browser or proxy server). >cE >    I don't have any particular objection to MSIE users using my FTPaF > server, but with all the FTP clients out there which work better, my0 > urge to accomodate the lame ones is not great. >IJ > ------------------------------------------------------------------------ >VE >    Steven M. Schweda               (+1) 651-699-9818  (voice, home)hE >    382 South Warwick Street        (+1) 763-781-0308  (voice, work)iI >    Saint Paul  MN  55105-2547      (+1) 763-781-0309  (facsimile, work) ; >    sms@antinode.org                sms@provis.com  (work)h >f   ------------------------------  # Date: Tue, 09 Apr 2002 16:27:01 GMTe% From: jlsue <jlsuexxxz@insightbb.com>X9 Subject: Re: another hint of post-merger direction, maybei8 Message-ID: <1c56bug14vnqp6ib3o58a1ssdlf4vbjuui@4ax.com>  E On Mon, 08 Apr 2002 17:49:45 +0100, Andrew Harrison SUNUK Consultancye3 <andrew_nospam.remove_this.harrison@sun.com> wrote:c   >s >e
 >jlsue wrote:h >p6 >> On Wed, 03 Apr 2002 15:22:21 +0100, Andrew Harrison6 >> <andrew_nospam.harrison_remove_this@sun.com> wrote: >>     >>>s >> rG >> Well, the particular SPEC isn't important to me.  I'm just trying toaF >> understand how you choose which benchmarks you consider to be valid >> for comparisons.  >> r >f >iA >As ever you missed the point. The top500.org list isn't compiledd? >by Sun or any other vendor. top500 have chosen to use Linpack.f  A As ever, YOU missed my point.  I never stated that top500.org was6C compiled by Sun.  What I was trying to understand is why you choose F which benchmark organizations to spout on about.  You offer no clue asA to why that's better than spec.org stats (which, similarly, isn't C compiled by Compaq or any other vendor).  However, you consistently E berate the spec benchmarks that, for some reason, put Compaq ahead of  Sun.  F My main, more subtle perhaps, point is that you seem to be very biased. in which benchmarks you lend your support for.  1 Not speaking for anyone, certainly not DEC/CompaqT- (get rid of the xxxz in my address to e-mail)I   ------------------------------  % Date: Tue, 09 Apr 2002 16:09:07 +0200h= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>t Subject: Re: API for FTP) Message-ID: <3CB2F603.8855C50F@gtech.com>i   JF Mezei wrote:t6 > How did Digital implement the COPY/FTP  in VMS 7.2 ? > L > Doesn't that give some big hint that there would be some form of FTP API ?   Not necesarrily.  > It could just as well (and apperently is) coded in the utility itself.P   Arne   ------------------------------  $ Date: Tue, 9 Apr 2002 10:14:29 -0400; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com> , Subject: BACKUP causes crash on VMS V5.5-2H4$ Message-ID: <3cb2f7bc$1@news.si.com>  > I've an older VAX 4000-500A running VMS V5.5-2H4..  It has theL VAXY2K01_U2055 and VAXF11X06_U2055 ECOs applied.  Yet, when I try to make anJ image BACKUP of certain disks, the system crashes.  I've looked at all theJ ECOs I can find looking for another to address this problem, but the aboveL two seem the most likely, and they're already applied, as far as I can tell.  K Is there anything I've missed and that you may recall that may address this  problem?  Thanks.e --A Brian Tillman                   Internet: tillman_brian at si.comyA Smiths Aerospace                          tillman at swdev.si.comd= 3290 Patterson Ave. SE, MS      Addresses modified to prevents< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  # Date: Tue, 09 Apr 2002 12:32:04 GMTb' From: "Hans Vlems" <hvlems@zfree.co.nz>i$ Subject: Re: behavior of working set$ Message-ID: <3cb2df3e$1@zfree.co.nz>  ? Working set sizes in VAX/VMS are expressed in pages, not bytes. 3 A working set gets adjusted in increments of pages.gJ A steadily increasing memory usage may be an indication of memory leakage,N but may just as well indicate that the program is logging things, e.g. logging( that is printed at the end of a program.J If the memory requirements increase but stay within the working set limitsF assigned to the process then perhaps there's no cause to get alarmed.   - "Schild Niklaus" <schin@hta-bi.bfh.ch> wrote:uK >To see if there are memory leaks I use to observer my programs over a longcM >time with the  >sh proc /cont /id=15d97 command. I was told that the workingoG >set should not increase after a certain time. If it does then I have a J >memory leak. I that correct? My programs are written in c and so I do notM >have full controll of the memory because malloc usually allocates internally.I >bigger blocks of memory which are not immediatly freed when my memory isn >freed.tJ >I made also some tests with test programs where I enforced my programs toE >alloc and free the memory in a loop, so that it is easier to see thehK >effects. Under the same conditions sometimes the working set was increasedrC >very slowly and sometimes the working set was absolutely constant.. >aH >How is the relation with working sets and allocated memory. (if I alloc 1024M >bytes how does that affect the working set?). Should I also take care anbouts >the virtual pages?  >p >t >Thanks NikV >  >X       http://www.zfree.co.nz   ------------------------------   Date: 9 Apr 2002 07:59:27 -0500e From: briggs@encompasserve.org$ Subject: Re: behavior of working set3 Message-ID: <d78hsYa34ut9@eisner.encompasserve.org>d  Y In article <a8tv02$qum$1@rex.ip-plus.net>, "Schild Niklaus" <schin@hta-bi.bfh.ch> writes:bL > To see if there are memory leaks I use to observer my programs over a longN > time with the  >sh proc /cont /id=15d97 command. I was told that the workingH > set should not increase after a certain time. If it does then I have aK > memory leak. I that correct? My programs are written in c and so I do not N > have full controll of the memory because malloc usually allocates internallyJ > bigger blocks of memory which are not immediatly freed when my memory is > freed.  B You were misinformed.  An increase in working set over time is notC symptomatic of a memory leak.  An increase in virtual address spaceo* over time is symptomatic of a memory leak.  D Your working set will increase automatically if your page fault rateC per CPU second exceeds a particular SYSGEN parameter (PFRATH).  The C idea is that if your process is thrashing around, accessing virtual ? memory locations that are not physically resident, you probablyeA need more physical memory.  So VMS automatically gives it to you.i  F Locality of reference is the key.  A program that has good locality ofG reference can live within a tiny working set.  Even though that programrB may be leaking virtual memory like a sieve.  Conversely, a programD that has poor locality of reference can find its working set growingB right up to the relevant limits (WSEXTENT, WSMAX, GROWLIM, process: virtual memory, etc) even in the absence of a memory leak.  I Now, if you're doing dynamic memory allocation and you've been allocating C and de-allocating memory for a while and your program has an access A pattern that jumps around willy-nilly from one block of allocated C memory to the next, your locality of reference will tend to suffer. A And you may see a corresponding increase in working set size overo1 time as your virtual memory pool gets stirred up.r  C I use $ SHOW PROCESS /CONTINUOUS and keep an eye on "Virtual pages"u* as a simple sanity check for memory leaks.   	John Briggs   ------------------------------  $ Date: Tue, 9 Apr 2002 13:38:31 +0200$ From: "labadie" <labadie_g@decus.fr>$ Subject: Re: behavior of working set3 Message-ID: <SoAs8.2034$fL6.54256@news.cpqcorp.net>e  7 "Schild Niklaus" <schin@hta-bi.bfh.ch> wrote in message $ news:a8tv02$qum$1@rex.ip-plus.net...L > To see if there are memory leaks I use to observer my programs over a longF > time with the  >sh proc /cont /id=15d97 command. I was told that the workingbH > set should not increase after a certain time. If it does then I have aK > memory leak. I that correct? My programs are written in c and so I do notOC > have full controll of the memory because malloc usually allocates 
 internallyJ > bigger blocks of memory which are not immediatly freed when my memory is > freed.K > I made also some tests with test programs where I enforced my programs totF > alloc and free the memory in a loop, so that it is easier to see theL > effects. Under the same conditions sometimes the working set was increasedD > very slowly and sometimes the working set was absolutely constant. >tI > How is the relation with working sets and allocated memory. (if I allocs 1024G > bytes how does that affect the working set?). Should I also take carea anbout > the virtual pages? >e >u > Thanks Nik >e Helloh  % You can try callmon from the freeware : http://www.openvms.compaq.com/freeware/freeware40/callmon/   Regards-   Grard   ------------------------------   Date: 9 Apr 2002 09:23:09 -0700B) From: P.Young@unsw.EDU.AU (Patrick Young)e$ Subject: Re: behavior of working set= Message-ID: <55f85d77.0204090823.5fafbcd1@posting.google.com>r  ^ "Schild Niklaus" <schin@hta-bi.bfh.ch> wrote in message news:<a8tv02$qum$1@rex.ip-plus.net>...L > To see if there are memory leaks I use to observer my programs over a longN > time with the  >sh proc /cont /id=15d97 command. I was told that the workingH > set should not increase after a certain time. If it does then I have aK > memory leak. I that correct? My programs are written in c and so I do not   G That would be an assumption that may, or may not be true - depending on  the program.  N > have full controll of the memory because malloc usually allocates internallyJ > bigger blocks of memory which are not immediatly freed when my memory is > freed.  < No - you get what you asked for, however they are not freed.  K > I made also some tests with test programs where I enforced my programs toeF > alloc and free the memory in a loop, so that it is easier to see theL > effects. Under the same conditions sometimes the working set was increasedD > very slowly and sometimes the working set was absolutely constant.  E Memory will not be released until image exit. Most programs will need H to re-allocate portions of what has been released during their run time,E the CRTL therefore will not release it. Actually it is not so much upeG to the CRTL as "free" is a wrapper for LIB$VM_FREE (undocumented) just 8G as malloc is a wrapper for LIB$VM_MALLOC (undocumented) which wraps ...oJ forget it - way too late at night (actually 2am for me to look this up :-)  N > How is the relation with working sets and allocated memory. (if I alloc 1024N > bytes how does that affect the working set?). Should I also take care anbout > the virtual pages?  J How this works depends on your process quotas and on the system parameters of the PQL type.   As a _very minimal_ example:  
 $ create mu.c    #include <jpidef>d #include <lib$routines>i #include <stdio> #include <stdlib>n #include <starlet> #include <strings>   #define MB (1024 * 1024)   printws(char *comment) {       unsigned int wspeak;  2     lib$getjpi(&JPI$_WSPEAK, 0, 0, &wspeak, 0, 0);A     printf("%s peaking at %d MB\n", comment, wspeak * 8192 / MB);        }.   main() {       char *ptr;       printws("Starting out");       ptr = malloc(2 * MB);t  "     printws("after grabbing 2MB");       free(ptr);  !     printws("after freeing 2MB");8       ptr = malloc(12 * MB);  #     printws("after grabbing 12MB");d       bzero(ptr, 12 * MB);  "     printws("after filling 12MB");       free(ptr);  "     printws("after freeing 12MB");       sys$hiber();   }    $ cc muu $ lin mu& $ r/det/work=10/out='f$trnlnm("tt") mu= %RUN-S-PROC_ID, identification of created process is 0000015D. $    Starting out peaking at 10 MBt# after grabbing 2MB peaking at 12 MBt" after freeing 2MB peaking at 12 MB$ after grabbing 12MB peaking at 12 MB# after filling 12MB peaking at 33 MB # after freeing 12MB peaking at 33 MBu $ sh proc/id=15d/acc  G 10-APR-2002 02:12:13.60   User: PATRICK          Process ID:   0000015DrA                           Node: HILUX            Process name: ""    Accounting information:eA  Buffered I/O count:         9  Peak working set size:       4272 A  Direct I/O count:           2  Peak virtual size:         1952480A  Page faults:             1658  Mounted volumes:                0M  Images activated:           0)  Elapsed CPU time:          0 00:00:00.12 )  Connect time:              0 00:00:13.64e  ( $ write sys$output 4272*8192/(1024*1024) 33
 $ stop/id=15d    ------------------------------   Date: 9 Apr 2002 07:58:41 GMT/( From: nmm1@cus.cam.ac.uk (Nick Maclaren)  Subject: Re: Blade architectures0 Message-ID: <a8u6vh$kvm$1@pegasus.csx.cam.ac.uk>  O In article <GuA17q.82K@world.std.com>, Alex Colvin <alexc@world.std.com> wrote:sH >>> I remember when some folks pooh-poohed Virtual Memory because of its= >>> overhead and performance concerns.  Now it is everywhere.U >I9 >what made VM work was lowering the price of real memory.  > # >i've heard a remark something likeM > ) >	Virtual memory is good for many things. , >	Simulating real memory is not one of them. >R >[anyone kow the source?]0 > H >sure, you can access more memory than you own. but it's no way to live.  D I have seen that before.  I know where the VIEWPOINT originated, butF none of the people that I know were involved are well-known for having a good way with catch phrases.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679m   ------------------------------   Date: 9 Apr 2002 06:22:28 -0500 + From: young_r@encompasserve.org (Rob Young)s  Subject: Re: Blade architectures3 Message-ID: <jNX5Ni$q3MiI@eisner.encompasserve.org>   m In article <Zbvs8.53233$l7.5013438@bin7.nnrp.aus1.giganews.com>, "Bill Todd" <billtodd@metrocast.net> writes:n > : > "Rob Young" <young_r@encompasserve.org> wrote in message   > 4 >> Seems 100 miles would be about 7 milliseconds for? >> a write to complete. . . which may or may not be acceptable.u > K > You're off by a factor of two:  Keith's graph was for round-trip times at[M > one-way separation values, so a dual-round-trip write to a 100-mile-distanttB > partner would complete in about 3.5 ms. (plus disk-access time). >    	Ah.  < 	I should of spotted that... that and your earlier statementB 	10 us per mile, but it is late and hard to always be dead-on :-).A 	And disk write access *must* be cache or you will feel the pain.t    L > Of course, if you'll re-read my statement it said "anything MUCH UNDER 100K > mile separation is measurable but not noticeable", which certainly brings N > the dual round-trip time down to under 3 ms. (and the single round-trip time! > for a read to under 1.5 ms.).  t  E 	You can force reads to be local and won't have to add "overhead" for7F 	those except in extreme (large local queues).  And you may be talkingC 	overhead as it is hard to service reads in 1.5 ms (as we know, but,D 	then again... if those are solid state disks it wold be better than 	platters...).  ( > So even with the dual-round-trip writeH > 'feature' the distance would be noticeable only for fairly write-heavy > activity.  >   > 	You can force or skewer things depending on DBMS.  Elsewhere,8 	Keith mentions judicious use of Global Buffers (lowers)D 	disk read traffic and makes the read/write ratio a less traditional@ 	number... but yes this would be all over the place depending on
 	environment.u   				Rob    ------------------------------  * Date: Tue, 9 Apr 2002 11:58:35 +0000 (UTC)/ From: Sander Vesik <sander@haldjas.folklore.ee>r  Subject: Re: Blade architectures3 Message-ID: <1018353514.731362@haldjas.folklore.ee>   5 In comp.arch Alex Colvin <alexc@world.std.com> wrote:hH >>> I remember when some folks pooh-poohed Virtual Memory because of its= >>> overhead and performance concerns.  Now it is everywhere.o > : > what made VM work was lowering the price of real memory. > $ > i've heard a remark something like > 0 >        Virtual memory is good for many things.3 >        Simulating real memory is not one of them.g >  > [anyone kow the source?] > I > sure, you can access more memory than you own. but it's no way to live.m >   J No, its the *ONLY* way to live. It may not be needed in a small amount of G "only one process running at a time, ever" situations, but in all other-D places, virtual memory and virtual memory based sharing is vital for normal operation.   F If I run 10 xterms I very much want them to take up 2 times the amount> running a single xterm would take and not 10 times the amount.   --   	Sander6   +++ Out of cheese error +++    ------------------------------   Date: 9 Apr 2002 12:10:37 GMT ( From: nmm1@cus.cam.ac.uk (Nick Maclaren)  Subject: Re: Blade architectures0 Message-ID: <a8ulnt$4l1$1@pegasus.csx.cam.ac.uk>  3 In article <1018353514.731362@haldjas.folklore.ee>,D1 Sander Vesik <sander@haldjas.folklore.ee> writes:n8 |> In comp.arch Alex Colvin <alexc@world.std.com> wrote:K |> >>> I remember when some folks pooh-poohed Virtual Memory because of itsj@ |> >>> overhead and performance concerns.  Now it is everywhere. |> >  = |> > what made VM work was lowering the price of real memory.  |> > >' |> > i've heard a remark something likep |> > t3 |> >        Virtual memory is good for many things.o6 |> >        Simulating real memory is not one of them. |> > s |> > [anyone kow the source?]m |> > hL |> > sure, you can access more memory than you own. but it's no way to live. |> aM |> No, its the *ONLY* way to live. It may not be needed in a small amount of tJ |> "only one process running at a time, ever" situations, but in all otherG |> places, virtual memory and virtual memory based sharing is vital foro |> normal operation. |> tI |> If I run 10 xterms I very much want them to take up 2 times the amount A |> running a single xterm would take and not 10 times the amount.r  = Yet another argument based on the interpretation of words :-)m  A I doubt that your are disagreeing, except in the exact meaning ofh< "access more memory than you own".  That quote is definitely? referring to the original claim that virtual memory allowed yous? to have programs that used more workspace than you had physicalp; memory.  As early tests showed, that works only for trivial<
 examples ....e  A Yes, one of the better uses of virtual memory is for data sharingeA (both read-only and read/write).  You don't NEED it for that, butF it makes things a lot easier.      Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679e   ------------------------------   Date: 9 Apr 2002 08:27:26 -0500n- From: koehler@encompasserve.org (Bob Koehler)s  Subject: RE: Blade architectures3 Message-ID: <IBdm623qDWSJ@eisner.encompasserve.org>5  _ In article <CIEJLCMNHNNDLLOOGNJIAEHBELAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes: 3 > Try moving an ethernet cable in a running clusters  G    Been there, done that.  On an LAVC I found I could keep the ethernet H    open for about 3 seconds without any of the systems deciding to leaveK    the cluster, just time enough to add or remove one piece of thin wire.  n*    Sure made the consoles wake up, though.   ------------------------------  * Date: Tue, 9 Apr 2002 15:15:14 +0000 (UTC)" From: greenaum@BOLLOCKSyahoo.co.uk  Subject: Re: Blade architectures3 Message-ID: <3cbaf5bf.5286024@news.btopenworld.com>   5 On Tue, 9 Apr 2002 11:58:35 +0000 (UTC), Sander VesikM& <sander@haldjas.folklore.ee> sprachen:  G >If I run 10 xterms I very much want them to take up 2 times the amount-? >running a single xterm would take and not 10 times the amount.e  B True, but that's not the same as paging to disk, you could (and it: would be better!) do that with just virtual memory in RAM.  E The 2 are separate, but one's sort-of an obvious progression from the  other.  A IMO paging is a good idea if you've got a database etc, but as an D alternative filing system. An easy way of keeping everything indexedB without having to do it yourself. Which if you did, would sorta be. implementing paging in the application anyway.  H ------------------------------------------------------------------------  H if love is a drug, then, ideally, it's a healing, healthful drug... it's@ kind of like prozac is supposed to work (without the sexual side5 effects and long-term damage to the brain and psyche)    ------------------------------  # Date: Tue, 09 Apr 2002 15:49:10 GMT.' From: CBFalconer <cbfalconer@yahoo.com>   Subject: Re: Blade architectures) Message-ID: <3CB30B54.C5BB5258@yahoo.com>r   Nick Maclaren wrote: >  ... snip ... > C > I doubt that your are disagreeing, except in the exact meaning ofr> > "access more memory than you own".  That quote is definitelyA > referring to the original claim that virtual memory allowed youeA > to have programs that used more workspace than you had physicalh= > memory.  As early tests showed, that works only for trivialr > examples ....t > C > Yes, one of the better uses of virtual memory is for data sharingwC > (both read-only and read/write).  You don't NEED it for that, butd > it makes things a lot easier..  > VM is simply a way of managing the memory you do have, without9 thinking too much about it. Any decent system will keep a-@ dynamically selected set of pages or segments in real memory for4 fast access, and try to minimize the disk thrashing.   -- t< Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net);    Available for consulting/temporary embedded and systems.n:    <http://cbfalconer.home.att.net>  USE worldnet address!   ------------------------------   Date: 9 Apr 2002 15:43:41 GMTo  From: hack@watson.ibm.com (hack)  Subject: Re: Blade architectures+ Message-ID: <a8v27d$prs$1@news.btv.ibm.com>s  5 In article <3cb1bab0$0$347$9b622d9e@news.freenet.de>,s, Patrick Schaaf <mailer-daemon@bof.de> wrote:# >hack@watson.ibm.com (hack) writes:oJ >>The *protocol* (TCP) includes bounded time-outs on the order of 45s, andE >>these are necessary because of the short (16-bit) sequence numbers.n >a& >Huh? TCP sequence numbers are 32 bit.  C Right -- 16 bits easily wrap in under 45s.  I meant "32 bits is nottD enough", sorry.  16 bits would have been ludicrous even in the 60's.  F It's been a while since I looked at our protocols in detail.  ThinkingE about this I realise that sequence numbers were not the issue anyway.aH TCP private port numbers *are* 16-bit, and that is definitely too small.  G Our protocols had 64-bit "channel" identifiers so they could be assumedgF to have indefinite lifetime, beyond the death of an instance.  This isE only indirectly related to time-out, but is directly related to being F able to distinguish new from resurrected connections.  In this contextF sequence numbers can be quite short, since they can be reused during aF transmission.  They simply bound the number of unacknowledged packets,E and they don't wrap -- they saturate (i.e. the last packet of a group E is transmitted over and over again, ever more slowly, until an ack istD received giving a list of missing packets for the group).  Pacing isA achieved by varying the interpacket delay (increases if lossy andaE decreases if lossless) as opposed to modulating the "window" (the maxr! amount of unacknowledged data).      Thanks for your other comments.    Michel.r   ------------------------------   Date: 9 Apr 2002 16:12:39 GMTT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)  Subject: Re: Blade architectures0 Message-ID: <a8v3tn$hdn$1@pegasus.csx.cam.ac.uk>  ) In article <3CB30B54.C5BB5258@yahoo.com>,+) CBFalconer <cbfalconer@yahoo.com> writes:  |> Nick Maclaren wrote:w |> > sF |> > Yes, one of the better uses of virtual memory is for data sharingF |> > (both read-only and read/write).  You don't NEED it for that, but" |> > it makes things a lot easier. |> lA |> VM is simply a way of managing the memory you do have, withouth< |> thinking too much about it. Any decent system will keep aC |> dynamically selected set of pages or segments in real memory fore7 |> fast access, and try to minimize the disk thrashing.n  ' Yes, that is so - but it wasn't always!l  A Back when virtual memory was first used, it was occasionally usedr@ to run programs that needed more memory that physically existed.@ There are lots of apocryphal stories, such as the one when a newA virtual memory system was installed and the customer ran the mainh@ benchmark code.  The trouble is that it was a sort, that started@ off by grabbing all of memory and using it for in-store sorting!> I was told that by someone who claimed to have been told it by' someone who claimed to have seen it :-)o  8 I can give a first-hand example, but it is not so pithy.  A The original quote is from that era and is a warning against thate sort of practice.      Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679    ------------------------------  % Date: Tue, 09 Apr 2002 13:14:41 -0400 - From: "Alan T. Bowler" <atbowler@thinkage.ca>n  Subject: Re: Blade architectures+ Message-ID: <3CB32181.80E01543@thinkage.ca><   Sander Vesik wrote:s   > H > If I run 10 xterms I very much want them to take up 2 times the amount@ > running a single xterm would take and not 10 times the amount.  ; True, that's what you want, but experience seems to be that = by the time things were done, you needed 20 times the memory.M# Fortunately, real memory got cheap.d   ------------------------------  # Date: Tue, 09 Apr 2002 17:49:20 GMTn* From: "Bill Todd" <billtodd@metrocast.net>  Subject: Re: Blade architectures@ Message-ID: <AQFs8.74234$w7.6153777@bin6.nnrp.aus1.giganews.com>  / <greenaum@BOLLOCKSyahoo.co.uk> wrote in messageo- news:3cbaf5bf.5286024@news.btopenworld.com... 7 > On Tue, 9 Apr 2002 11:58:35 +0000 (UTC), Sander Vesik ( > <sander@haldjas.folklore.ee> sprachen: > I > >If I run 10 xterms I very much want them to take up 2 times the amountrA > >running a single xterm would take and not 10 times the amount.y >oD > True, but that's not the same as paging to disk, you could (and it< > would be better!) do that with just virtual memory in RAM. >oG > The 2 are separate, but one's sort-of an obvious progression from theh > other.  I Indeed.  DEC's PDP-11 operating systems didn't page, but they could sharerL (pure, static, read-only) code and data sections among multiple instances of? a program (actually, I can't remember now whether they did thisaK automatically, but if not they certainly allowed such sections to be loadedr as a sharable library).    >lC > IMO paging is a good idea if you've got a database etc, but as anrF > alternative filing system. An easy way of keeping everything indexedD > without having to do it yourself. Which if you did, would sorta be0 > implementing paging in the application anyway.  D No:  OS-controlled paging may be easy, but it tends to be relativelyF expensive in performance because the general page-replacement policiesL implemented by the OS are completely unaware of application access patterns.K When you 'do it yourself' (using overlay mechanisms that take at least somepK of the effort out of doing so), you control both the (variable) granularity L and (to some degree) the timing of what you page in, and you control what it
 displaces.  H That said, good OS virtual memory managers aren't horrible, and creatingK ideal overlay structures in complex programs involves decidedly non-trivial I effort.  And to at least some degree VM mechanisms allow a program to userH more physical memory when it's available and thus are more flexible than/ (normal) statically-defined overlay mechanisms."   - bill   ------------------------------  % Date: Tue, 09 Apr 2002 12:11:36 -0500o From: awvov@mail.ru F Subject: Boost Your Windows Reliability!!                         4884> Message-ID: <000041553cc9$0000154a$00001314@lib.kochi-u.ac.jp>   <html> <body>   <p>Dear Windows User,<br>M <br>K Now you can boost the reliability of ordinary Windows 95, 98 and ME to nea=s rly K the level of Windows NT or 2000, Microsoft's professional and industrial v=  ersion of Windows.<br>n <br>K The new WinFix is a very effective way to improve the reliability of Windo=  ws,:? because it makes Windows fault-tolerant and self-repairing.<br>  <br>K And WinFix is very safe, because it operates completely independent of Win= 	 dows.<br>U <br>K <u><font color=3D"#0000ff"><a href=3D"http://biz02873.hostu4free.net">CLIC=tK K HERE</a></font></u> to find out more about WinFix, the safest, most effe=-G ctive way to keep you working, by keeping your PC working non-stop.<br>i <br> Arlen Dixon, CEO<br> Pinewood Software Marketing<br>a <br>; -------------------------------------------------------<br>s <br>C This announcement is intended for PC users who asked to be kept<br>e: informed about new developments in Windows technology.<br> </p>I <p><font size=3D"1">We apologize for any email you may have inadvertentlyl
 received.<br>cK Please <a href=3D"http://biz02873.hostu4free.net/remove.htm">CLICK HERE</a=e0 > to be removed from future mailings.</font></p> </body>  </html>f   ------------------------------   Date: 9 Apr 2002 08:24:55 -0500e- From: koehler@encompasserve.org (Bob Koehler)  Subject: Re: Daylight savingsn3 Message-ID: <aLFCoXBlvCq4@eisner.encompasserve.org>e  V In article <3CB1BAA3.6070401@compaq.com>, John Reagan <john.reagan@compaq.com> writes: > Charlie Hammond wrote:C >> OpenVMS systems prior to V7.3 do not have a mechanism to change bE >> between standard and daylight time.  DTSS can do this if installed  >> on these systemsa >  > Hmmm?3 > J > My V7.2-1 system (I don't have DTSS installed) changed automatically on  > Sunday.  How am I special? > 
 > John Reaganc  F    Maybe like many of us you're using NTP?  DTSS took care of it on my/    home cluster and Multinet XNTP at my office..   ------------------------------  % Date: Tue, 09 Apr 2002 12:17:59 -0400 * From: John Reagan <john.reagan@compaq.com> Subject: Re: Daylight savings1) Message-ID: <3CB31437.4070807@compaq.com>    Bob Koehler wrote:   >  > H >    Maybe like many of us you're using NTP?  DTSS took care of it on my1 >    home cluster and Multinet XNTP at my office.  >   - Ah, never mind.  I found my DTSS process. :-)0   -- i John Reagant' Compaq Pascal/{A|I}MACRO Project Leader    ------------------------------   Date: 9 Apr 2002 05:58:11 -0700n' From: h_zhang98@yahoo.com (Roger Zhang),+ Subject: DCL and batch procedure conversion < Message-ID: <b07858c7.0204090458.c953fb8@posting.google.com>   Hello,C Anyone here successfully migrated their DCL and batch procedures toUE Windows NT? If you have done so and wouldn't mind telling the rest ofh; us about the general methods you have used that'd be great.3D Information like name of any NT simulation package used for VMS-likeB batch queues, language in which any converting program is written,, such as Perl, etc. is very much appreciated. Thanks,. Rogero   ------------------------------   Date: 9 Apr 2002 08:57:28 -0500(- From: koehler@encompasserve.org (Bob Koehler)o/ Subject: Re: DCL and batch procedure conversion.3 Message-ID: <o2IOYIr2kgtP@eisner.encompasserve.org>f  f In article <b07858c7.0204090458.c953fb8@posting.google.com>, h_zhang98@yahoo.com (Roger Zhang) writes: > Hello,E > Anyone here successfully migrated their DCL and batch procedures toGG > Windows NT? If you have done so and wouldn't mind telling the rest of = > us about the general methods you have used that'd be great.hF > Information like name of any NT simulation package used for VMS-likeD > batch queues, language in which any converting program is written,. > such as Perl, etc. is very much appreciated.	 > Thanks,c > Rogers  :    Tried it, gave up.  Migrated from VAX to Alpha instead.   ------------------------------   Date: 9 Apr 2002 04:26:37 -0700e& From: andysands@yahoo.com (Andy Sands) Subject: Exporting SYSUAF.LISd= Message-ID: <9bbdf680.0204090326.11a2aac9@posting.google.com>   A I asked a client to use the following script to download the usermC profiles from their VMS box to allow me to perform some audit work.g   Prompt		Commandt $ 		SET DEFAULT SYS$SYSTEM $ 		RUN AUTHORIZEy UAF> 		LIST [*,*]/FULL UAF> 		EXIT   D The SYSUAF.LIS file that I have been given has no structure in it atF all (no carriage returns specified etc) so cannot be read into a fixedC width interpreter like Monarch or a delimited file interpreter liket
 ACL/Excel.  E I have received VMS downloads of these files in the past with clearlyeE structured records in them.  Does anyone know what has gone wrong and E if so what is the correct command syntax for obtaining the sysuaf.lisn files in a structured form?u   Thanks in advance,   Andy   ------------------------------  % Date: Tue, 09 Apr 2002 08:26:30 -0400s2 From: rdeininger@mindspring.com (Robert Deininger)! Subject: Re: Exporting SYSUAF.LIS@K Message-ID: <rdeininger-0904020826310001@1cust169.tnt1.nashua.nh.da.uu.net>D  = In article <9bbdf680.0204090326.11a2aac9@posting.google.com>,a' andysands@yahoo.com (Andy Sands) wrote:t  B >I asked a client to use the following script to download the userD >profiles from their VMS box to allow me to perform some audit work. >r >Prompt          Command' >$               SET DEFAULT SYS$SYSTEMe >$               RUN AUTHORIZE  >UAF>            LIST [*,*]/FULL >UAF>            EXITo    This looks perfectly reasonable.    E >The SYSUAF.LIS file that I have been given has no structure in it attG >all (no carriage returns specified etc) so cannot be read into a fixed D >width interpreter like Monarch or a delimited file interpreter like >ACL/Excel.e > F >I have received VMS downloads of these files in the past with clearlyF >structured records in them.  Does anyone know what has gone wrong andF >if so what is the correct command syntax for obtaining the sysuaf.lis >files in a structured form?  G Was the file transferred via ftp?  I would guess either binary transfert> mode was used, or the ftp client (or server) is simply busted.  - Always use ASCII mode to transfer text files.   B When possible, avoid the "ftp client" that is included in many webG browsers.  Most of them are stupid.  I get better results with real ftpt clients.  YMMV.l  I If you used something other than ftp, I don't know what to suggest.  More  details would be useful here.    ------------------------------  # Date: Tue, 09 Apr 2002 12:35:34 GMTe' From: "Hans Vlems" <hvlems@zfree.co.nz> ! Subject: Re: Exporting SYSUAF.LISo$ Message-ID: <3cb2e013$1@zfree.co.nz>  < Just had a similar problem with a DECwindows/Motif LOG file.6 Remedy: import it into Word and save it as a TXT file. YMMV though,  ' andysands@yahoo.com (Andy Sands) wrote:cB >I asked a client to use the following script to download the userD >profiles from their VMS box to allow me to perform some audit work. >e >Prompt		Command >$ 		SET DEFAULT SYS$SYSTEM: >$ 		RUN AUTHORIZE >UAF> 		LIST [*,*]/FULLS >UAF> 		EXIT >sE >The SYSUAF.LIS file that I have been given has no structure in it at G >all (no carriage returns specified etc) so cannot be read into a fixed D >width interpreter like Monarch or a delimited file interpreter like >ACL/Excel.e >hF >I have received VMS downloads of these files in the past with clearlyF >structured records in them.  Does anyone know what has gone wrong andF >if so what is the correct command syntax for obtaining the sysuaf.lis >files in a structured form? >, >Thanks in advance,+ >0 >Andy        http://www.zfree.co.nz   ------------------------------  % Date: Tue, 09 Apr 2002 14:05:36 +0100b' From: Elliott Roper <elliott@yrl.co.uk>:! Subject: Re: Exporting SYSUAF.LISs2 Message-ID: <090420021405364784%elliott@yrl.co.uk>  B In article <9bbdf680.0204090326.11a2aac9@posting.google.com>, Andy" Sands <andysands@yahoo.com> wrote:  C > I asked a client to use the following script to download the user,E > profiles from their VMS box to allow me to perform some audit work.i >  > Prompt    Commanda > $      SET DEFAULT SYS$SYSTEMp > $      RUN AUTHORIZE > UAF>      LIST [*,*]/FULLi > UAF>      EXIT > F > The SYSUAF.LIS file that I have been given has no structure in it atH > all (no carriage returns specified etc) so cannot be read into a fixedE > width interpreter like Monarch or a delimited file interpreter like- > ACL/Excel. > G > I have received VMS downloads of these files in the past with clearlyeG > structured records in them.  Does anyone know what has gone wrong andsG > if so what is the correct command syntax for obtaining the sysuaf.list > files in a structured form?- >  >-D That should have worked. I suspect the .LIS got broken on the way to0 your Billybox. Try ftp-ing the original as text.   ------------------------------  % Date: Tue, 09 Apr 2002 16:55:26 +0200 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>n! Subject: Re: FTP captive account? ) Message-ID: <3CB300DD.E4F61C4B@gtech.com>q   Tom Simpson wrote:I > I would like to setup FTP accounts for clients that need to drop off or  > pickupL > files from/to our ES40 cluster (OpenVMS 7.2-1, TCP 5.0a).  The problem I'mI > trying to solve is setting up the proper security.  I want to give eacheM > client their own user name and password and FTP file area and restrict themiN > to only that area.  As it seems to work now, once they log in, they are freeG > to CD to anywhere they want, restricted only by the file protections.mL > Giving them unique UICs will solve some of the problem, but not all of it.L > The FTP Anonymous account does what I need as far as effectively disablingL > the CD command, but this is a special case for the Anonymous account only.I > It protects our system from "browsing" clients, but I can see no way toc& > protect the clients from each other. > M > Is there any way to do this or is there a third party software package thata > will do this?   C Install HGFTP (previously MGFTP) on top of TCP/IP (previously UCX).s  : That FTP server supports multiple anonymous usernames with: different rigths. I think it should be possible to make it$ fullfill your specific requirements.   Arne   ------------------------------  $ Date: Tue, 9 Apr 2002 10:28:12 -0400; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>n3 Subject: Re: Help! VT100 emulators - some problems. $ Message-ID: <3cb2faf4$1@news.si.com>  J >It's been many years, but I'm pretty sure that the VT200 and VT300 didn't& >have a mechanism to "invert" letters.  F There was a command procedure on a DECUS tape long ago that used SIXELD programming to change all the characters on the screen of a VT to be9 backward (or upside down, I can't recall which any more).  --A Brian Tillman                   Internet: tillman_brian at si.comoA Smiths Aerospace                          tillman at swdev.si.coms= 3290 Patterson Ave. SE, MS      Addresses modified to preventt< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Tue, 09 Apr 2002 19:24:57 +0010e% From: paddy.o'brien@zzz.tg.nsw.gov.aud4 Subject: Re: How to know if EDS hires for VMS Eng. ?5 Message-ID: <01KGDD5PBXLE000DYE@tgmail.tg.nsw.gov.au>l  K I *had* to reply to this, even though it's a bit old -- I've just got back ) from 5 weeks away.  L Incidentally, I do not understand how Didier's original post related to the 
 subject line.    JF Mezei wrote:w >Paul Sture wrote:G >> Agreed. One limitation is that comments take a full line - you can'tiG >> simply append a comment to the end of a line, which can lead to hugeT# >> chunks of code without comments.a  N Lost track of Paul's reply and am assuming it is still about Cobol, but isn't H that what C programmers are guilty of.  I have seen little C codes with   greater than 0 lines of comment.  M >But COBOL code is much easier to understand on a line by line basis so there + >is a bit less need for line-documentation.r  B I do not subscribe to the belief that in any language code can be 0 self-documenting -- well perhaps Knuth's WEB :-)  N >I still remember at school, one person got a 0 for his assignment even thoughN >it produced the right results. The names he had used for variables/paragraphs >yielded stuff such as:o1 >	PERFORM AROUSAL THROUGH ORGASM UNTIL EXHAUSTED.  > G >and:	MOVE HANDS TO BREASTS, MOVE <male bodypart> TO <female bodypart>.a >nM >Interestingly, the output was perfectly serious, but the program, while veryhF >fun to read, was incomprehensible because the variable names did not 
 >correctly >reflect they contents.u  O Since I'm guessing that JF was at school no later than the 1980's,  I can only tJ surmise that the 0 mark was given by a humourless prude, and from prudish M spite.  Did the assignment state that variable nomenclature must be relevant o to the variable's function?i  J In my main language (Fortran), back then there was a 6-character limit on N variable names and they were, of essence, cryptic.  Since by default I-N were O integers and the rest reals, many variables which would have had integer names eJ beginning with C were re-written with K.  I still write many mathematical O routines using simple and short names, much as I would writing a formula.  For ,L geometry, I am happy to use X and Y as variable names for co-ordinates in a F succinct and well-defined subroutine.  Or A, B and C in simple matrix O manipulation routines.  I find brevity of names easier to read in mathematical t1 routines, akin to the formulae I would handwrite.e  I What, in a verbose way, I am trying to say is that variable names do not  I really matter.  If variables HANDS and BREASTS are used consistently, it iN should not take any program maintainer with an IQ greater than semi-moron any N time to understand the routine.  I remember having to correct a major problem O in a largish routine with about 50 variables and none with a name of more than 7  two characters: I1, I2, I3, etc.   Regards, Paddy   ------------------------------  # Date: Tue, 09 Apr 2002 16:13:27 GMTm% From: jlsue <jlsuexxxz@insightbb.com>0O Subject: Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinux 8 Message-ID: <4n46bu8nmv6h5k3gtcq2lkd5o93oe15olb@4ax.com>  E On Mon, 08 Apr 2002 16:23:55 +0100, Andrew Harrison SUNUK Consultancy 3 <andrew_nospam.remove_this.harrison@sun.com> wrote:,   >c >d
 >jlsue wrote:l >e >t >>  F >> I DO remember my background and it is, by and large, my challengingG >> you on technical FUD & BS about VMS & VMSclusters that you spout offsE >> in here.  My technical experience with VMS and VMSclusters againstaI >> your FUD will win every time.  And you have yet to address any of youraI >> analysis problems that I have raised on these subjects - which usuallyt@ >> consists of finding ONE (or a smallish group < 10) person whoG >> complains in very vague terms about some issue, and you harp on thataH >> as if it's the rule rather than the exception.  I have challenged youH >> time and again to provide more evidence than this small, self-skewing  >> data set, but you never have. >> n >e >p >a9 >Check deja, you will find that you didn't win any of thei7 >arguments you claim to have done. In fact you ended upe9 >getting corrected by Bill a couple of times and the restO: >of the time you appear to think that the person who posts! >the last ad-hominem attack wins.l >m9 >As I said earlier on your posting style is very like thei >lamented Christof keep it up.    A Hey bozo, you make the accusations, it's up to you to provide theiF evidence.  I'm certainly not doing your work for you.  But if you wantC to have me e-mail you several megs of messages, just say the word. i  1 Not speaking for anyone, certainly not DEC/Compaqf- (get rid of the xxxz in my address to e-mail)s   ------------------------------  # Date: Tue, 09 Apr 2002 16:20:50 GMTu% From: jlsue <jlsuexxxz@insightbb.com>iO Subject: Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinux 8 Message-ID: <6q46bukn97ac1tp544hu5iq0bh5fu1o85a@4ax.com>  E On Mon, 08 Apr 2002 18:06:24 +0100, Andrew Harrison SUNUK Consultancy43 <andrew_nospam.remove_this.harrison@sun.com> wrote:h   >c >> >Paul Sture wrote:   >> aL >> Where are your facts? Evidence? URLs will often do. Inquiring minds wish 
 >> to know...M >> n >s >l  # >My stance on WildFire performance:c9 >      My claim that it would require OPS in a box to gets; >      half decent performance, proved right by Compaqs ownv3 >      benchmark results and the number of returns.W  C This is one of the biggest laughs of your so-called analysis.  As IyF stated in an earlier discussion about this very topic (which I, again,F offer to send you if you wish), you picked one (or maybe two) specificE examples to try to prove your opinion.  But you completely ignore allfD the data of the many, many satisfied (nay, very happy) customers who1 use these systems *without* needing OPS in a box.i  F If you'll recall, the last time I even went through the effort to do aB google search and could only find ONE note in this group that came> close to supporting your opinion.  And even that one was NOT aC clear-cut case because it wasn't at all clear that he'd even gottenf5 his system  tuned properly after the initial install.c  F So, once again, if you're going to claim that OPS-in-a-box is required@ to get the performance, it is incumbent upon you to provide that
 evidence.    >-7 >There are numerous other examples as you know well and09 >in each of these cases members of the choir claimed that:8 >I was the person providing the BS, we all know that the >opposite was the case.a  E If there are 'numerous examples', then it shouldn't be that difficult>B for you to present that evidence.  If you're not going to do that,D then just shut up because your own position in this discussion is at& least as biased as anyone from Compaq.   > 7 >One memorable example was a choir member claiming thato8 >Compaq would not need to use OPS in a box to get decent@ >OLTP performance from WildFire, the claimant being unaware that> >Compaqs own TPC results used exactly this trick. I was called: >a FUDSTER which I now take as a compliment. In this group; >anything that a choir member calls FUD almost always turnso >out to be the truth.c >a  D Again, provide this proof.  I have no idea what you're talking about: here, so I'd be willing to review any of this information.  E Time and again you make claims that you do not back up with evidence.eB If you want us support his obviously high opinion you have of your: abilities, please allow us to enjoy these winning moments.  1 Not speaking for anyone, certainly not DEC/Compaq-- (get rid of the xxxz in my address to e-mail)    ------------------------------  # Date: Tue, 09 Apr 2002 16:11:40 GMTe% From: jlsue <jlsuexxxz@insightbb.com>8O Subject: Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxC8 Message-ID: <so36bu01vr12uqms0it1i2fo33vm2demd3@4ax.com>  E On Mon, 08 Apr 2002 16:17:05 +0100, Andrew Harrison SUNUK Consultancyu3 <andrew_nospam.remove_this.harrison@sun.com> wrote:f   >N >i
 >jlsue wrote:e >eH >> On Fri, 05 Apr 2002 16:59:39 +0100, Andrew Harrison SUNUK Consultancy6 >> <andrew_nospam.remove_this.harrison@sun.com> wrote:  6 >>>Ohh I thought not that would expose you to ridicule& >>>allong with your fellow Choristers. >>>n >>>s >> .B >> Examples of what, exactly?  Can you pose a coherent question or, >> request so that I can actually answer it? >> g >  >n< >Ok it was a coeherant question and one that all the readers7 >of this newsgroup understood perfectly just as you dide, >but let me reprase it for you specifically. >r4 >Did you consistently and systematically missled the= >inhabitants of this newsgroup about the status and future ofs >the Alpha processor ? >e@ >Did you consistently and systematically missled the inhabitants7 >of this newsgroup about the market position of Alpha ?i >p@ >Did you consistently and systematically missled the inhabitants< >of this newsgroup about the performance of Alpha systems in2 >particular I refer to the GS320 performance hype. >-: >Did you consistently and systematically flame posters who >expressed a counter view ?p >iA >Do you now have any missgivings having discovered belatedly thato6 >the people you flamed were right and you were wrong ? >		   F I believe the answer to all these questions is a 'NO'.  I don't recallE putting too much marketing fluff in my posts.  Mostly my intent is to B comment on the so-called analysis that I see in here (when I'm notF commenting directly on the technology I have first-hand knowledge of).C And yet, I certainly don't recall anyone else in here being missledeD [sic] about any of this either.  Certainly any performance claims of@ the GS320 are at *least* as valid as your claims that it doesn'tF perform (and yes, I've got many of these old messages saved as well if you'd like copies of them).    >n >yE >> Your entire rant is chock full of non sequitors, so it serves as an >> fine example in itself. >> -E >> Now, if you want examples that prove my previous points about your F >> hypocrisy in that you began the name calling in this thread... justG >> let me know and I'll exract the entire thing and send it to ya.  I'mhG >> not about to re-post it here:  It's entirely too long and the drivelo( >> was bad enough the first time around. >>   >I >k> >As I said earlier it ill behoves anyone from Compaq to accuse@ >anyone of hypocracy. If you had an ounce of concience you would, >realise this and quit while you are behind.  E I see.  Once again all you can do is make grandiose claims of "anyone A from Compaq" and not quote specific facts to back up your claims.r( This is typical Andrew discussion logic.   >e? >Just to illustrate this I have just met with the CIO of one of @ >your customers, he is consolidating all his existing datacenterB >applications that currently run on Alpha (OpenVMS and Tru64) onto0 >either IBM or Sun, he hasn't decided which yet. >e	 >Reasons.c >r> >Nothing wrong with the platform but as a long term Alpha user< >he has decided to move from Alpha to something else because9 >the Alphacide left him with a bad taste in his mouth andh> >looking exposed (he chose Alpha origionally) and the takeover> >of Compaq by HP was the final nail in the coffin. He will not= >be buying any more Alpha servers or Compaq x86 boxes either.s  E So what?  Do you think we couldn't find specific examples of just thesD opposite (i.e., migrating from Sun)?  Hand-picking your data doesn't9 make your so-called "statistical" analysis more accurate.w  E But, what the heck does that all have to do with me (or Fred for that-D matter) anyway?  More specifically, how does your logic go from thisD customer's attitudes about Alpha to saying that we're all hypocrites (note the spelling)?   >o= >It was activities like the ones indulged in by you and otherf< >members of the choir that put him in the possition he is in: >now ironically he remarked that he didn't realise that 25  >years could slip by so quickly.  F Bull S****!  Nobody in here put him in any position.  And his positionD has not changed ONE IOTA with any of the events you name above.  TheC fact is that any AlphaServer he bought today would still be working > just great 10 - hell, probably 20 - years from now.  This is aF smokescreen of incongrous information stitched together to try to drawE some bizarre conclusion (that, interestingly enough, just plays rightp into your person opinion).   >h: >People like you who would like to pretend that no one was: >missled either deliberately or though lack of scrutiny of9 >the information they were psssing on, a bit more humiltya= >on your part might make some of your soon to be ex customers,
 >re-consider.v >e  @ Still waiting for facts on the missled part.  At the very worst,= Compaq was no more misleading than Sun in their own marketingiE materials.  And more than likely, this customer wasn't mislead in anysD way that really affects his business.  There is no evidence that his4 previous decisions on Alpha have caused any problem.  D You're on your typical high-horse again, and once again using prettyC loose data collection and interpretation techniques to try to provem your own opinions.  E People won't have to like the decision on Alpha.  Hell, I'm sure many B don't.... but that doesn't change the fact that it is still a veryE powerful system for most applications.  And I can tell you that, eveniB with all this stuff you're throwing about, we're selling $millionsF (US) in Alphaservers, both Tru64 and VMS.  So obviously there are some/ who don't hold the same attitude as your littleM! example-to-try-to-prove-the-rule.a    1 Not speaking for anyone, certainly not DEC/Compaqn- (get rid of the xxxz in my address to e-mail)t   ------------------------------    Date: 09 Apr 2002 03:23:23 -04000 From: cstacy@theworld.com (Christopher C. Stacy)  Subject: Re: IA64 is not the VAX( Message-ID: <uhemlfi8k.fsf@theworld.com>  @ >>>>> On 08 Apr 2002 18:56:43 -0700, Eric Smith ("Eric") writes:    Eric> I wrote::J  >> Trying to use special-purpose hardware to compete with general purposeL  >> CPUs is like trying to run in front of a steamroller.  You might be ableH  >> to do it for a short while, but you're going to get tired before the@  >> steamroller does.  This is why LISP machines were ultimately  >> unsuccessful.,  4  Eric> Tim Shoppa <shoppa@trailing-edge.com> writes:K  >> But to most embedded-type folks a LISP machine is not a special-purposee,  >> machine, it's a general-purpose machine.  K  Eric> It depends on your perspective.  But a VAX or Sun was certainly more J  Eric> general-purpose than a LISP machine.  Not enough people wanted LISPK  Eric> machines, so the engineering costs couldn't be amortized over enoughi=  Eric> units to make them both cost-effective and profitable.c  C I'm not sure where you get the idea that the Lisp Machine companiesAB folded due to engineering costs.  That's not what happened at all.? The problem had more to do with mis-management of the companiestD in areas like real estate investment, and certain failures of focus.F The main feature of the Lisp Machine systems was always the software, E not the hardware, but there was management confusion about that, too.lD It's quite a long and complicated saga, but it had very little to do with the technology itself.u  G The Lisp Machines were strictly single-user workstations, so it doesn't I make total sense to compare them to the VAX, which used by most customersnF as a timesharing system.  On the other hand, very atypically, we had aD farm of VAX-11/750s used as individual workstations at MIT, and theyC were far less powerful than the Lisp Machines.  I don't remember ifnD the early SUN systems were typically single-user workstations (whichF is what I saw) or timesharing machines.  But the Lisp Machine was also+ a much more powerful machine than the SUNs.   A In the context of a workstation, the Lisp Machine was certainly anB fully "general purpose" system -- what would type of computing do + you imagine was not possible to do with it?s  E If it's the word "Lisp", maybe you're unaware that the "Lisp" Machine2F also included FORTRAN, C, ADA, PASCAL, PROLOG, MACSYMA, and some otherD languages, and that it interoperated over the network with all knownE operating systems (that is, transparent network file access using the I foreign host's own native file system syntax and semantics, automaticallyBB selecting the various protocols such as FTP, NFILE, NFS, etc; alsoB supporting all standard network protocols such as TELNET and RPC).  B I am pretty sure the Lisp Machine was the first system to have theA attributes cited in the original message here (the original XeroxA@ machines were much less powerful by comparison), and it was the 4 first such workstation to be available commercially.  < When I say "powerful", I'm talking about number of features,< computing speed, memory size, disk size, network, graphics, 6 and of course the software (which was the main thing).  < Lisp Machines were used for an infinite variety of problems.  B During the Gulf War, I remember watching on TV and being amused toD note that the bombing sorties (in particular, but along with lots ofC other strategic operations) were being scheduled by a Lisp machine,dC and the military was communicating over networks that were designedhE using a Lisp Machine, and that even the TV network news graphics thateH I was looking at were done on a Lisp Machine.  (Of course, Lisp MachinesD were used for lots more than wars and television...that was just oneE day when they happened to be simltaneously used for multiple purposes>E in everybody's living room, so "general purpose" brought it to mind.)a  7 What do you mean by "general purpose" that was lacking?t= They weren't multi-user timesharing, and they weren't cheap, u if that's what you mean.   ------------------------------    Date: 09 Apr 2002 09:53:35 -0700 From: Greg Finn <finn@isi.edu>  Subject: Re: IA64 is not the VAX( Message-ID: <6wpu18yfsg.fsf@cnn.isi.edu>  2 cstacy@theworld.com (Christopher C. Stacy) writes:  B > >>>>> On 08 Apr 2002 18:56:43 -0700, Eric Smith ("Eric") writes:   >  Eric> I wrote: L >  >> Trying to use special-purpose hardware to compete with general purposeN >  >> CPUs is like trying to run in front of a steamroller.  You might be ableJ >  >> to do it for a short while, but you're going to get tired before theB >  >> steamroller does.  This is why LISP machines were ultimately >  >> unsuccessful.m  6 >  Eric> Tim Shoppa <shoppa@trailing-edge.com> writes:M >  >> But to most embedded-type folks a LISP machine is not a special-purposem. >  >> machine, it's a general-purpose machine.  M >  Eric> It depends on your perspective.  But a VAX or Sun was certainly morehL >  Eric> general-purpose than a LISP machine.  Not enough people wanted LISPM >  Eric> machines, so the engineering costs couldn't be amortized over enoughy? >  Eric> units to make them both cost-effective and profitable.   E > I'm not sure where you get the idea that the Lisp Machine companiestD > folded due to engineering costs.  That's not what happened at all.A > The problem had more to do with mis-management of the companies F > in areas like real estate investment, and certain failures of focus.H > The main feature of the Lisp Machine systems was always the software, G > not the hardware, but there was management confusion about that, too. F > It's quite a long and complicated saga, but it had very little to do > with the technology itself.H   [snip]  D None of the language-specific workstations was what one would call aG commercial success a far as I am aware.  Pascal, Lisp and Mesa machinesdH tried their hands at becoming successful and folded.  There were severalG Lisp workstations.  For example, Symbolics, LMI and Xerox come to mind.e< Something was obviously wrong with all of them commercially.  I Many of these early workstations took a page out of the Xerox PARC method B and tried to optimize CPU performance for a particular language byL microcoding the architecture to suit that language.  While a Xerox DandelionI chassis could run multiple languages, it did so by rebooting into anothern machine architecture.   D That was an attractive design option only so long as microprocessorsG remained small and slow by comparison with a multi-chip, customized CPUeK design.  Also, with an entirely new hardware/software architecture, it tooktK years to build a respectable operating system and suite of useful packages.mG But by that time micros had passed through another generation or two ofsK advancement.  As 1980 rolled around it was already clear that 32-bit microsgH such as the 68000 (@ 8MHz) would soon render bit-slice designs obsolete.  I It's hard to beat the physics advantages of a single-chip microprocessor.n) Eventually, even Cray couldn't beat that.    ------------------------------  # Date: Tue, 09 Apr 2002 06:42:19 GMTi  From: cjt <cheljuba@prodigy.net> Subject: Re: Itanium troubleso+ Message-ID: <3CB28D3A.CD00FC48@prodigy.net>-   bill davidsen wrote: > 2 > In article <a81nml$s8n$1@pegasus.csx.cam.ac.uk>,+ > Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:i > G > | More seriously, what I am referring to is the stage of more-or-less4F > | open shipment of development workstations using more-or-less finalH > | versions of the chip to all and sundry ISVs and in-house developmentI > | teams.  It is VERY hard to keep those secret, and they hadn't startedy. > | when a little bird last cheeped in my ear. > |nD > | What is more, I can tell you that it will take at least 3 monthsD > | (and more likely 6) FOLLOWING that stage before there are enoughE > | third-party applications to make it worth while for non-developernA > | users (i.e. 99% of them) to buy a new system.  And I speak asg$ > | someone in just such a position. > F >   Actually I would think the market at the moment is servers, and ifF > only ONE application is (a) stable and (b) runs faster than the sameF > dollars invested in Pentium whatever, then it makes sense to buy the > more cost effective server.y > I >   If you have a large cluster of {wed,mail,usenet,dns} servers, you run  > what is cost effective.   K That's a superficially attractive statement, but "cost" is a bit vague when K a change of architecture is involved.  Do you include training costs, costsiG of the inevitable instability of any new and unproven system, costs of lN building a spare parts inventory, the cost of having to deal with a new set ofP vendors and tech support people, etc.?  Little things, like the absence of your L favorite software utility, not yet ported, can present a snag whose cost is  hard to quantify.  r  D In some ways, the devil you know is better than the devil you don't.   >  > --; > bill davidsen <davidsen@tmr.com> CTO, TMR Associates, InckK >   Programming without software engineering is like sculpting with a chainhM > saw. The very talented can produce a work of art, the mediocre wind up witheH > a misshapen lump in a pile of rubble, and in neither case does the endE > result have more than a passing resemblance to the original intent.m   ------------------------------   Date: 9 Apr 2002 08:01:32 GMTo( From: nmm1@cus.cam.ac.uk (Nick Maclaren) Subject: Re: Itanium troubles,0 Message-ID: <a8u74s$l2m$1@pegasus.csx.cam.ac.uk>  : In article <a8tpf4$2hvg$1@newssvr05-en0.news.prodigy.com>,' bill davidsen <davidsen@tmr.com> wrote:o1 >In article <a81nml$s8n$1@pegasus.csx.cam.ac.uk>,a* >Nick Maclaren <nmm1@cus.cam.ac.uk> wrote: >iF >| More seriously, what I am referring to is the stage of more-or-lessE >| open shipment of development workstations using more-or-less finalyG >| versions of the chip to all and sundry ISVs and in-house developmentpH >| teams.  It is VERY hard to keep those secret, and they hadn't started- >| when a little bird last cheeped in my ear.i >| rC >| What is more, I can tell you that it will take at least 3 monthssC >| (and more likely 6) FOLLOWING that stage before there are enough0D >| third-party applications to make it worth while for non-developer@ >| users (i.e. 99% of them) to buy a new system.  And I speak as# >| someone in just such a position.i >3E >  Actually I would think the market at the moment is servers, and ifoE >only ONE application is (a) stable and (b) runs faster than the sameeE >dollars invested in Pentium whatever, then it makes sense to buy the, >more cost effective server. >yH >  If you have a large cluster of {wed,mail,usenet,dns} servers, you run >what is cost effective.  B It isn't really cost-effective to buy a unique system for a singleA application, outside HPC, process control and similar areas.  You @ don't need everything, but you do need enough range to cover the? requirements of a target market.  Hence 3 months - if it were ahC complete range of software I wwas referring to, that 3 would be 12.s     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679t   ------------------------------   Date: 9 Apr 2002 08:07:18 GMT ( From: nmm1@cus.cam.ac.uk (Nick Maclaren) Subject: Re: Itanium troublesd0 Message-ID: <a8u7fm$l7l$1@pegasus.csx.cam.ac.uk>  @ In article <NKus8.66823$w7.5656854@bin6.nnrp.aus1.giganews.com>,) Bill Todd <billtodd@metrocast.net> wrote:  >t4 >"bill davidsen" <davidsen@tmr.com> wrote in message5 >news:a8tp4h$40n6$1@newssvr05-en0.news.prodigy.com...v3 >> In article <a81dll$l04$1@pegasus.csx.cam.ac.uk>,h, >> Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:1 >> | In article <3CA3A0AE.538DD287@videotron.ca>,e4 >> | JF Mezei  <jfmezei.spamnot@videotron.ca> wrote: >>: >> | >How much of MS's bloatware has been ported to IA64 ?< >> | >How much of Linux's software has been ported to IA64 ? >> |I >> | Quite a lot and most.  Remember that most Linux software is portablesH >> | and already runs on 64-bit systems.  From an application viewpoint,J >> | the only differences between IA-64 and Alpha are performance and bugs6 >> | related to optimisation and similar complexities. >>K >>   I was under the impression that the Alpha was a big-endian port, whicheK >> presents more chances for "learning experiences" than one might imagine.i >lE >Not sure what you're saying above.  Though both may be very slightlyoK >little-endian-preferring internally, Alpha and Itanic (and for that matteruL >POWER4 and I think SPARC and maybe MIPS) will run either-endian.  Linux wasL >developed primarily on little-endian machines (x86s), runs little-endian onE >Alpha (though it might run big-endian somewhere - I don't know), andiH >presumably will also run little-endian on Itanic.  Windows runs (to the> >degree it runs at all) little-endian everywhere that it runs.  @ Try MIPS, PowerPC and others.  I don't know if Linux will run onA mixed-endian systems or if it has separate switches, but it isn'to3 hard to write ordering-neutral code if you want to.s  = In fact, my experience is that most clean codes are naturallyhA ordering-neutral, and it is only people who insist on playing thetA most disgusting games who have serious troubles.  Of course, notea@ that I include the interfaces of Internet/TCP/etc. networking asA a disgusting design, because they introduce ordering dependenciest& even in places where there is no need.  A This doesn't mean that clean programs that have to munge bits andbB bytes don't have ordering-dependent code, but that there is likely@ to be a small amount of moderately well-isolated code.  Which is8 particularly critical if the bit order can also vary :-)     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679s   ------------------------------  # Date: Tue, 09 Apr 2002 08:27:03 GMT . From: "aaron spink" <aaronspink@earthlink.net> Subject: Re: Itanium troublesfD Message-ID: <rBxs8.2326$3P4.186792@newsread2.prod.itd.earthlink.net>  3 "bill davidsen" <davidsen@tmr.com> wrote in messages4 news:a8tp4h$40n6$1@newssvr05-en0.news.prodigy.com...J >   I was under the impression that the Alpha was a big-endian port, whichJ > presents more chances for "learning experiences" than one might imagine. >SL IIRC, the only in use OS that runs on Alpha in big endian mode is UNICOS/mkTJ on the T3E/T3D.  The only other OS that was planned to run on Alpha in bigK endian mode was the Tandem NonStop Kernal.  I don't how far they got beforebI the "new alpha age" began, but I would assume they at least got something 9 booting.  NT, DUNIX,VMS, and Linux are all little endian.t   Aaron Spinkg speaking for myself    ------------------------------  $ Date: Tue, 9 Apr 2002 12:44:43 +0200/ From: "Maciej W. Rozycki" <macro@ds2.pg.gda.pl>e Subject: Re: Itanium troublesnI Message-ID: <Pine.GSO.3.96.1020409123748.397C-100000@delta.ds2.pg.gda.pl>p  # On 9 Apr 2002, Nick Maclaren wrote:e  B > Try MIPS, PowerPC and others.  I don't know if Linux will run onC > mixed-endian systems or if it has separate switches, but it isn'tn5 > hard to write ordering-neutral code if you want to.e  J  At least on MIPS you can configure Linux to run either endian as for manyF MIPS systems the kernel-mode endianness is fixed by hardware.  ReverseB endianness in the user mode as available on MIPS processors is notJ supported by Linux -- too much hassle to handle for a questionable gain --? endianness of user binaries must match endianness of a kernel.     -- n@ +  Maciej W. Rozycki, Technical University of Gdansk, Poland   +@ +--------------------------------------------------------------+@ +        e-mail: macro@ds2.pg.gda.pl, PGP key available        +   ------------------------------   Date: 9 Apr 2002 10:42:26 GMTo From: Dan.Pop@ifh.de (Dan Pop) Subject: Re: Itanium troublest* Message-ID: <a8ugii$lr7$1@sunnews.cern.ch>  d In <NKus8.66823$w7.5656854@bin6.nnrp.aus1.giganews.com> "Bill Todd" <billtodd@metrocast.net> writes:    L >POWER4 and I think SPARC and maybe MIPS) will run either-endian.  Linux wasL >developed primarily on little-endian machines (x86s), runs little-endian onE >Alpha (though it might run big-endian somewhere - I don't know), andn  B The first Linux port was to m68k, which is a big-endian processor.  2 >presumably will also run little-endian on Itanic.  I Why "presumably" and why the future tense?  Linux has been running littleeG endian on Itanium since 1999.  When the chip was officially released bygG Intel, there were multiple production Linux distributions available forh it, including Red Hat.   Danc -- Dan PopE DESY Zeuthen, RZ group Email: Dan.Pop@ifh.dee   ------------------------------  # Date: Tue, 09 Apr 2002 17:22:49 GMTu* From: "Bill Todd" <billtodd@metrocast.net> Subject: Re: Itanium troublesn@ Message-ID: <JrFs8.73981$w7.6133376@bin6.nnrp.aus1.giganews.com>  + "Dan Pop" <Dan.Pop@ifh.de> wrote in message $ news:a8ugii$lr7$1@sunnews.cern.ch...E > In <NKus8.66823$w7.5656854@bin6.nnrp.aus1.giganews.com> "Bill Todd"s  <billtodd@metrocast.net> writes:   ...f  4 > >presumably will also run little-endian on Itanic. >s, > Why "presumably" and why the future tense?  K Because I have not been sufficiently interested in Itanic or (at least yet)- Linux to know.     Linux has been running littletI > endian on Itanium since 1999.  When the chip was officially released bypI > Intel, there were multiple production Linux distributions available for  > it, including Red Hat.  G And when/if people actually begin to purchase that platform, then Linuxe 'will' run there...n   - bill   >  > Dani > --	 > Dan Pops > DESY Zeuthen, RZ group > Email: Dan.Pop@ifh.dea >t   ------------------------------  % Date: Tue, 09 Apr 2002 08:50:57 +0200t) From: Bart Zorn <B.Zorn@xs4all.nospam.nl> 0 Subject: Re: Just another linux nightmare day #2/ Message-ID: <3CB28F51.5090501@xs4all.nospam.nl>r   Bob Ceculski wrote: 7 > these posts are endless, but not for a vms webserver!m > H > I just got a call from my DSL provider informing me that my machine isJ > being used to perform port scans on a bank. I thought that my system was> > pretty secure but apparently I was wrong. Here is the setup, >  > Redhat Linux 7.2 + Ximianl; > 2.4.18 kernel with Win4Lin patch (I compiled this myself)gG > All current patches installed (at least all of the patchs that Ximianu > supplies for Redhat 7.2)I > PMfirewall, all ports closed except for SSH, HTTPS and HTTP (I've sinceo > closed HTTP)@ > Machine is acting as a router for several other Linux systems. > @ > I've run the port scans on the Sheilds Up site and the ones on; > DSLReports and they all report that my machine is secure.a > 8 > A suspicious message that I found in /var/log/messages > E > Apr  5 18:07:31 dreadnought kernel: eth0: Setting promiscuous mode.dJ > Apr  5 18:07:31 dreadnought kernel: device eth0 entered promiscuous mode >  >  > Here are my questions, > C > 1) How do I figure out if there is a trojan running on my system?o > ; > 2) Is there anything I else I can do to secure my system?e > J > 3) Is there anyway to clean it up short of doing a complete reinstall of	 > the OS?p > I > 4) Recommendations for a hardware firewall? I need to be able to SSH toh& > my network and SFTP to my machines.  > = > 5) How do I disable promiscous mode on the NIC (3Com 905B)?  > 	 > Thanks,   > Hey Bob, care to explain what this all has to do with OpenVMS?  	 Bart Zornr   ------------------------------   Date: 9 Apr 2002 03:51:26 -0700d) From: P.Young@unsw.EDU.AU (Patrick Young)w0 Subject: Re: Just another linux nightmare day #2= Message-ID: <55f85d77.0204090251.1b3516c5@posting.google.com>i  ` Bart Zorn <B.Zorn@xs4all.nospam.nl> wrote in message news:<3CB28F51.5090501@xs4all.nospam.nl>... > @ > Hey Bob, care to explain what this all has to do with OpenVMS? >   A I'm sure Bob is well meaning in his efforts, even if they may notv  be in the "best possible taste".  H I am confused as to why he "preaches to the converted" (ie: comp.os.vms)  H I'm sure his postings would be better placed in the advocacy departmentsC of newsgroups relating to the various operating systems to which he0$ sometimes quite correctly refers to.   ------------------------------  % Date: Tue, 09 Apr 2002 12:52:17 +0100n( From: Nic Clews <sendspamhere@127.0.0.1>0 Subject: Re: Just another linux nightmare day #2) Message-ID: <3CB2D5F1.CBC89691@127.0.0.1>    Patrick Young wrote:  C > I'm sure Bob is well meaning in his efforts, even if they may noto" > be in the "best possible taste".  < ...recollections of Kenny Everett and his TV show, smiles...  aJ > I am confused as to why he "preaches to the converted" (ie: comp.os.vms)  & Something needs to balance the trolls.  yJ > I'm sure his postings would be better placed in the advocacy departmentsE > of newsgroups relating to the various operating systems to which hec& > sometimes quite correctly refers to.  E Bob, why not set up a web site for all to see? A man of your calibre,A. with EDT and RUNOFF skills, it'll be a breeze.   -- t( Regards, Nic Clews CSC Computer Sciences nclews at csc dot comi   ------------------------------   Date: 9 Apr 2002 05:56:15 -0700E( From: bob@instantwhip.com (Bob Ceculski)0 Subject: Re: Just another linux nightmare day #2= Message-ID: <d7791aa1.0204090456.74bd2b09@posting.google.com>i  n P.Young@unsw.EDU.AU (Patrick Young) wrote in message news:<55f85d77.0204090251.1b3516c5@posting.google.com>...b > Bart Zorn <B.Zorn@xs4all.nospam.nl> wrote in message news:<3CB28F51.5090501@xs4all.nospam.nl>... > > B > > Hey Bob, care to explain what this all has to do with OpenVMS? > >  > C > I'm sure Bob is well meaning in his efforts, even if they may noti" > be in the "best possible taste". > J > I am confused as to why he "preaches to the converted" (ie: comp.os.vms) > J > I'm sure his postings would be better placed in the advocacy departmentsE > of newsgroups relating to the various operating systems to which hef& > sometimes quite correctly refers to.  E and I do exactly that ... this is for Andrew and all those vms winerseF who keep mentioning linux as a place to port to from vms ... I however am not stupid ...s   ------------------------------   Date: 9 Apr 2002 05:45:31 -0700 ( From: bob@instantwhip.com (Bob Ceculski)0 Subject: Re: Just another linux nightmare day #2= Message-ID: <d7791aa1.0204090445.3fefdb76@posting.google.com>s  ` Bart Zorn <B.Zorn@xs4all.nospam.nl> wrote in message news:<3CB28F51.5090501@xs4all.nospam.nl>... > @ > Hey Bob, care to explain what this all has to do with OpenVMS? >  > Bart Zornc  < it shows Andrew and everyone else who thinks linux is a good" alternative to vms to think again!   ------------------------------   Date: 9 Apr 2002 15:34:56 GMTa1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)h1 Subject: Re: Just another linux nightmare day ...8, Message-ID: <a8v1n0$2htn$1@info.cs.uofs.edu>  0 That's what I like about this group.  The logic.  @ You take a toy OS, installed and administered by a some kid withA less than 0 days of experience and no idea how to spell security.oD And from this you draw the conclusion that any OS that has a similar* user interface is insecure no matter what.  ? By the way, using the logic applied here, the most secure OS is.A still Primos. I don't believe there has ever been a CERT advisory8E that mentioned it being susceptable. (Yes, Primos is still avaialable>A and in use.)  Come to think of it, I'll bet there aren't any CERTg' advisories for RT-11 or RSTS/E either!!d   bill   -- sJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Tue, 09 Apr 2002 16:07:32 +0200m= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> ! Subject: Re: Macro-64 and Vax VMS ) Message-ID: <3CB2F5A4.A7ED9340@gtech.com>i   Larry Kilgallen wrote:] > In article <3CAD0A91.CF37E1EF@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:iG > > Actually, Macro64 still requires a license. However, the license is  > > freely available.  > > > Macro64 requires a license _PAK_, which is freely available. > D > _Any_ software requires a _license_, unless you wrote it yourself.D > That includes Freeware.  A license might even be verbal, but it is > required.c  + Strictly speaking you are ofcourse correct.t  . But I think everyone understood what was said.   Arne   ------------------------------  % Date: Tue, 09 Apr 2002 16:46:47 +0200l= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>i Subject: Re: Memory Corruption) Message-ID: <3CB2FED7.3F6A4862@gtech.com>e   wing wrote:sC > I am writing a program, in which, if I compile it in with nodebugiF > version, its behaviour is not the same as in debug mode.  Thus, I am> > thinking of memory corruption. Is I am in a right direction?   Probably yes !  A > Is there any guideline/tool to find out if there is any memeorys > corruption in a program.  9 I would start by looking a bit at the source code all thei places where you use pointers !p   Arne   ------------------------------  # Date: Tue, 09 Apr 2002 16:56:10 GMT % From: jlsue <jlsuexxxz@insightbb.com>d Subject: Re: Memory Corruption8 Message-ID: <2076bug5pelochmht3ff2h1egjn4p0tih0@4ax.com>  / On Tue, 09 Apr 2002 16:46:47 +0200, Arne Vajhj  <arne.vajhoej@gtech.com> wrote:c   >wing wrote:D >> I am writing a program, in which, if I compile it in with nodebugG >> version, its behaviour is not the same as in debug mode.  Thus, I amv? >> thinking of memory corruption. Is I am in a right direction?u >s >Probably yes !N > B >> Is there any guideline/tool to find out if there is any memeory >> corruption in a program.o >o: >I would start by looking a bit at the source code all the  >places where you use pointers ! >f  C Here's another thing to try.  When compiling optimized, turn on allIC checks.  For Fortran, this was /check=all iirc.  At the very least,  put in /checks=bounds.  D What you will then get is an accvio if/when your code tries to stuff too much into a variable.   ? I had this problem once-upon-a-time when I declared a characteriD variable in Fortran to some size (say 30 characters), and then triedF to stuff in 31 (or more) characters of data.  At first it worked fine,B but one day it started doing strange things after a recompile.  ItC seems that the variables were mapped differently in memory, and the D last byte got stuffed into an integer variable that just happened toC get mapped directly after the character variable.  And the order ofyA declaration did not determine which variable mapped where (again,yE iirc).  If I had been luckier, the extra data would have been stuffeduB into nowrite pages (i.e., code) and I'd have gotten an ACCVIO, butD instead I spent days trying to figure it via debug (kept seeing this@ variable's value change when I was deep down into a an area that didn't use it).t  @ Compiling with the checks can help detect this situation easier.  1 Not speaking for anyone, certainly not DEC/Compaqo- (get rid of the xxxz in my address to e-mail)e   ------------------------------  $ Date: Tue, 9 Apr 2002 08:14:43 -0700, From: "James Gessling" <jgessling@yahoo.com>! Subject: Merge Two CMS libraries? 4 Message-ID: <a8v0gp$vk87u$1@ID-46415.news.dfncis.de>  A Anyone know how to do this?  It would be nice to keep the historyu intact if possible..   Any suggestions appreciated,   Jim7   ------------------------------  # Date: Tue, 09 Apr 2002 11:59:18 GMT># From: "John Smith" <a@nonymous.com>s' Subject: Re: More positive OpenVMS newshH Message-ID: <qIAs8.28473$cN1.26596@news01.bloor.is.net.cable.rogers.com>  H If Walter Hewlett gets his way, we won't have Carly to worry about, just Curly.    : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3CB23D57.F554E3D1@videotron.ca... > "John N." wrote: > > What I was happyI > > to see was a press release that mentioned a recognizable customer and|H > > actually mentioned OpenVMS rather proudly in a public press release. That# > > to me looks like real progress!. >r8 > Renaissance was allowed to dry up with no explanation.( > Now, HP is about to change everything. >sI > Perhaps the top VMS guys have realised that Carly has some bad news ford VMSeE > and they are taking the time before the vote is certified to try too convince# > her to change her mind about VMS.m >tL > Strikes me as odd that they would start now to admit that they have VMS as
 a product. >lG > Also, note that since the announcement of the murder of Alpha,  PressP Releases# > about Alpha sales have increased.n >tJ > Sorry, but until I hear/see Carly's 3 year product roadmap, I won't make/ > anything from stuff that comes out of Compaq.  > K > If VMS survives the initial round, then HP will have 3 years to show whate itK > wants with VMS. If during those 3 years, VMS remains as obscure as it wasrI > under Compaq, then it won't last much longer once the 3 years are over.x But ifH > during those 3 years, we see sustained marketing and exposure for VMS, then$ > perhaps we can start to celebrate. > H > I just wish Carly opened the damned enveloppe and revealed her product roadmap.   ------------------------------  $ Date: Tue, 9 Apr 2002 09:05:17 -05001 From: "Dave Gudewicz" <david.gudewicz@abbott.com>h' Subject: Re: More positive OpenVMS newsC8 Message-ID: <a8up77$f7b$1@fizban.fizban.pprd.abbott.com>  K Well said Larry, and I find it interesting that even when some good news isd/ delivered about VMS, some here take it to task.e  H A friend of mine was known to say on such occasions:  "some people would' bitch if you hung 'em with a new rope."l   Dave...c  : "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message- news:1mbt8XEFj1xw@eisner.encompasserve.org...rK > In article <a0ps8.130129$K52.20480662@typhoon.tampabay.rr.com>, "John N."r <JNixon@cfl.rr.com> writes::J > > As far as I was concerned, it wasn't the fact that it was only 3 es40s and a-I > > couple of DSxxs.  Heck, I bought more than that recently.  What I wask happynI > > to see was a press release that mentioned a recognizable customer and2H > > actually mentioned OpenVMS rather proudly in a public press release. That# > > to me looks like real progress!0 >1B > There were insufficient press releases to cover the early monthsD > of ES45 deliveries, and rumors they were going to people who would6 > never agree to press releases about their purchases.   ------------------------------   Date: 9 Apr 2002 06:03:53 -0700S( From: bob@instantwhip.com (Bob Ceculski)' Subject: Re: More positive OpenVMS newsn= Message-ID: <d7791aa1.0204090503.663f870f@posting.google.com>g  a JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message news:<3CB23D57.F554E3D1@videotron.ca>...  > "John N." wrote: > > What I was happyI > > to see was a press release that mentioned a recognizable customer and N > > actually mentioned OpenVMS rather proudly in a public press release.  That$ > > to me looks like real progress!  > 8 > Renaissance was allowed to dry up with no explanation.( > Now, HP is about to change everything. > M > Perhaps the top VMS guys have realised that Carly has some bad news for VMSnN > and they are taking the time before the vote is certified to try to convince# > her to change her mind about VMS.w > W > Strikes me as odd that they would start now to admit that they have VMS as a product.s > P > Also, note that since the announcement of the murder of Alpha,  Press Releases# > about Alpha sales have increased.r > J > Sorry, but until I hear/see Carly's 3 year product roadmap, I won't make/ > anything from stuff that comes out of Compaq.a > N > If VMS survives the initial round, then HP will have 3 years to show what itK > wants with VMS. If during those 3 years, VMS remains as obscure as it was P > under Compaq, then it won't last much longer once the 3 years are over. But ifM > during those 3 years, we see sustained marketing and exposure for VMS, theno$ > perhaps we can start to celebrate. > Q > I just wish Carly opened the damned enveloppe and revealed her product roadmap.i  H if Carly was going to kill vms, she would have done it the same time mpeF and tru64 was killed, but she didn't, and the itanium port is still on and EV7 is out ...  . Compaq has announced the EV7 alphaserver line:  1    http://www.compaq.com/hps/announce/apr_02.htmld    A MARVELous picture is included.   ------------------------------  % Date: Tue, 09 Apr 2002 10:16:51 -0400 1 From: Michael Austin <maustin@firstdbasource.com>d' Subject: Re: More positive OpenVMS newsn2 Message-ID: <3CB2F7D3.25F5E73E@firstdbasource.com>   Bob Ceculski wrote:e > t > "Bill Todd" <billtodd@metrocast.net> wrote in message news:<PFos8.62918$w7.5185433@bin6.nnrp.aus1.giganews.com>...: > > "Kenneth Farmer" <kfarmer@farmer.org> wrote in message= > > news:bFms8.26983$XP2.10431805@typhoon.southeast.rr.com...4 > > > I > > > Compaq Selected for Department of Veteran's Affairs PCHS-2 Contract ? > > > http://www.openvms.org/stories.php?story=02/04/08/4148271  > > >PP > > > Tyson Foods Chooses Compaq High Availability Solution for Mission Critical
 > > > Apps? > > > http://www.openvms.org/stories.php?story=02/04/08/1897447@ > >cM > > I mean it's nice and all, but somehow I doubt that the sale of 3 ES40s, a.M > > DS20, and a DS10 to Tyson (if they didn't already own some of those) willuO > > even fund Fred for a year.  Nor is it clear that any significant portion of O > > that VA contract involves VMS (though, again, it's nice to see it in use atl	 > > all).b > > 
 > > - bill > H > does that mean that because my company only uses Alphaserver 800's andI > workstations that that's not important ... large accounts are nice, buteK > lots and lots of small and medium accounts are just as important ... it'sEH > called volume ... and as those small and medium accounts grow, so doesL > your large customer base ... if you worked for dec, no wonder they failed, > you don't know how to market!J  H Now if Oracle could convince Tyson to dump Sybase and go back to Rdb.  IH once was onsite to assist in the 4.1-5.1 upgrade. 45+ databases (100+ GBG - 7 years ago) backed up, Rdb upgraded, databases converted and back inmD production  in a little under 3.5 hours.  The poor Adabase guy had aE 40Mb -- yes 80,000 blocks  took more than 14 hours to do his upgrade.   G To do the backups they had the 15 DEC tape drives made by IBM with tapeWF loader that held 10 tapes each adn the backups were peformed using the) Rdb ability to multithread the backups...    -- s Regards,  7 Michael Austin            Registered Linux User #261163S7 First DBA Source, Inc.    http://www.firstdbasource.comn Sr. Consultant 704-947-1089 (Office)y 704-236-4377 (Mobile)t   ------------------------------  % Date: Tue, 09 Apr 2002 12:59:08 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> ' Subject: Re: More positive OpenVMS newso, Message-ID: <3CB31DDA.C3AE68EE@videotron.ca>   Bill Todd wrote:N > The point is not how long *you* give *them* to prove something, the point isN > that regardless of whether they show a 3-year roadmap for VMS next week they9 > may well still kill it next fall, as they did with MPE.h    N If Carly really does intend to see VMS survive, grow and compete, I suspect weF will know fairly quickly because we'll start to see flashing balls andN advertising again. And that will be followed perhaps by some agressive pricing@ changes to make VMS compete more effectively against Windows NT.  : If they want to make VMS succeed, it won't be hard to see.   ------------------------------  % Date: Tue, 09 Apr 2002 13:32:41 -0400-- From: JF Mezei <jfmezei.spamnot@videotron.ca>2' Subject: Re: More positive OpenVMS newsi, Message-ID: <3CB325B4.F6CD4A58@videotron.ca>   John Smith wrote:g > J > If Walter Hewlett gets his way, we won't have Carly to worry about, just > Curly.  M If HP doesn't get Compaq, Curly has at most one year to find a job elsewhere.,N Usually, when such merger fail, there follows a change in management some timeU later (later enough that it doesn't look to be a direct result of the failed merger).    ------------------------------  % Date: Tue, 09 Apr 2002 14:00:46 -0400'- From: JF Mezei <jfmezei.spamnot@videotron.ca>h' Subject: Re: More positive OpenVMS newsa, Message-ID: <3CB32C48.7FC71569@videotron.ca>   Bob Ceculski wrote:oJ > if Carly was going to kill vms, she would have done it the same time mpeH > and tru64 was killed, but she didn't, and the itanium port is still on > and EV7 is out ...  M When Compaq seriously considered killing VMS 2 years ago, I suspect that they:I really looked at the metrics and how much "churn" they would get (loss of I customer to competitor). They knew then that killing VMS right away wouldi% disrupt Compaq's financials too much.e  K Palmer may have hoped to convert VMS customers to NT. It wasn't so much his E execution that was flawed, but his timing. NT was nowhere near ready.c  5 Remember the "VMS is now where NT 8 wants to be". ???t  L Is Windows XP , NT-5 or NT-6 ?  Either way, it is starting to get near NT-8.? Unix is gaining clustering, has far more business applications.oG Tandem is moving to that IA64 thing and offers the reliability some VMSa customers need.C  N Times have changed. Options which were suicidal in the palmer years may become more palatable over time.0   ------------------------------  # Date: Tue, 09 Apr 2002 08:03:04 GMTi' From: "Hans Vlems" <hvlems@zfree.co.nz>a# Subject: Re: Motif startup problem.c$ Message-ID: <3cb2a035$1@zfree.co.nz>  7 Probably not, Motif was started at boot time as well aso1 interactively when the system was up and running.s; The monitor (a Digital PCXCV-xx) turns its green active led.< into orange and stays that way. In a sense you may be right,5 it shows the behaviour of being in screen saver mode.a# But not owing to lack of activity.  A Are there any diagnostics/tools/utilities to check the videocard?   6 "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote:C >The log looks fine.  Are you sure you aren't in screen saver mode?m >s > 8 >Hans Vlems wrote in message <3cb1bba6$1@zfree.co.nz>... >>H >>Hardware: white box Alpha 3000, 256 MB memory and an S3 Trio/32/Trio64 >video
 >>adapter.3 >>Software: AXP/VMS 7.3, Motif and DECnet phase IV.  >>H >>As soon as Motif is started the the display turns black and that's it.K >>The windows server (SYS$MANAGER:DECW$SERVER_0_ERROR.LOG) looks like this:i >>4 >> 8-APR-2002 13:45:26.4 Hello, this is the X serverG >>This is the DECwindows X11 display server for OpenVMS AXP V7.1-010317m& >> compiled on Mar 17 2001 at 09:10:09 >>Main address = 000386A0B, >>Activating extension image DECW$SVEXT_Xie,- >>extension name: Xie, entry address 002E6780t2 >>Activating extension image DECW$SVEXT_DEC_XTRAP,3 >>extension name: DEC-XTRAP, entry address 003B8548 8 >>Activating extension image DECW$SVEXT_Multi_Buffering,9 >>extension name: Multi-Buffering, entry address 003FA0A0f4 >>DECW$TRANSPORT_COMMON image base address: 7C7300004 >>DECW$TRANSPORT_DECNET image base address: 0043E000< >>%DECW-I-ATTACHED, transport DECNET attached to its network3 >>DECW$TRANSPORT_LOCAL image base address: 0059E000 3 >>DECWINDOWS DigitalEquipmentCorp. AXP, Release 7.1d7 >>Shareable Image DDX GQ, InitOutput loaded at 005E1160t >>* >>s3vmsScreenInit: Init S3 screen number 0& >>vgaMapScreen: Call vmsInitDeviceData# >>vgaMapScreen: Opening device GQA0e# >>vgaMapScreen: Controller Number 0  >>vgaMapScreen: DVI info -  >>    Width = 1024, Height = 768. >>    Video PFN = 41C0000, Video Pagelets 2048' >>    CSR PFN 42C0000, CSR pagelets 128e >>    Planes 8I >>  Swizzle Shift 5, Swizzle Length 3, Swizzle Short Mask 8, Long Mask 18 & >>VGAVMS_MapScreen: CSRs are at 800000M >>VGAVMS_MapScreen: VRAM C00000:2C00000, PFN ^x41C0000, pages 65536, onscreen  >>C00000! >> cputype = 5 (2 = EV4, 5 = EV5)v1 >>s3vmsScreenInit: - Set the Default Visual Class ( >>s3vmsScreenInit: - Call cfbInitVisuals- >>s3vmsScreenInit: - mask 0x80 bits-per-rgb 8 ) >>s3vmsScreenInit: - Visuals Information:8 >>  - Root depth        = 8o >>  - Number of depths  = 2M >>  - Number of visuals = 6- >>  - Default Visual    = 20; >>s3vmsScreenInit: Actual memory size reported is: 0x100000E: >>s3vmsScreenInit: Set screen geometry to match user setup: >>s3vmsScreenInit: Display Mode Index 7, Refresh Rate 72Hz; >>s3vmsScreenInit: Screen X 1024, Screen Y 768, density 100u& >>s3vmsScreenInit: - Call miScreenInit2 >>s3vmsScreenInit: Create screen private structureG >>s3vmsScreenInit: Mapping Enhanced Registers (font buffer VGA address)tB >>s3vmsScreenInit: 2097152 bytes of enhanced register space mapped& >>                 mapped at 0x2C00000& >>                 PFN is at 0x4280A00; >>s3vmsScreenInit: No DENSE space, using SPARSE memory only:; >>s3vmsScreenInit: Chip type S3 Trio64, Rev: 0x54, PNP == 1o) >>s3vmsScreenInit: M Parameter value is 2i+ >>s3vmsScreenInit: N Parameter value is 255e( >>s3vmsScreenInit: FIFO checking ENABLED= >>s3vmsScreenInit: Memory Layout (from start of FB in bytes):h* >> - memory total               0x001000000 >> - number of offscreen lines  0x000000BF (191)* >> - offscreen starts at        0x000C0000* >> - offscreen ends at          0x000EFC00* >> - fonts cache starts at      0x000EFC00* >> - fonts cache ends at        0x000FFC00* >> - cursor bits are at         0x000FFC00/ >>s3vmsScreenInit: - Init the common VGA screen-4 >>VGAVMS_InitScreenCommon: Init the Screen structure7 >>s3vmsScreenInit: Installing S3-specific optimizationsr% >>s3vmsScreenInit: Using EV5 routinesf  >>s3vmsScreenInit: Init hardware
 >>CR55 = 0
 >>CR56 = 0 >>SR18 = 40j >>DAC Mask = FFe? >>s3HWInit: linear addr window = 1 (0=64kb, 1=1mb, 2=2mb, 3=4mbi >>s3HWInit: Set MMIO mode 3 >>Initial M Parameter = 6, Setting M Parameter to 2m6 >>Initial N Parameter = 15, Setting N Parameter to 255" >>End of init, turn screen back onB >>s3vmsScreenInit: Cursor Setup: shape and color is done by server2 >>s3vmsScreenInit: Cursor is being moved by server, >>s3vmsScreenInit: Call cfbCreateDefColormap+ >>s3vmsScreenInit: - Init the backing storeo9 >>VGAVMS_SetupConsole: Console remove value (seconds) 300c, >>VGAVMS_SetupConsole: Issue AUTO_OPDOWN QIO; >>VGAVMS_SetupConsole: Setup the console suspend routine... . >>s3vmsScreenInit: Exit from init with successK >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR32] removed from font path, fontsa >>unavailable.K >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR16] removed from font path, fontse >>unavailable.I >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_100DPI] removed from font path, fontst
 >unavailable.,H >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_75DPI] removed from font path, fonts
 >unavailable.aI >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_COMMON] removed from font path, fontso
 >unavailable.iI >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_SPEEDO] removed from font path, fontse
 >unavailable.rH >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_TYPE1] removed from font path, fonts
 >unavailable.d2 >> 8-APR-2002 13:45:36.4 Calling the dispatcher...J >> 8-APR-2002 13:45:36.8 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.75DPI] removed$ >>from font path, fonts unavailable.K >> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.100DPI] removed $ >>from font path, fonts unavailable.I >> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C] removed fromi >fonta >>path, fonts unavailable. >> >>P.S.  >>I have no control over the url >> >> >> >> >>http://www.zfree.co.nz >> >  >,       http://www.zfree.co.nz   ------------------------------  # Date: Tue, 09 Apr 2002 08:05:11 GMTi' From: "Hans Vlems" <hvlems@zfree.co.nz>5# Subject: Re: Motif startup problem.o$ Message-ID: <3cb2a0b3$1@zfree.co.nz>  > Good question; the answer is it is an old Digital PCXCV-BW and! I do not know its characteristis.-" Where can I change the attributes?   Hans  , Jeff Campbell <jcampbell@ins-msi.com> wrote:> >Log looks normal. Does your monitor support 1024 by 768 mode? >6 >> vgaMapScreen: DVI info -g! >>     Width = 1024, Height = 768h >w >Jeff Campbell >n8wxs@arrl.neta >t >Hans Vlems wrote: >> eI >> Hardware: white box Alpha 3000, 256 MB memory and an S3 Trio/32/Trio641 videor >> adapter.J4 >> Software: AXP/VMS 7.3, Motif and DECnet phase IV. >> sI >> As soon as Motif is started the the display turns black and that's it.dL >> The windows server (SYS$MANAGER:DECW$SERVER_0_ERROR.LOG) looks like this: >>  5 >>  8-APR-2002 13:45:26.4 Hello, this is the X serverhH >> This is the DECwindows X11 display server for OpenVMS AXP V7.1-0103176 >>                 compiled on Mar 17 2001 at 09:10:09 >> Main address = 000386A0- >> Activating extension image DECW$SVEXT_Xie, . >> extension name: Xie, entry address 002E67803 >> Activating extension image DECW$SVEXT_DEC_XTRAP,t4 >> extension name: DEC-XTRAP, entry address 003B85489 >> Activating extension image DECW$SVEXT_Multi_Buffering, : >> extension name: Multi-Buffering, entry address 003FA0A05 >> DECW$TRANSPORT_COMMON image base address: 7C730000o5 >> DECW$TRANSPORT_DECNET image base address: 0043E000c= >> %DECW-I-ATTACHED, transport DECNET attached to its networkt4 >> DECW$TRANSPORT_LOCAL image base address: 0059E0004 >> DECWINDOWS DigitalEquipmentCorp. AXP, Release 7.18 >> Shareable Image DDX GQ, InitOutput loaded at 005E1160 >> c+ >> s3vmsScreenInit: Init S3 screen number 0o' >> vgaMapScreen: Call vmsInitDeviceDatad$ >> vgaMapScreen: Opening device GQA0$ >> vgaMapScreen: Controller Number 0 >> vgaMapScreen: DVI info - ! >>     Width = 1024, Height = 768 / >>     Video PFN = 41C0000, Video Pagelets 2048e( >>     CSR PFN 42C0000, CSR pagelets 128 >>     Planes 8vJ >>   Swizzle Shift 5, Swizzle Length 3, Swizzle Short Mask 8, Long Mask 18' >> VGAVMS_MapScreen: CSRs are at 800000tN >> VGAVMS_MapScreen: VRAM C00000:2C00000, PFN ^x41C0000, pages 65536, onscreen	 >> C00000i" >>  cputype = 5 (2 = EV4, 5 = EV5)2 >> s3vmsScreenInit: - Set the Default Visual Class) >> s3vmsScreenInit: - Call cfbInitVisualsy. >> s3vmsScreenInit: - mask 0x80 bits-per-rgb 8* >> s3vmsScreenInit: - Visuals Information: >>   - Root depth        = 8 >>   - Number of depths  = 2 >>   - Number of visuals = 6 >>   - Default Visual    = 20.< >> s3vmsScreenInit: Actual memory size reported is: 0x100000; >> s3vmsScreenInit: Set screen geometry to match user setup ; >> s3vmsScreenInit: Display Mode Index 7, Refresh Rate 72Hze< >> s3vmsScreenInit: Screen X 1024, Screen Y 768, density 100' >> s3vmsScreenInit: - Call miScreenInit23 >> s3vmsScreenInit: Create screen private structurepH >> s3vmsScreenInit: Mapping Enhanced Registers (font buffer VGA address)C >> s3vmsScreenInit: 2097152 bytes of enhanced register space mappedE' >>                  mapped at 0x2C00000g' >>                  PFN is at 0x4280A00e< >> s3vmsScreenInit: No DENSE space, using SPARSE memory only< >> s3vmsScreenInit: Chip type S3 Trio64, Rev: 0x54, PNP == 1* >> s3vmsScreenInit: M Parameter value is 2, >> s3vmsScreenInit: N Parameter value is 255) >> s3vmsScreenInit: FIFO checking ENABLEDF> >> s3vmsScreenInit: Memory Layout (from start of FB in bytes):+ >>  - memory total               0x00100000 1 >>  - number of offscreen lines  0x000000BF (191)s+ >>  - offscreen starts at        0x000C0000e+ >>  - offscreen ends at          0x000EFC00 + >>  - fonts cache starts at      0x000EFC00 + >>  - fonts cache ends at        0x000FFC00s+ >>  - cursor bits are at         0x000FFC00o0 >> s3vmsScreenInit: - Init the common VGA screen5 >> VGAVMS_InitScreenCommon: Init the Screen structure 8 >> s3vmsScreenInit: Installing S3-specific optimizations& >> s3vmsScreenInit: Using EV5 routines! >> s3vmsScreenInit: Init hardwaret >> CR55 = 0  >> CR56 = 0  >> SR18 = 40 >> DAC Mask = FF@ >> s3HWInit: linear addr window = 1 (0=64kb, 1=1mb, 2=2mb, 3=4mb >> s3HWInit: Set MMIO mode4 >> Initial M Parameter = 6, Setting M Parameter to 27 >> Initial N Parameter = 15, Setting N Parameter to 255'# >> End of init, turn screen back oneC >> s3vmsScreenInit: Cursor Setup: shape and color is done by serveri3 >> s3vmsScreenInit: Cursor is being moved by serverh- >> s3vmsScreenInit: Call cfbCreateDefColormap , >> s3vmsScreenInit: - Init the backing store: >> VGAVMS_SetupConsole: Console remove value (seconds) 300- >> VGAVMS_SetupConsole: Issue AUTO_OPDOWN QIOe< >> VGAVMS_SetupConsole: Setup the console suspend routine.../ >> s3vmsScreenInit: Exit from init with success L >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR32] removed from font path, fonts >> unavailable.sL >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR16] removed from font path, fonts >> unavailable.uJ >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_100DPI] removed from font path, fonts unavailable.I >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_75DPI] removed from font path, fonts  unavailable.J >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_COMMON] removed from font path, fonts unavailable.J >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_SPEEDO] removed from font path, fonts unavailable.I >> DECW$SYSCOMMON:[SYSFONT.DECW.USER_TYPE1] removed from font path, fontsl unavailable.3 >>  8-APR-2002 13:45:36.4 Calling the dispatcher...$K >>  8-APR-2002 13:45:36.8 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.75DPI] removedm% >> from font path, fonts unavailable.8L >>  8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.100DPI] removed% >> from font path, fonts unavailable.CJ >>  8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C] removed from font >> path, fonts unavailable.o >> o >> P.S. ! >> I have no control over the urlr >> l >> http://www.zfree.co.nzw       http://www.zfree.co.nz   ------------------------------  # Date: Tue, 09 Apr 2002 09:47:44 GMTo' From: "Hans Vlems" <hvlems@zfree.co.nz>0# Subject: Re: Motif startup problem. $ Message-ID: <3cb2b8bd$1@zfree.co.nz>   Fred,   $ I modified DECW$DEVICE_CONFIG_GQ.COM  6 "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote:C >The log looks fine.  Are you sure you aren't in screen saver mode?  >  >i8 >Hans Vlems wrote in message <3cb1bba6$1@zfree.co.nz>... >>H >>Hardware: white box Alpha 3000, 256 MB memory and an S3 Trio/32/Trio64 >video
 >>adapter.3 >>Software: AXP/VMS 7.3, Motif and DECnet phase IV.s >>H >>As soon as Motif is started the the display turns black and that's it.K >>The windows server (SYS$MANAGER:DECW$SERVER_0_ERROR.LOG) looks like this:o >>4 >> 8-APR-2002 13:45:26.4 Hello, this is the X serverG >>This is the DECwindows X11 display server for OpenVMS AXP V7.1-010317 & >> compiled on Mar 17 2001 at 09:10:09 >>Main address = 000386A00, >>Activating extension image DECW$SVEXT_Xie,- >>extension name: Xie, entry address 002E6780n2 >>Activating extension image DECW$SVEXT_DEC_XTRAP,3 >>extension name: DEC-XTRAP, entry address 003B8548s8 >>Activating extension image DECW$SVEXT_Multi_Buffering,9 >>extension name: Multi-Buffering, entry address 003FA0A0 4 >>DECW$TRANSPORT_COMMON image base address: 7C7300004 >>DECW$TRANSPORT_DECNET image base address: 0043E000< >>%DECW-I-ATTACHED, transport DECNET attached to its network3 >>DECW$TRANSPORT_LOCAL image base address: 0059E000c3 >>DECWINDOWS DigitalEquipmentCorp. AXP, Release 7.137 >>Shareable Image DDX GQ, InitOutput loaded at 005E1160e >>* >>s3vmsScreenInit: Init S3 screen number 0& >>vgaMapScreen: Call vmsInitDeviceData# >>vgaMapScreen: Opening device GQA0 # >>vgaMapScreen: Controller Number 0r >>vgaMapScreen: DVI info -  >>    Width = 1024, Height = 768. >>    Video PFN = 41C0000, Video Pagelets 2048' >>    CSR PFN 42C0000, CSR pagelets 128y >>    Planes 8I >>  Swizzle Shift 5, Swizzle Length 3, Swizzle Short Mask 8, Long Mask 18E& >>VGAVMS_MapScreen: CSRs are at 800000M >>VGAVMS_MapScreen: VRAM C00000:2C00000, PFN ^x41C0000, pages 65536, onscreene >>C00000! >> cputype = 5 (2 = EV4, 5 = EV5)t1 >>s3vmsScreenInit: - Set the Default Visual Classo( >>s3vmsScreenInit: - Call cfbInitVisuals- >>s3vmsScreenInit: - mask 0x80 bits-per-rgb 8 ) >>s3vmsScreenInit: - Visuals Information:b >>  - Root depth        = 8- >>  - Number of depths  = 2u >>  - Number of visuals = 6- >>  - Default Visual    = 20; >>s3vmsScreenInit: Actual memory size reported is: 0x100000 : >>s3vmsScreenInit: Set screen geometry to match user setup: >>s3vmsScreenInit: Display Mode Index 7, Refresh Rate 72Hz; >>s3vmsScreenInit: Screen X 1024, Screen Y 768, density 100p& >>s3vmsScreenInit: - Call miScreenInit2 >>s3vmsScreenInit: Create screen private structureG >>s3vmsScreenInit: Mapping Enhanced Registers (font buffer VGA address) B >>s3vmsScreenInit: 2097152 bytes of enhanced register space mapped& >>                 mapped at 0x2C00000& >>                 PFN is at 0x4280A00; >>s3vmsScreenInit: No DENSE space, using SPARSE memory onlye; >>s3vmsScreenInit: Chip type S3 Trio64, Rev: 0x54, PNP == 1 ) >>s3vmsScreenInit: M Parameter value is 2g+ >>s3vmsScreenInit: N Parameter value is 255 ( >>s3vmsScreenInit: FIFO checking ENABLED= >>s3vmsScreenInit: Memory Layout (from start of FB in bytes): * >> - memory total               0x001000000 >> - number of offscreen lines  0x000000BF (191)* >> - offscreen starts at        0x000C0000* >> - offscreen ends at          0x000EFC00* >> - fonts cache starts at      0x000EFC00* >> - fonts cache ends at        0x000FFC00* >> - cursor bits are at         0x000FFC00/ >>s3vmsScreenInit: - Init the common VGA screen 4 >>VGAVMS_InitScreenCommon: Init the Screen structure7 >>s3vmsScreenInit: Installing S3-specific optimizations"% >>s3vmsScreenInit: Using EV5 routines?  >>s3vmsScreenInit: Init hardware
 >>CR55 = 0
 >>CR56 = 0 >>SR18 = 40i >>DAC Mask = FFb? >>s3HWInit: linear addr window = 1 (0=64kb, 1=1mb, 2=2mb, 3=4mbf >>s3HWInit: Set MMIO modeM3 >>Initial M Parameter = 6, Setting M Parameter to 2h6 >>Initial N Parameter = 15, Setting N Parameter to 255" >>End of init, turn screen back onB >>s3vmsScreenInit: Cursor Setup: shape and color is done by server2 >>s3vmsScreenInit: Cursor is being moved by server, >>s3vmsScreenInit: Call cfbCreateDefColormap+ >>s3vmsScreenInit: - Init the backing store 9 >>VGAVMS_SetupConsole: Console remove value (seconds) 300 , >>VGAVMS_SetupConsole: Issue AUTO_OPDOWN QIO; >>VGAVMS_SetupConsole: Setup the console suspend routine...w. >>s3vmsScreenInit: Exit from init with successK >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR32] removed from font path, fonts  >>unavailable.K >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR16] removed from font path, fontse >>unavailable.I >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_100DPI] removed from font path, fontsa
 >unavailable. H >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_75DPI] removed from font path, fonts
 >unavailable.lI >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_COMMON] removed from font path, fontst
 >unavailable..I >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_SPEEDO] removed from font path, fontsi
 >unavailable.cH >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_TYPE1] removed from font path, fonts
 >unavailable.h2 >> 8-APR-2002 13:45:36.4 Calling the dispatcher...J >> 8-APR-2002 13:45:36.8 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.75DPI] removed$ >>from font path, fonts unavailable.K >> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.100DPI] removedX$ >>from font path, fonts unavailable.I >> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C] removed frome >fonta >>path, fonts unavailable. >> >>P.S.  >>I have no control over the url >> >> >> >> >>http://www.zfree.co.nz >> >u >r       http://www.zfree.co.nz   ------------------------------  # Date: Tue, 09 Apr 2002 09:52:06 GMTC' From: "Hans Vlems" <hvlems@zfree.co.nz>-# Subject: Re: Motif startup problem. $ Message-ID: <3cb2b9c3$1@zfree.co.nz>   Fred,s  N I modified DECW$DEVICE_CONFIG_GQ.COM, changed the screen resolution to 800x6004 and that worked. The changes are now put in the fileH DECW$PRIVATE_SERVER_SETUP.COM and the original *GQ.COM file was put back	 in place. F The monitor does support 1024x800 at 75 Hz, so the question is what is5 the name of the variable that controls the frequency?gH What kind of things can be put into the private server setup file, is it documented?    Hans  6 "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote:C >The log looks fine.  Are you sure you aren't in screen saver mode?  >0 > 8 >Hans Vlems wrote in message <3cb1bba6$1@zfree.co.nz>... >>H >>Hardware: white box Alpha 3000, 256 MB memory and an S3 Trio/32/Trio64 >video
 >>adapter.3 >>Software: AXP/VMS 7.3, Motif and DECnet phase IV.i >>H >>As soon as Motif is started the the display turns black and that's it.K >>The windows server (SYS$MANAGER:DECW$SERVER_0_ERROR.LOG) looks like this:  >>4 >> 8-APR-2002 13:45:26.4 Hello, this is the X serverG >>This is the DECwindows X11 display server for OpenVMS AXP V7.1-0103173& >> compiled on Mar 17 2001 at 09:10:09 >>Main address = 000386A0e, >>Activating extension image DECW$SVEXT_Xie,- >>extension name: Xie, entry address 002E678082 >>Activating extension image DECW$SVEXT_DEC_XTRAP,3 >>extension name: DEC-XTRAP, entry address 003B8548e8 >>Activating extension image DECW$SVEXT_Multi_Buffering,9 >>extension name: Multi-Buffering, entry address 003FA0A0c4 >>DECW$TRANSPORT_COMMON image base address: 7C7300004 >>DECW$TRANSPORT_DECNET image base address: 0043E000< >>%DECW-I-ATTACHED, transport DECNET attached to its network3 >>DECW$TRANSPORT_LOCAL image base address: 0059E000=3 >>DECWINDOWS DigitalEquipmentCorp. AXP, Release 7.137 >>Shareable Image DDX GQ, InitOutput loaded at 005E1160i >>* >>s3vmsScreenInit: Init S3 screen number 0& >>vgaMapScreen: Call vmsInitDeviceData# >>vgaMapScreen: Opening device GQA00# >>vgaMapScreen: Controller Number 0s >>vgaMapScreen: DVI info -  >>    Width = 1024, Height = 768. >>    Video PFN = 41C0000, Video Pagelets 2048' >>    CSR PFN 42C0000, CSR pagelets 128  >>    Planes 8I >>  Swizzle Shift 5, Swizzle Length 3, Swizzle Short Mask 8, Long Mask 183& >>VGAVMS_MapScreen: CSRs are at 800000M >>VGAVMS_MapScreen: VRAM C00000:2C00000, PFN ^x41C0000, pages 65536, onscreenI >>C00000! >> cputype = 5 (2 = EV4, 5 = EV5)S1 >>s3vmsScreenInit: - Set the Default Visual ClassI( >>s3vmsScreenInit: - Call cfbInitVisuals- >>s3vmsScreenInit: - mask 0x80 bits-per-rgb 8a) >>s3vmsScreenInit: - Visuals Information:i >>  - Root depth        = 8  >>  - Number of depths  = 2t >>  - Number of visuals = 6t >>  - Default Visual    = 20; >>s3vmsScreenInit: Actual memory size reported is: 0x100000m: >>s3vmsScreenInit: Set screen geometry to match user setup: >>s3vmsScreenInit: Display Mode Index 7, Refresh Rate 72Hz; >>s3vmsScreenInit: Screen X 1024, Screen Y 768, density 100I& >>s3vmsScreenInit: - Call miScreenInit2 >>s3vmsScreenInit: Create screen private structureG >>s3vmsScreenInit: Mapping Enhanced Registers (font buffer VGA address)nB >>s3vmsScreenInit: 2097152 bytes of enhanced register space mapped& >>                 mapped at 0x2C00000& >>                 PFN is at 0x4280A00; >>s3vmsScreenInit: No DENSE space, using SPARSE memory onlyR; >>s3vmsScreenInit: Chip type S3 Trio64, Rev: 0x54, PNP == 1S) >>s3vmsScreenInit: M Parameter value is 2r+ >>s3vmsScreenInit: N Parameter value is 255S( >>s3vmsScreenInit: FIFO checking ENABLED= >>s3vmsScreenInit: Memory Layout (from start of FB in bytes):N* >> - memory total               0x001000000 >> - number of offscreen lines  0x000000BF (191)* >> - offscreen starts at        0x000C0000* >> - offscreen ends at          0x000EFC00* >> - fonts cache starts at      0x000EFC00* >> - fonts cache ends at        0x000FFC00* >> - cursor bits are at         0x000FFC00/ >>s3vmsScreenInit: - Init the common VGA screen54 >>VGAVMS_InitScreenCommon: Init the Screen structure7 >>s3vmsScreenInit: Installing S3-specific optimizations.% >>s3vmsScreenInit: Using EV5 routinest  >>s3vmsScreenInit: Init hardware
 >>CR55 = 0
 >>CR56 = 0 >>SR18 = 40T >>DAC Mask = FFi? >>s3HWInit: linear addr window = 1 (0=64kb, 1=1mb, 2=2mb, 3=4mbt >>s3HWInit: Set MMIO mode>3 >>Initial M Parameter = 6, Setting M Parameter to 2p6 >>Initial N Parameter = 15, Setting N Parameter to 255" >>End of init, turn screen back onB >>s3vmsScreenInit: Cursor Setup: shape and color is done by server2 >>s3vmsScreenInit: Cursor is being moved by server, >>s3vmsScreenInit: Call cfbCreateDefColormap+ >>s3vmsScreenInit: - Init the backing storee9 >>VGAVMS_SetupConsole: Console remove value (seconds) 300f, >>VGAVMS_SetupConsole: Issue AUTO_OPDOWN QIO; >>VGAVMS_SetupConsole: Setup the console suspend routine...6. >>s3vmsScreenInit: Exit from init with successK >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR32] removed from font path, fontst >>unavailable.K >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR16] removed from font path, fonts  >>unavailable.I >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_100DPI] removed from font path, fontsi
 >unavailable.eH >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_75DPI] removed from font path, fonts
 >unavailable.0I >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_COMMON] removed from font path, fontss
 >unavailable. I >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_SPEEDO] removed from font path, fonts>
 >unavailable. H >>DECW$SYSCOMMON:[SYSFONT.DECW.USER_TYPE1] removed from font path, fonts
 >unavailable. 2 >> 8-APR-2002 13:45:36.4 Calling the dispatcher...J >> 8-APR-2002 13:45:36.8 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.75DPI] removed$ >>from font path, fonts unavailable.K >> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.100DPI] removede$ >>from font path, fonts unavailable.I >> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C] removed from> >fontR >>path, fonts unavailable. >> >>P.S.  >>I have no control over the url >> >> >> >> >>http://www.zfree.co.nz >> >P >_       http://www.zfree.co.nz   ------------------------------  $ Date: Tue, 9 Apr 2002 13:11:06 -04005 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>v# Subject: Re: Motif startup problem.o3 Message-ID: <pmFs8.2043$fL6.55352@news.cpqcorp.net>S  ( $ define/sys decw$server_refresh_rate 66    ! The choices are 60, 66, 70 and 72D  I The screen geometry choices are 640x480, 800x600, 1024x768 (you only havei 1mb of memory).e   _FredC  7 Hans Vlems wrote in message <3cb2b9c3$1@zfree.co.nz>...i >e >Fred, >wG >I modified DECW$DEVICE_CONFIG_GQ.COM, changed the screen resolution toC 800x600t5 >and that worked. The changes are now put in the file I >DECW$PRIVATE_SERVER_SETUP.COM and the original *GQ.COM file was put back 
 >in place.G >The monitor does support 1024x800 at 75 Hz, so the question is what isc6 >the name of the variable that controls the frequency?I >What kind of things can be put into the private server setup file, is it  >documented? >e >Hansp > 7 >"Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote: D >>The log looks fine.  Are you sure you aren't in screen saver mode? >> >>9 >>Hans Vlems wrote in message <3cb1bba6$1@zfree.co.nz>...s >>>cI >>>Hardware: white box Alpha 3000, 256 MB memory and an S3 Trio/32/Trio64e >>video, >>>adapter.,4 >>>Software: AXP/VMS 7.3, Motif and DECnet phase IV. >>>sI >>>As soon as Motif is started the the display turns black and that's it.nL >>>The windows server (SYS$MANAGER:DECW$SERVER_0_ERROR.LOG) looks like this: >>>f5 >>> 8-APR-2002 13:45:26.4 Hello, this is the X serveraH >>>This is the DECwindows X11 display server for OpenVMS AXP V7.1-010317' >>> compiled on Mar 17 2001 at 09:10:09r >>>Main address = 000386A0- >>>Activating extension image DECW$SVEXT_Xie,v. >>>extension name: Xie, entry address 002E67803 >>>Activating extension image DECW$SVEXT_DEC_XTRAP,t4 >>>extension name: DEC-XTRAP, entry address 003B85489 >>>Activating extension image DECW$SVEXT_Multi_Buffering, : >>>extension name: Multi-Buffering, entry address 003FA0A05 >>>DECW$TRANSPORT_COMMON image base address: 7C73000005 >>>DECW$TRANSPORT_DECNET image base address: 0043E000 = >>>%DECW-I-ATTACHED, transport DECNET attached to its networka4 >>>DECW$TRANSPORT_LOCAL image base address: 0059E0004 >>>DECWINDOWS DigitalEquipmentCorp. AXP, Release 7.18 >>>Shareable Image DDX GQ, InitOutput loaded at 005E1160 >>> + >>>s3vmsScreenInit: Init S3 screen number 0o' >>>vgaMapScreen: Call vmsInitDeviceDataV$ >>>vgaMapScreen: Opening device GQA0$ >>>vgaMapScreen: Controller Number 0 >>>vgaMapScreen: DVI info -F! >>>    Width = 1024, Height = 768 / >>>    Video PFN = 41C0000, Video Pagelets 2048M( >>>    CSR PFN 42C0000, CSR pagelets 128 >>>    Planes 8oJ >>>  Swizzle Shift 5, Swizzle Length 3, Swizzle Short Mask 8, Long Mask 18' >>>VGAVMS_MapScreen: CSRs are at 800000tE >>>VGAVMS_MapScreen: VRAM C00000:2C00000, PFN ^x41C0000, pages 65536,  onscreen	 >>>C00000 " >>> cputype = 5 (2 = EV4, 5 = EV5)2 >>>s3vmsScreenInit: - Set the Default Visual Class) >>>s3vmsScreenInit: - Call cfbInitVisualso. >>>s3vmsScreenInit: - mask 0x80 bits-per-rgb 8* >>>s3vmsScreenInit: - Visuals Information: >>>  - Root depth        = 8 >>>  - Number of depths  = 2 >>>  - Number of visuals = 6 >>>  - Default Visual    = 20S< >>>s3vmsScreenInit: Actual memory size reported is: 0x100000; >>>s3vmsScreenInit: Set screen geometry to match user setupr; >>>s3vmsScreenInit: Display Mode Index 7, Refresh Rate 72HzM< >>>s3vmsScreenInit: Screen X 1024, Screen Y 768, density 100' >>>s3vmsScreenInit: - Call miScreenInitD3 >>>s3vmsScreenInit: Create screen private structureiH >>>s3vmsScreenInit: Mapping Enhanced Registers (font buffer VGA address)C >>>s3vmsScreenInit: 2097152 bytes of enhanced register space mappedr' >>>                 mapped at 0x2C00000.' >>>                 PFN is at 0x4280A00P< >>>s3vmsScreenInit: No DENSE space, using SPARSE memory only< >>>s3vmsScreenInit: Chip type S3 Trio64, Rev: 0x54, PNP == 1* >>>s3vmsScreenInit: M Parameter value is 2, >>>s3vmsScreenInit: N Parameter value is 255) >>>s3vmsScreenInit: FIFO checking ENABLEDD> >>>s3vmsScreenInit: Memory Layout (from start of FB in bytes):+ >>> - memory total               0x00100000D1 >>> - number of offscreen lines  0x000000BF (191)t+ >>> - offscreen starts at        0x000C0000 + >>> - offscreen ends at          0x000EFC00>+ >>> - fonts cache starts at      0x000EFC00r+ >>> - fonts cache ends at        0x000FFC00 + >>> - cursor bits are at         0x000FFC00H0 >>>s3vmsScreenInit: - Init the common VGA screen5 >>>VGAVMS_InitScreenCommon: Init the Screen structuren8 >>>s3vmsScreenInit: Installing S3-specific optimizations& >>>s3vmsScreenInit: Using EV5 routines! >>>s3vmsScreenInit: Init hardware  >>>CR55 = 0y >>>CR56 = 0s >>>SR18 = 40 >>>DAC Mask = FF@ >>>s3HWInit: linear addr window = 1 (0=64kb, 1=1mb, 2=2mb, 3=4mb >>>s3HWInit: Set MMIO mode4 >>>Initial M Parameter = 6, Setting M Parameter to 27 >>>Initial N Parameter = 15, Setting N Parameter to 255 # >>>End of init, turn screen back on C >>>s3vmsScreenInit: Cursor Setup: shape and color is done by serverA3 >>>s3vmsScreenInit: Cursor is being moved by server>- >>>s3vmsScreenInit: Call cfbCreateDefColormapv, >>>s3vmsScreenInit: - Init the backing store: >>>VGAVMS_SetupConsole: Console remove value (seconds) 300- >>>VGAVMS_SetupConsole: Issue AUTO_OPDOWN QIOa< >>>VGAVMS_SetupConsole: Setup the console suspend routine.../ >>>s3vmsScreenInit: Exit from init with successaL >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR32] removed from font path, fonts >>>unavailable.tL >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_CURSOR16] removed from font path, fonts >>>unavailable.AJ >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_100DPI] removed from font path, fonts >>unavailable.I >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_75DPI] removed from font path, fontsn >>unavailable.J >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_COMMON] removed from font path, fonts >>unavailable.J >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_SPEEDO] removed from font path, fonts >>unavailable.I >>>DECW$SYSCOMMON:[SYSFONT.DECW.USER_TYPE1] removed from font path, fontsn >>unavailable.3 >>> 8-APR-2002 13:45:36.4 Calling the dispatcher...pK >>> 8-APR-2002 13:45:36.8 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.75DPI] removed0% >>>from font path, fonts unavailable.2L >>> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C.100DPI] removed% >>>from font path, fonts unavailable. J >>> 8-APR-2002 13:45:36.9 CDE$USER_DEFAULTS:[CONFIG.XFONTS.C] removed from >>font >>>path, fonts unavailable.r >>>  >>>P.S. ! >>>I have no control over the url  >>>m >>>e >>>: >>>t >>>http://www.zfree.co.nz  >>>m >> >> >: >C >  >http://www.zfree.co.nzr >I   ------------------------------    Date: 09 Apr 2002 00:28:50 -07003 From: Eric Smith <eric-no-spam-for-me@brouhaha.com> 2 Subject: Re: Predictions - just for the hell of it0 Message-ID: <qhadsde3f1.fsf@ruckus.brouhaha.com>  H Jeff Campbell <jcampbell@ins-msi.com> writes about the inability of most+ DEC floppy disk subsystems to format disks:MF > Not piracy, but interchange. Not allowing customer formatting almost4 > eliminates service calls for interchange problems.  G Can't see why.  Floppy drives use open-loop positioning, usually with ai? stepper motor.  So a drive that couldn't format properly due toSG positioner or head alignment error will still have problems reading and ( writing even with preformatted floppies.  H In fact, the problems with writing can be worse, because you can wind upC with media that has the address fields and data fields at differing E radii, which is much harder to recover data from than a disk that hasm> been formatted and written on the same out-of-alignment drive.  G I very much doubt that either piracy or interchange was the motivation.oD I think the initial RX01 design would have been more complicated andB expensive if it had been designed to support formatting, so it wasA simply a cost-cutting measure.  Then someone noticed that it alsoEG allowed them to charge more money for formatted diskettes, and that led B them to deliberately leave out formatting capability in almost allG floppy subsystems they designed thereafter.  But of course this is only: supposition on my part.p   ------------------------------  $ Date: Tue, 9 Apr 2002 09:50:48 +0200* From: Thord Nilson <thordn@stacken.kth.se>2 Subject: Re: Predictions - just for the hell of itM Message-ID: <Pine.GSO.4.21.0204090937080.28846-100000@chicken.stacken.kth.se>   & On Mon, 8 Apr 2002, Kevin Handy wrote:   > Bill Todd wrote:N > > "Malcolm MacArthur" <malcolmix@neverness.freeserve.co.uk> wrote in messageA > > news:Xns91EAD23DB7B73malcolmixnevernessfree@195.92.193.157...r > M > >>peripherals were made (and for which even fewer can still be found), and,m3 > >>as mentioned, couldn't format its own floppies.G > >  > > G > > I just can't remember the rationale for this.  There were certainlyAO > > piracy-protection efforts back then that involved special floppy formating,30 > > and that might have been part of the reason. > >  > < > I always heard that it was DEC's attempt at forcing you to? > buy the floppies from them. I was also told that you couldn'to< > use floppies with the reinforced hubs as they could damage > the RX50 drive itself. > F In one project we had a microvax-ii that we had to backup on floppies.C We found that about 5 foppies of 100 could not be written. First weSA thought that these was defective, but later we found that if you  7 removed and then inserted the floppy again, it was ok. O  C Thus the problem was that the floppy was not always centered on theF@ hub in the RX50. Since a saveset was some 80-100 floppies, if weH had format/backup/verified in one go, the probability that we ever could# restore the set again was very low..  = Soloution to the problem... don't fix the floppy drives, sell2 preformated floppies instead.S   /Thord.O   ------------------------------   Date: 9 Apr 2002 04:32:16 -0700R) From: P.Young@unsw.EDU.AU (Patrick Young)X2 Subject: Re: Predictions - just for the hell of it= Message-ID: <55f85d77.0204090332.1cfbc83c@posting.google.com>>  ^ Jeff Campbell <jcampbell@ins-msi.com> wrote in message news:<3CB2506B.833EF689@ins-msi.com>...  -< > RX01, RX02 and RX50 subsystems cannot format their media.   B RX01 was a really screwed up mess - I can remember unsoldering and8 swapping those unreliable PROMS in it to get it working.  6 It always reminds me of an Apple II floppy in just how non standard it was.   RX02 could be formatted....    .FORMAT DY:I   from the RT-11 days :-)h  @ IIRC you could (re)format RX01 single density media for the RX01 on the RX02:   .FORMAT DY:/SINGLE  A Yup, _just like_ VMS you can leave out the unit number and have 0y assumed :-)    ------------------------------   Date: 9 Apr 2002 11:40:06 GMTu( From: peter@taronga.com (Peter da Silva)2 Subject: Re: Predictions - just for the hell of it1 Message-ID: <a8ujum$cti$1@citadel.in.taronga.com>   0 In article <qhadsde3f1.fsf@ruckus.brouhaha.com>,5 Eric Smith  <eric-no-spam-for-me@brouhaha.com> wrote: I >Jeff Campbell <jcampbell@ins-msi.com> writes about the inability of most , >DEC floppy disk subsystems to format disks:G >> Not piracy, but interchange. Not allowing customer formatting almosti5 >> eliminates service calls for interchange problems.s  H >Can't see why.  Floppy drives use open-loop positioning, usually with a@ >stepper motor.  So a drive that couldn't format properly due toH >positioner or head alignment error will still have problems reading and) >writing even with preformatted floppies.   I I ran into this... back in the early '80s I used to use some Apple IIs atDI the university to work on, and I started getting a lot of "bad" floppies. J It turned out that one of the apples had a bad drive, way out of alignmentJ but the Apple "software controller" scheme was able to compensate. Mostly.  I So even floppies formatted in other drives would gradually go bad if theyD! were written to on that computer.t   -- o@ Rev. Peter da Silva, ULC.	                                 WWFD?  F "Be conservative in what you generate, and liberal in what you accept" 	-- Matthew 10:16 (l.trans)0   ------------------------------   Date: 9 Apr 2002 08:36:22 -0500e- From: koehler@encompasserve.org (Bob Koehler)n2 Subject: Re: Predictions - just for the hell of it3 Message-ID: <yo6H3V3t0yHC@eisner.encompasserve.org>e   In article <Xns91EAD23DB7B73malcolmixnevernessfree@195.92.193.157>, Malcolm MacArthur <malcolmix@neverness.freeserve.co.uk> writes:  > @ > Judging from the fact he talks about the PC competing with theC > "'real' 11", he must be referring to the DEC Professional line ofPE > PDP-11 based personal computers. It's based around a microprocessor,H > implementation of the PDP-11 (J-11??) and shipped with a not-very-goodC > menu-driven operating system called P/OS. Some people may cruellyeC > suggest that the initials stand for 'a quantity of excrement' ;-)h  @    Pro 350 were F-11, 380 were J-11.  I've heard of a Pro 325 (a    stripped down 350).  M > They look like Rainbows, and, rather oddly, use a Yale-type key switch for S > the power.  E    Funny, I've never seen a key switch on a Pro.  Certainly not on mye    350.e  E > P/OS is a cut down version of RSX, with some extra features as wellID > (subdirectories come to mind). Development and CLI access was madeJ > possible with the P/OS Toolkit. RT-11 was also available and some third  > party unix called Venix.  A    There are a couple of UNIX out there, but who'd want to ruin a C    perfectly good PDP with this stuff?  Sparc 2 Suns go for about 20    bucks at hamfests.   A    Anybody know which was the earliest version of P/OS to provideSB    subdirectories and where I can download it?  Currently I've gotC    P/OS 2.0 (IIRC), and I need to update it anyway because I've goti,    a later version of DECnet I want to load.   > F > -Malcolm (Proud owner of a Professional 380, 512 KW core (well, OK, L > they're DRAMS...), P/OS (unfortunately. Want RSX. Anyone know how to make 6 > it work? :) and Venix (requires a hard drive swap)).   ------------------------------  % Date: Tue, 09 Apr 2002 17:23:23 +0100e% From: Alan Greig <a.greig@virgin.net>i2 Subject: Re: Predictions - just for the hell of it8 Message-ID: <it46bu4gj772t8s9cj8fg2nhbcgliu7a8j@4ax.com>  - On 8 Apr 2002 20:28:20 GMT, Malcolm MacArthurl, <malcolmix@neverness.freeserve.co.uk> wrote:     >i >DCL agrees: >$ call dcl sexyF >%DCL-E-INVCALL, invalid CALL nesting structure or data inconsistency 	 >detectedn   :-)e    J >... which isn't surprising, considering that with UCX 1.0 you had to pay H >for something the Unix people were bundling for free (and UCX probably J >wasn't as good). What I guess no-one realised (and it's easy to see with   A Never mind Unix. Tenex/TOPS-20 had Arpanet support built into theS! operating system back around 1970   H >hindsight, which of course has 20/20 vision...) is the extent to which J >networking computers and getting them to talk to each other would become 1 >just as important as the computers themselves...O  D But that was exactly what DEC computers were doing back then. It was@ interesting that, in the early 80s, DEC provided online Internet? support for TOPS-20 but not VMS. A DEC-20 called something like.F market.lcg.dec.com. In fact .dec.com may even have been the very first ".COM"  > DEC had no excuses for the lack of internet support other than arrogance. a       >> --u >> Alane >> > >W
 >-Malcolm. >F- >Subtract nine for e-mail (anti-spam measure)n   ------------------------------  % Date: Tue, 09 Apr 2002 13:10:00 -0400m- From: JF Mezei <jfmezei.spamnot@videotron.ca>O2 Subject: Re: Predictions - just for the hell of it, Message-ID: <3CB32065.901156FA@videotron.ca>   Eric Smith wrote:EF > I think the initial RX01 design would have been more complicated andD > expensive if it had been designed to support formatting, so it was  > simply a cost-cutting measure.  H Did Digital actually design and build the diskette drives ? I would haveJ though that they would have simply oem'd them and put their onw board with) their own firmware and interface on them.L  L Also, someone told me that the diskette drives on MACs (whi have the abilityM to eject) are pretty standard and it is just a question of Windows not having K the right software to tell the standard diskette drives to "eject". Is that  true ?   ------------------------------  # Date: Tue, 09 Apr 2002 18:06:18 GMTo+ From: Jeff Campbell <jcampbell@ins-msi.com>m2 Subject: Re: Predictions - just for the hell of it+ Message-ID: <3CB323DA.AB81DABB@ins-msi.com>    Patrick Young wrote: > ` > Jeff Campbell <jcampbell@ins-msi.com> wrote in message news:<3CB2506B.833EF689@ins-msi.com>... > = > > RX01, RX02 and RX50 subsystems cannot format their media.f > D > RX01 was a really screwed up mess - I can remember unsoldering and: > swapping those unreliable PROMS in it to get it working. > 8 > It always reminds me of an Apple II floppy in just how > non standard it was. >  > RX02 could be formatted....c > 
 > .FORMAT DY:e >  > from the RT-11 days :-)  > B > IIRC you could (re)format RX01 single density media for the RX01 > on the RX02: >  > .FORMAT DY:/SINGLE > C > Yup, _just like_ VMS you can leave out the unit number and have 0 
 > assumed :-)g  E The RT-11 format command does NOT format RX02 diskettes. The physicals@ track format already exists. Degauss an unimportant diskette and try it. 8-)b  
 Jeff Campbelln n8wxs@arrl.net   ------------------------------  # Date: Tue, 09 Apr 2002 17:56:20 GMT * From: "Bill Todd" <billtodd@metrocast.net>2 Subject: Re: Predictions - just for the hell of itB Message-ID: <8XFs8.302683$Gf.27510358@bin2.nnrp.aus1.giganews.com>  : "Bob Koehler" <koehler@encompasserve.org> wrote in message- news:yo6H3V3t0yHC@eisner.encompasserve.org...l   ...   C >    Anybody know which was the earliest version of P/OS to provideV >    subdirectories>  F As best I can remember I added that in V3 to support multi-workstation5 servers (and in the process multi-user workstations).n    and where I can download it?   I Sorry.  But weren't there some references earlier in this thread to that?X   - bill   ------------------------------  $ Date: Tue, 9 Apr 2002 10:33:17 +0200, From: "Toine Dirven" <tdirven@volvocars.com>! Subject: Processes in state RWSCSR3 Message-ID: <a8u8vu$67f6@eccws12.dearborn.ford.com>P   Hello,  K We have a OpenVMS cluster with 4 alpha nodes ( 3 * AS8200 and one ES40). Wet run OpenVMS 7.2-2DF with Oracle Rdb SQL V7.0-62. We have one lock master in our cluster to prevent lock remastering.7  H  At the moment we have a problem with batch-applications which runs on a" node which is not the lock master.J These Rdb batch programs takes a long time (15 hours) to do their job. The' process is all the time in state RWSCS.0  G I have increased the cluster credits to 128 on each alpha node but thisV
 doesn't help.0  $ What should I do to get this faster.  ) Put LOCKDIRWT on each node to value 1 ???   
 Best regards,t   Toine Dirven Volvo Cars Gentr   ------------------------------  # Date: Tue, 09 Apr 2002 09:56:28 GMTM' From: "Hans Vlems" <hvlems@zfree.co.nz>1% Subject: Re: Processes in state RWSCS0" Message-ID: <3cb2bac9@zfree.co.nz>   Toine,  . pls post the results of MC SYSGEN SHO /CLUSTER7 Take a look at MSCP_CREDITS (# of outstanding IO's) andt MSCP_CMD_TMO (should be 0). @ Changing LOCKDIRWT might do more harm since lock remastering is  an expensive process. + What kind of a cluster are you running, CI?>   Hans  - "Toine Dirven" <tdirven@volvocars.com> wrote:m >Hello,e >pI >We have a OpenVMS cluster with 4 alpha nodes ( 3 * AS8200 and one ES40).e We >run OpenVMS 7.2-2G >with Oracle Rdb SQL V7.0-62. We have one lock master in our cluster to1 >prevent lock remastering. > I > At the moment we have a problem with batch-applications which runs on a # >node which is not the lock master.tK >These Rdb batch programs takes a long time (15 hours) to do their job. The0( >process is all the time in state RWSCS. >rH >I have increased the cluster credits to 128 on each alpha node but this >doesn't help. >s% >What should I do to get this faster.E >m* >Put LOCKDIRWT on each node to value 1 ??? >6 >Best regards, >=
 >Toine Dirvene >Volvo Cars Gent >a >        http://www.zfree.co.nz   ------------------------------  % Date: Tue, 09 Apr 2002 11:43:27 +0100c( From: Nic Clews <sendspamhere@127.0.0.1>% Subject: Re: Processes in state RWSCS ) Message-ID: <3CB2C5CF.B1801DDB@127.0.0.1>    Toine Dirven wrote:tM > We have a OpenVMS cluster with 4 alpha nodes ( 3 * AS8200 and one ES40). We  > run OpenVMS 7.2-2oH > with Oracle Rdb SQL V7.0-62. We have one lock master in our cluster to > prevent lock remastering.r > J >  At the moment we have a problem with batch-applications which runs on a$ > node which is not the lock master.L > These Rdb batch programs takes a long time (15 hours) to do their job. The) > process is all the time in state RWSCS.c > I > I have increased the cluster credits to 128 on each alpha node but this  > doesn't help.: > & > What should I do to get this faster. > + > Put LOCKDIRWT on each node to value 1 ???M  E More information would be helpful here, particularly what the clustert interconnects are.  F I have seen this effect before but I'm not sure you're doing the rightF thing to prevent lock remastering. Old RDB used to master and keep theE locks to itself wherever it started and this could have the effect of>> making distributed locks be mastered on an inappropriate node.  F I'm not sure why you are saying you have only one lock manager runningF in the cluster. This should not happen in a cluster. (Are you implying, one node with LOCKDIRWT=1 and all others 0?)  H Each node has a lock manager, and the lock manager will manage all localG (node private) locks and a "designated proportion" (depending on SYSGEN]F parameters and current operating conditions) of the distributed locks.  F The distribution attempts to gets the locks to be mastered on the mostB appropriate way, and RWSCS in some situations will be unavoidable.  G However, you've said that [or I have interpreted as] you have trying too0 prevent lock remastering *and* are seeing RWSCS.  H RWSCS means "this process is waiting for the SCS layer of the cluster toE do its work". This is probably (but not conclusively) due to the lockoH manager having to communicate with the lock directory manager and/or theH lock master for a lock. (Unless you've understood that last sentence, DO< NOT try to force the cluster to do anything with its locks!)  D By preventing the cluster from taking the remastering decisions, youH could be inducing this effect. That is my first comment and observation,D so my first suggestion is *don't* force the cluster to master or not3 remaster locks in any way, and merely observe then.>  E I would then say only force mastering conditions if you're absolutely  sure what you're doing.r  G My next comment revolves around no matter what you do, you still end up H with RWSCS giving you application issues. What you need to do is address the cluster interconnect.   F For example, if you have one network card in each Alpha connected to aG 10MBs LAN, shared with other systems (so it is busy) it is not going touH help. Consider a private 100MBs connection through a private hub betweenA the systems, or even for lower latency, Memory Channel and a hub.e  H After re-reading my response to you, I *guess* you have set LOCKDIRWT atH 1 on the ES40, and 0 on the 8400s, and your batch application is runningG on one of the 8200s. I suggest setting LOCKDIRWT the same on all nodes,n and see what happens.r  G (Reason: It is flawed to assume that lock management potential is basedc on CPU power alone)   @ I think I'm starting to get too deep and complex when (with someA assumptions on my behalf) this could be an application issue. For E example, you can have a node which is the lock master for a lock, yet E due to the lock directory (and failing to use the NL lock) inter nodeo= communication has to take place. This could boil down to sometF application redesign, but I think you should take one stage at a time.   As to your one line (repeated)  + > Put LOCKDIRWT on each node to value 1 ???h   Yes. (PE1 should also be 0).  4 I expect Keith Parris will enter at this juncture...   -- f( Regards, Nic Clews CSC Computer Sciences nclews at csc dot come   ------------------------------  % Date: Tue, 09 Apr 2002 08:13:46 -0400-2 From: norm lastovica <norman.lastovica@oracle.com>% Subject: Re: Processes in state RWSCSr* Message-ID: <3CB2DAFA.3930B947@oracle.com>   Nic Clews wrote: >  > Toine Dirven wrote:SO > > We have a OpenVMS cluster with 4 alpha nodes ( 3 * AS8200 and one ES40). We  > > run OpenVMS 7.2-2 J > > with Oracle Rdb SQL V7.0-62. We have one lock master in our cluster to > > prevent lock remastering.X > >AL > >  At the moment we have a problem with batch-applications which runs on a& > > node which is not the lock master.N > > These Rdb batch programs takes a long time (15 hours) to do their job. The+ > > process is all the time in state RWSCS.o > >sK > > I have increased the cluster credits to 128 on each alpha node but thise > > doesn't help.r > >g( > > What should I do to get this faster. > >a- > > Put LOCKDIRWT on each node to value 1 ???C > G > More information would be helpful here, particularly what the clusterh > interconnects are. > H > I have seen this effect before but I'm not sure you're doing the rightH > thing to prevent lock remastering. Old RDB used to master and keep theG > locks to itself wherever it started and this could have the effect ofw@ > making distributed locks be mastered on an inappropriate node.  @ 	When a database is first opened, the lock tree for the databaseE is created.  If lock remastering is disabled, the lock tree will stay G on that node (this has nothing to do with Rdb and everything to do withb the lock manager).  o  * > Regards, Nic Clews CSC Computer Sciences > nclews at csc dot comw   -- o> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  % Date: Tue, 09 Apr 2002 08:15:24 -0400 2 From: norm lastovica <norman.lastovica@oracle.com>% Subject: Re: Processes in state RWSCS:* Message-ID: <3CB2DB5C.EF29AD01@oracle.com>   Toine Dirven wrote:9 >  > Hello, > M > We have a OpenVMS cluster with 4 alpha nodes ( 3 * AS8200 and one ES40). We: > run OpenVMS 7.2-2sH > with Oracle Rdb SQL V7.0-62. We have one lock master in our cluster to > prevent lock remastering.e > J >  At the moment we have a problem with batch-applications which runs on a$ > node which is not the lock master.L > These Rdb batch programs takes a long time (15 hours) to do their job. The) > process is all the time in state RWSCS.t  F 	Try running this process on the node where the lock tree is mastered;I this will result in all lock requests being local with no DLM SCS trafficu between nodes.  G 	This is the performance effect of running in a cluster; locks requests N get passed around.  The more remote (off-node) locking you do within a single  tree, the worse it can look.   ------------------------------  $ Date: Tue, 9 Apr 2002 13:34:16 +0100* From: "Richard Brodie" <R.Brodie@rl.ac.uk>% Subject: Re: Processes in state RWSCSo, Message-ID: <a8un3r$1o5o@newton.cc.rl.ac.uk>  Y "Nic Clews" <sendspamhere@127.0.0.1> wrote in message news:3CB2C5CF.B1801DDB@127.0.0.1...:  H > For example, if you have one network card in each Alpha connected to aI > 10MBs LAN, shared with other systems (so it is busy) it is not going toFJ > help. Consider a private 100MBs connection through a private hub between > the systems,  O ...and check that the LAN paths are behaving. Half/full duplex grief, and otherr1 low-level network problems can manifest as RWSCS.    ------------------------------  $ Date: Tue, 9 Apr 2002 15:44:03 +0200$ From: "labadie" <labadie_g@decus.fr>% Subject: Re: Processes in state RWSCSd3 Message-ID: <AeCs8.2036$fL6.54197@news.cpqcorp.net>d  7 "Toine Dirven" <tdirven@volvocars.com> wrote in message - news:a8u8vu$67f6@eccws12.dearborn.ford.com...l > Hello, > J > We have a OpenVMS cluster with 4 alpha nodes ( 3 * AS8200 and one ES40). We > run OpenVMS 7.2-2tH > with Oracle Rdb SQL V7.0-62. We have one lock master in our cluster to > prevent lock remastering.y >gJ >  At the moment we have a problem with batch-applications which runs on a$ > node which is not the lock master.L > These Rdb batch programs takes a long time (15 hours) to do their job. The) > process is all the time in state RWSCS.o >tI > I have increased the cluster credits to 128 on each alpha node but thisE > doesn't help.i >v& > What should I do to get this faster. >u+ > Put LOCKDIRWT on each node to value 1 ???s >  > Best regards,m >e > Toine Dirven > Volvo Cars Gentn>  What is PE1 on all nodes ? May be you should run on all nodes this program	 begin url1L http://www.compaq.com/support/asktima/operating_systems/009F6DF1-BC7EC4E1-1C	 02A1.htmlD end url1D and, according to results, lower PE1, or put it to -1 on some nodes.8 Quite often, with Rdb, the best thing to do is PE1 to -1   Regardsn   Grard   ------------------------------  % Date: Tue, 09 Apr 2002 16:39:41 +0100a( From: Nic Clews <sendspamhere@127.0.0.1>% Subject: Re: Processes in state RWSCSh) Message-ID: <3CB30B3D.E7F96A7A@127.0.0.1>    norm lastovica wrote:  >  > Nic Clews wrote:  J > > I have seen this effect before but I'm not sure you're doing the rightJ > > thing to prevent lock remastering. Old RDB used to master and keep theI > > locks to itself wherever it started and this could have the effect of1B > > making distributed locks be mastered on an inappropriate node. > I >         When a database is first opened, the lock tree for the databaseyG > is created.  If lock remastering is disabled, the lock tree will stayfI > on that node (this has nothing to do with Rdb and everything to do withS > the lock manager).  H I'm not an RDB expert here, however I have observed lock remastering NOTE working, and all my attempts to force it to a particular node failed.IH When I took it up with our internal guy who works with RDB he says "yeahH it does that". To put the version in context it is circa the V6.1 VMS. ID can't question what he says, he can walk the lock database in SDA asB good as anyone, and he offered the only explanation for what I wasH observing. It could actually be more around the lock directory function,A this was the part of the application code giving grief (raising .8  D For the record SYSGEN parameters were fine, but even loading them to5 force a preferred lock master [for all trees] failed.i  G All I can say is that the observed evidence disagrees with the expectedu behaviour, hence my comment. -- c( Regards, Nic Clews CSC Computer Sciences nclews at csc dot comi   ------------------------------   Date: 9 Apr 2002 12:31:21 GMTd1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon) G Subject: Re: Relative invulnerability of VMS to buffer-overflow attacksi, Message-ID: <a8umup$21gl$1@info.cs.uofs.edu>  / In article <ub3vqudbki6g6c@corp.supernews.com>,s(  "RLC" <rlc@magicalsoftware.com> writes: |>B |> The nice thing about VMS for buffer overflow attacks is that if@ |> the web application (say HTTP server) is not given privilegesD |> and file system privs are set correctly, it doesn't really matter> |> that you can execute code.  Access woudl be limited so that- |> nothing significant could be done anyway.    C Which of course is true of a Unix system or any other OS as well.  cC The general lack of security in most Unix systems is due to mindsetrA (an old one that sadly is no longer acceptable for social reasonsuC rather than technical reasons) and not real technical shortcomings.-   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   E   ------------------------------   Date: 9 Apr 2002 10:33:28 -0700a From: mrotgeri@web.de (ROM)n Subject: remote mailboxa= Message-ID: <b3ad37dc.0204090933.2ef17331@posting.google.com>a  E I have a C++ program that wants to communicate from one node "A" to aa remote node "B" with a mailbox.eE On "B" I create a mailbox via sys$crembx() and a process can open them> mailbox with the fopen() function to write data to the mailbox (fwrite()).i  B From "A" I get a connection to the remote mailbox with the fopen()	 function.n. This all works fine, but there is one problem:  F I can't read from the mailbox (fread()) while the process on "B" has a connection to the mailbox.A Only after the process on "B" closes the connection (fclose()) an:/ other process on "A" can read from the mailbox.g  6 What is wrong? (The used Code fragment included below)   Create mailbox on node "B":p   unsigned char prmflg = 1;s unsigned int  chan   = 0;9# unsigned long maxmsg = sizeof(int);a  unsigned long bufquo = 0x10000L; unsigned long promsk = 0;s unsigned long acmode = 0;     ' $DESCRIPTOR ( MBXName, "MailBoxName" );o   status = sys$crembx(             prmflg,              &chan,             maxmsg,              bufquo,              promsk,              acmode,o             &logname );A   To connect to the mailboxd on "B":   ( fMBX = fopen( MAILBOX, "w", "ctx=bin" );   On "A":i  . fMBX = fopen( szMailBoxName, "r", "ctx=bin" );   To write to the mailbox on "B":y   int iWrite = 33;9 int numwritten = fwrite( &iWrite, sizeof(int), 1, fMBX );t    To read from the mailbox on "A":   int iResult[256];r' memset( &iResult, 0, sizeof(iResult) );mC int iCount = fread( &iResult, sizeof(int), sizeof(iResult), fMBX );   C The fread() function return only when the process on "B" closes the  connection:    fclose( fMBX );r  A I need to read simultaneous without closing the connection to theg mailbox. I hope someone can help me.r ROM1   ------------------------------  % Date: Tue, 09 Apr 2002 09:40:31 -0400e1 From: Michael Austin <maustin@firstdbasource.com> ! Subject: Re: Server Side Includes 2 Message-ID: <3CB2EF4F.BD7DBEC9@firstdbasource.com>   Rob Buxton wrote:  >  > For anyone interested.% > Had some offline e-mails with Alan.rB > And still finding it difficult to reconcile the behaviour of SSI > across Web Servers.. >  > The call:r > <!--#execnF > cmd="@dsa1:[apache.htdocs.wcc.command]CGI__FL_SHOW_SYS_DISK.COM" -->D > works fine under Apache. The .COM file is returning html formatted > text. Which is what I want.c  F Because of the way that IE and some other browsers handle certain file@ types, I ALWAYS leave off the .COM on OpenVMS systems as it just& confuses the heck out of the browsers.  ? I have had to do this on Cern, Purveyor, OSU and APACHE.  It isi0 generally not the web server, but the browser.    D I would be interested in seeing your code.   I have an anonymous ftpH server you can push it to me, just contact me offline for the site... it is not the domain listed.  -- T Regards,  7 Michael Austin            Registered Linux User #261163c7 First DBA Source, Inc.    http://www.firstdbasource.com' Sr. Consultant 704-947-1089 (Office)s 704-236-4377 (Mobile)W     > WASD displays the html codes.t >  > On Alans advice I've tried:aB > set /htdocs/wcc/command/*.com CONTENT=text/html in the Map File.6 > (htdocs is already set up as a path in the Map file) > ? > When that failed I copied the .COM as a .CMD and set up a new.1 > Content Type in the http$config.conf file, e.g.E > + > .CMD  text/html   DCL File to return HTML  > : > also changed the above set command to reference the .CMD >  All to no avail.eE > The only thing that delivers formatted text under WASD is to have aaC > type="text/html" as part of the exec call. Alas, this then breaksv) > Apache which fails to produce the page.a > G > In the Apache docs I have as part of the CSWS install, the conclusioneD > is that SSI is not a replacement for CGI. At this point I heartilyD > agree, there's too many differences between SSI interpretations on > different Web Servers. > G > As a postscript (and I don't get any kickbacks 8-) ) if Alans book isbC > anywhere near as informative as Alan has been, and as clear, thenoF > it'll be a very useful guide to anyone newish to VMS Web Servers and> > wanting things explained in terminology they can understand. > B > On Fri, 05 Apr 2002 08:21:05 GMT, winston@SSRL.SLAC.STANFORD.EDU0 > ("Alan Winston - SSRL Admin Cmptg Mgr") wrote: > a > >In article <3cad15f6.1291485798@news.wcc.govt.nz>, rob.buxton@wcc.govt.nz (Rob Buxton) writes: I > >>On Fri, 05 Apr 2002 02:58:38 GMT, rob.buxton@wcc.govt.nz (Rob Buxton) 
 > >>wrote: > >> > >>>Hi All, > >>> G > >>>I've been playing around with Server Side Includes on both my CSWShE > >>>(on OpenVMS Alpha) and WASD (on OpenVMS VAX) Web Servers and I'm 3 > >>>coming unstuck on a number of inconsistencies.. > >>>eE > >>>First off, there seems to be several SSI directives that are not 5 > >>>available on the CSWS Server (e.g. <!--#flastmodsJ > >>>file="dsa1:[www]shtml_1.shtml" --> will work on WASD but not on CSWS. > >s > >sF > >What's nailing you here is that CSWS (and OSU) restrict the file inR > >SSI statements with file="filespec" to being relative to the current directory.Q > >You can't move up the tree and you can't move to a different root. Want to getu= > >your own last modification date?  file="my-filename.shtml"u > > L > >WASD doesn't inflict that restriction on you; you can get to any file the9 > >webserver can get to (has appropriate privilege, etc).m > >sN > >You can use the virtual="/URLpath" to get to any mapped file, so you shouldK > >be able to do <!-- #flastmod virtual="/shtml_1.shtml" --> (assuming yourdM > >root is mapped to your /www/ directory).  This will work in WASD, OSU, anda > >CSWS. > > M > >(If you do an #include virtual of a CGI script, Apache will execute it andcL > >stick the results into your output document; the CGI script path can evenK > >include URL-encoded parameters in a query string, which is kind of cool.mC > >But be careful if you were actually trying to include the file.)r > >eR > >(Interestingly, all file access in WASD goes through the cache processor, whichM > >takes paths as well as filespecs, so - if I read the code right - the file 1 > >and virtual tags end up doing the same thing.)e > >  > >I > >' > >>>iK > >>>The WASD and CSWS share the same UIC so there are no ownership issues. D > >>>And I know SSI is working on CSWS because some directives work. > >>>bJ > >>>I've also found that the exec cmd SSI seems to funnction differently.I > >>>The Command file I'm calling from the .shtml page formats the output H > >>>with various html codes e.g. <TR> etc. On the CSWS Server this thenH > >>>displays as I would hope, alas on the WASD Server the <TR> etc. are# > >>>displayed as part of the text.o > >. > >c > >>>dK > >>>So, anyone know of any CSWS documentation that actually lists what SSI- > >>>directives can be used? > >1O > >I've been having good luck with _Professional Apache_ from Wrox Press.  It'slN > >Unix-oriented, not VMS-oriented, but the design goal for CSWS appears to beI > >to be as compatible as possible with Apache on Unix while still taking O > >advantage of VMS features "under the hood."  If you buy it for this purpose,f0 > >be aware that only about a third is relevant. > >(K > >You could also just go to www.apache.org, select the HTTP server, selectrN > >documentation, select 1.3, and look at the SSI tutorial and the mod_include > >documentation.> > >i > >>>iI > >>>Anyone know of why WASD presents the data back from an exec cmd callo@ > >>>unhtml'ed? or if there are ways of modifying the behaviour? > >y > >>>sI > >>>Before suggestions are made as regards replacing Apache with WASD on I > >>>the Alphas or similar, the main exercise I'm doing here is trying tosE > >>>learn more about Web Servers. So far I've learn't quite a lot asa& > >>>regards html and shtml standards! > >>>  > >nL > >Basically what's happened is that the CERN and NCSA servers invented waysJ > >to do a few things, and those few things ended up in OSU and WASD yearsK > >ago.  (NCSA eventually became Apache, although it's more like at the DNAML > >level than in actual inheritance of code.) Then all three servers evolvedM > >semi-independently, pretty much as their developers either felt a need for"A > >features or found things they'd be interested in playing with.o > > " > >>>As always, thanks in advance. > >>>a	 > >>>Rob.tH > >>Just a follow up, found that adding the type="text/html" to the execJ > >>cmd directive allowed the WASD Server to correctly display the output.F > >>Alas, now the CSWS Server gives an error occurred while processingF > >>directive error message but does still present the data correctly.J > >>I guess it still comes down to trying to determine what SSI directives% > >>are common between WASD and CSWS.x > > H > >If the content-type of an included/exec-ed file isn't text/html, WASDJ > >wraps the output in <PRE> </PRE> and you don't get your formatting tagsK > >interpreted.  If you really want to make a cross-platform compatible SSIlD > >document, the trick is not to use that tag, but instead in WASD's > >HTTPD$MAP.CONF to > >t/ > >SET /path-to-exec'ed-file  CONTENT=text/html  > >aN > >which will set the content-type of the file.  Then leave out the type= tag. > >o > >  > > J > >I've been sorting out which directives are common between those two andM > >OSU.  The killer is that even when the directives are the same, the syntaxxK > >or the semantics are different.   WASD adds a lot of extra functionalitytN > >through additional tags in similar directives, which makes Apache complain,0 > >although it does its best to keep processing. > >R7 > >>Might have to wait for Alans book to hit the shops!z > >RQ > >You don't have to wait for the book to hit the shops, but don't let getting an 8 > >answer now stop you from buying one when it does. :-) > >a
 > >-- Alan > >s > >hR > >===============================================================================3 > > Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDU0P > > Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056P > > Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210R > >=============================================================================== > >d   ------------------------------  % Date: Tue, 09 Apr 2002 16:17:19 +0200 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>.! Subject: Re: Server Side Includes-) Message-ID: <3CB2F7EE.DF5A260E@gtech.com>e   Rob Buxton wrote:rD > I've been playing around with Server Side Includes on both my CSWSB > (on OpenVMS Alpha) and WASD (on OpenVMS VAX) Web Servers and I'm0 > coming unstuck on a number of inconsistencies.  > SSI has is somewhat a hack per design, which was quite usefull when the WWW started..  4 Today that functionality is usually provide by other6 technologies like ASP (MS world) and JSP (Java world).   Arne   ------------------------------   Date: 9 Apr 2002 12:46:09 GMTo1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon) = Subject: Re: Spring Forward, Fall Back - unfortunately not :-1, Message-ID: <a8unqh$21gl$2@info.cs.uofs.edu>  8 In article <pnueauscimdopq24748rj2epohj8blf83m@4ax.com>,-  Jamie Stallwood <news@project76.net> writes:tH |> >Excuse me, but isn't the Daylight savings time the first *Sunday* inJ |> >*April* -- which isn't until next week - April 7th -- at least that isB |> >what it is here in the USA... Unless you guys are different... |> a= |> The ISO standard is the last sundays in March and October.e  G This is just like the K&R vs. ISO argument.  An American (Ben Franklin)tI invented the idea, a European changed it. And now people say the inventory was wrong.  ;-)e   bill   -- xJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   a   ------------------------------   Date: 9 Apr 2002 12:11:57 -0500m+ From: young_r@encompasserve.org (Rob Young)n> Subject: Re: VMS boottime - how to calculate # of days running3 Message-ID: <FVELX1gcEo4h@eisner.encompasserve.org>n   In article <3CB22A7C.DB7E9E13@spam-be-gone.intel.com.net>, "Kenneth H. Fairfield" <Kenneth.H.Fairfield@spam-be-gone.intel.com.net> writes: > Rob Young wrote: > k >> In article <dbcb2b6e.0204031051.6bd0aa1b@posting.google.com>, BHALPERN@UMCAZ.EDU (BRITT HALPERN) writes:WF >> > I am wondering if there is a quick way to determine the number ofK >> > days/time since VMS was last booted.  I know I can get the boottime bynI >> > using f$getsyi("boottime").  Not being a VMS guru, is there a way torE >> > subtract the current date/time from the boottime and display therH >> > number of days VMS has been running.  This would eventually be in a@ >> > daily report file.  If you can point me in the direction of/ >> > documentation, that would also be helpful.r >> > >>B >>         Assuming 7.1 or higher and borrowing a bit from Gotfryd7 >>         (by the way, it didn't like his f$trnlnm)...1 >> >> $ type up.com >> >> $ !R >> $ SetData       :==     (READ SYS$INPUT DATA ; DEFINE/JOB/NOLOG DATA_LOG &DATA)M >> $ GetData       :==     DATA_SYM == F$TRNLNM("""DATA_LOG""","""LNM$JOB""")u >> $ !> >> $ PIPE SHOW SYSTEM | SEARCH SYS$INPUT " uptime" | 'SetData' >> $ 'GetData'2 >> $ clean_data = f$edit(data_sym,"COMPRESS,TRIM") >> $ !5 >> $ days_up = f$integer(f$element(8," ",clean_data))s >> $ if days_up .lt. 2	 >> $ then  >> $       s = ""c	 >> $ elset >> $       s = "s"
 >> $ endifF >> $ write sys$output  "This node has been up for ",days_up," day''s'" >> $ @up > I > And the whole if-block and write can be condensed to a single statementr > using the F$FAO lexical: > J > $ Write Sys$Output F$Fao("This node has been up for !SL day!%S",days_up) > F > Note that !SL and !%S must be upper case, but the output of !%S will6 > correctly follow the case of the previous character. > E > I think $FAO (SYS and F versions) is one of the most under-utilizedcG > functions around.  It easily does almost any formayting task you wantE > _except_ floating-point... >    Ken,  0 	That is slick!!  Under-utilized?  I am weak... 1 	When it comes time for f$fao, here is what I do:o  / NODENAME_ROCKET_ROBBIE> diren/grand [...]*.com;i    * Grand total of 128 directories, 834 files.   		$ search [...]*.com f$faoh    < DISK$DISK99:[RR.RYOUNG.ROB.DISKSPACE]SPANISH_BACK_ARCH.COM;1  L $       ending_hour =  f$string(f$fao("!2ZL",start_hour + 4)) ! Run for four hour rs at a stretchb   ******************************8 DISK$DISK99:[RR.RYOUNG.ROB.DSN]GOOD_SYSTEM_MONITOR.COM;1  2 $  text = F$FAO("!32AS  !9UL  !9UL  !9UL  !9UL", -   ******************************4 DISK$DISK99:[RR.RYOUNG.ROB.DSN]VMSMAIL_PROFILE.COM;1  . $ Say F$FAO(Faoctl, Mail_Username, Mail_Count)D $            Say F$FAO("!_The following flags are set:", Mail_Flags)   ******************************9 DISK$DISK99:[RR.RYOUNG.ROB.VAXMAGS.PLAYLAND]DISKCOM.COM;1s  * $writerec:      outsiz = f$fao("!6ZL",sum)0 $               outroot = f$fao("!30AS",rootnam)   ******************************  > 	and modify what I find for current use.  Same for f$getqui.     				Robn   ------------------------------   Date: 9 Apr 2002 08:44:58 -0500l- From: koehler@encompasserve.org (Bob Koehler)  Subject: Re: VMS doesn't sucko3 Message-ID: <k2ySPhBlGrrd@eisner.encompasserve.org>   c In article <HD1y5RF1bVJK@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:re > In article <a8slp1$v5pt6$1@ID-46415.news.dfncis.de>, "James Gessling" <jgessling@yahoo.com> writes:o >> See here: >> b >> http://srom.zgp.org/  > @ > Imagine the scores if their parser were able to deal with more > dignified language.y  B    Since they won't search for PH-UX, they're somewhat missing the	    point.T   ------------------------------   Date: 9 Apr 2002 08:49:55 -0500a- From: koehler@encompasserve.org (Bob Koehler)  Subject: Re: VMS doesn't suck 3 Message-ID: <B$n4ry6cA612@eisner.encompasserve.org>   c In article <a8slp1$v5pt6$1@ID-46415.news.dfncis.de>, "James Gessling" <jgessling@yahoo.com> writes:  > See here:  >  > http://srom.zgp.org/  >    I see on thier link Fortran rules.  Maybe reality seeps in.   ------------------------------  # Date: Tue, 09 Apr 2002 17:03:28 GMTt% From: jlsue <jlsuexxxz@insightbb.com>  Subject: Re: VMS doesn't suck.8 Message-ID: <ck76bucdegf1a39aqfdnsl1ndf8gfcf0el@4ax.com>  3 On Mon, 8 Apr 2002 10:59:06 -0700, "James Gessling". <jgessling@yahoo.com> wrote:  
 >See here: >n >http://srom.zgp.org/d >N  > Um... As much as I like the sentiment, this site should not beF confused with any resemblance to real statistical analysis.  But it is fun to see that graph.  1 Not speaking for anyone, certainly not DEC/Compaq - (get rid of the xxxz in my address to e-mail)o   ------------------------------   Date: 9 Apr 2002 08:33:27 -0700n0 From: the.janitor@zero-knowledge.de (tom vedder)! Subject: VMS License PAK Problems < Message-ID: <5fa28f1e.0204090733.6589956@posting.google.com>   Hardware: DEC 3000/300X  Software: VMS 7.2  Knowledge: newbieo  K When logging into CDE for the first time as "system", this message appears:i   "LMF license check has failed."d  L I already read MGMT59 (Resolving License PAK Problems?) of the OpenVMS faqs. But that's of no help.   Tomo   ------------------------------  # Date: Tue, 09 Apr 2002 16:38:10 GMTe% From: jlsue <jlsuexxxz@insightbb.com> ) Subject: Re: VMS To Tru64 Migration Plansa8 Message-ID: <g566bus5j7honlk220qr1vqs1ebl56tajh@4ax.com>  , On Mon, 08 Apr 2002 12:30:01 -0400, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:r   >Bob Ceculski wrote:F >> Itanium has a future because Intel needs a 64 bit platform and theyF >> wouldn't have picked up alpha rights and engineers if they were not4 >> serious about getting into high end computing ... >oJ >Or simply eliminate a chp design that was far superior to its own becauseN >Intel knew it would never be able to match Alpha. There are two ways to go upR >in a competition. Be better than your competitors, or eliminate your competitors. > [ >Intel chose to eliminate Alpha because IA64's core design is just too full of compromises.m  F Well, this is almost more of a personal prediction than a statement ofC any known facts.  We really don't know, yet, what Intel's plans are-< for the Alpha technology it purchased.  It may or may not be# "eliminated" - whatever that means.u  1 Not speaking for anyone, certainly not DEC/CompaqB- (get rid of the xxxz in my address to e-mail)e   ------------------------------  % Date: Tue, 09 Apr 2002 16:45:28 +0200A= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>dK Subject: Re: What OS Was www.openvms.digital.com Running On n 22-Sep-1998 ?s) Message-ID: <3CB2FE88.DE28162C@gtech.com>o   Jerry Leslie wrote:t> > What operating system was www.openvms.digital.com running on> > September 22, 1998, the day that the OpenVMS home page read: > B >   "OpenVMS is today what Microsoft wants Windows NT v8.0 to be!" > , > Was it OpenVMS, Digital Unix, or Windows ?   I think it has always been VMS.r  0 It started with Purveyor but switched to Apache.   Arne   ------------------------------   Date: 9 Apr 2002 08:30:00 -0500q- From: koehler@encompasserve.org (Bob Koehler)f, Subject: Re: Who has (ever) used DECmigrate?3 Message-ID: <aiWCRG5LFGyv@eisner.encompasserve.org>-  _ In article <8APR02.16394901@feda01.fed.ornl.gov>, Dave Greenwood <greenwoodde@ornl.gov> writes:: > G > I VESTed a number of utilities and shareable images to jump-start ourNJ > migration from VAX to alpha.  Those have all been natively built by now.J > However, I did VEST a 3rd party product because the vendor was no longerC > producing a VMS version of the code - let alone an alpha version.i > G > And, I did VEST a couple of programs for one of our uses who had losti > his source code.  D    I VESTed a DECUS utility written in SCAN, since there was no SCAN>    compiler on Alpha and we didn't want to reinvent the wheel.  E    Then I went back and patched the original VAX image and re-VESTed,5    still worked fine.b   ------------------------------  % Date: Tue, 09 Apr 2002 10:34:54 -0400-1 From: Michael Austin <maustin@firstdbasource.com>i, Subject: Re: Who has (ever) used DECmigrate?2 Message-ID: <3CB2FC0E.A2C12A8B@firstdbasource.com>   Dave Greenwood wrote:v > G > In a previous article, Didier Morandi <Didier.Morandi@Free.fr> wrote: E > > I'm interested to get field feedback from users of the DECmigrateiF > > utility, aka VEST, which, for those who do not know about it, is aK > > binary image translator produced by the DEC Migration Engineering GrouphL > > when the Alpha came to life, to allow VAX executable images to be run on > > Alpha systems. > > L > > After 20 years of VMS support, as I said in another post here, I know of > > noone in Europe. > >tD > > I personally think that the users did find their sources and did+ > > compile/link them on their new systems.? > G > I VESTed a number of utilities and shareable images to jump-start our1J > migration from VAX to alpha.  Those have all been natively built by now.J > However, I did VEST a 3rd party product because the vendor was no longerC > producing a VMS version of the code - let alone an alpha version.t > G > And, I did VEST a couple of programs for one of our uses who had lostt > his source code. >  > Dave > --------------; > Dave Greenwood                Email: Greenwoodde@ORNL.GOViJ > Oak Ridge National Lab        %STD-W-DISCLAIMER, I only speak for myself    D I was contracted to port a FORTRAN program running on a VAX that wasG originally written on a PDP11 to Alpha.  After a few weeks of not beingYE able to get things to work correctly, I stumbled on to something thatu@ made me go look at the dates of the sources vs. the dates of theF executables.   We had source code that was newer than the dates of theC executables -- and there was only one version.  Somewhere along thetG line, some changes were going to be made tht never got compiled and thetH disks were purged leaving me with no alternative but to VEST the currentA images and let it run. I had make some minor logical name changes B because the original code ran on a VAX 3400 with 2 CXY08's and allD terminals/PC's were connected via these serial connections.  The newD Alpha uses some DECServers and network connections.  The application? generates code for a plasma cutter for an air-conditioning ducthF manufacturing process that gets downloaded via serial connection via aD DECServer90)  The DS20 (they could have easily used a DS10L) is fastB enough, that I do some additional processing of the files that get. generated and it is still faster than the VAX.  F BTW, If anyone close to Charlotte NC wants a VAX3400 running VMS 5.0-0F w/decnet only TK70/32 MB of memory and 3 disks (2DSSI and 1 3rd party); let me know.  I am sure they would part with it very cheap.e   -- e Regards,  7 Michael Austin            Registered Linux User #261163>7 First DBA Source, Inc.    http://www.firstdbasource.comc Sr. Consultant 704-947-1089 (Office)  704-236-4377 (Mobile)I   ------------------------------  % Date: Tue, 09 Apr 2002 16:48:28 +0200l= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>S, Subject: Re: Who has (ever) used DECmigrate?) Message-ID: <3CB2FF3C.3D74E0E9@gtech.com>n   Didier Morandi wrote:eC > I'm interested to get field feedback from users of the DECmigraterD > utility, aka VEST, which, for those who do not know about it, is aI > binary image translator produced by the DEC Migration Engineering GroupeJ > when the Alpha came to life, to allow VAX executable images to be run on > Alpha systems. > J > After 20 years of VMS support, as I said in another post here, I know of > noone in Europe. > B > I personally think that the users did find their sources and did) > compile/link them on their new systems.w   I have played a bit with it.  , But I have not used it for production usage.   Arne   ------------------------------  % Date: Tue, 09 Apr 2002 17:06:37 +0100 % From: Alan Greig <a.greig@virgin.net> , Subject: Re: Who has (ever) used DECmigrate?8 Message-ID: <ac46buo2574n6ll9ak55c84v6722n1l0at@4ax.com>  2 On Sat, 06 Apr 2002 09:30:55 +0200, Didier Morandi <Didier.Morandi@Free.fr> wrote:   B >I'm interested to get field feedback from users of the DECmigrateC >utility, aka VEST, which, for those who do not know about it, is anH >binary image translator produced by the DEC Migration Engineering GroupI >when the Alpha came to life, to allow VAX executable images to be run onA >Alpha systems.i >dI >After 20 years of VMS support, as I said in another post here, I know ofe >noone in Europe.   D We used it on a number of executables back at my previous universityC employer.  At my current employer we have one vested commercial app F dating from 1986. During the Alpha upgrade we finally managed to trackE down the owners of the successor product - no longer available on VMS 2 - and they agreed we could translate the original.  A >I personally think that the users did find their sources and didx( >compile/link them on their new systems. >P >Thanks. >o >D.e   ------------------------------  # Date: Tue, 09 Apr 2002 16:35:20 GMTs% From: jlsue <jlsuexxxz@insightbb.com>,1 Subject: Re: Why VMS is "unhackable" lesson 1 ...A8 Message-ID: <um56bu41pc1qqrge81nevihd2pvfjl2rnv@4ax.com>  @ On 8 Apr 2002 18:13:56 -0700, bob@instantwhip.com (Bob Ceculski) wrote:   >Andrew Harrison SUNUK Consultancy <andrew_nospam.remove_this.harrison@sun.com> wrote in message news:<3CB1D6C2.9000900@sun.com>...UD >> Perhaps you could illuminate the choir on this one, the pervading9 >> belief is that OpenVMS is immune from buffer overruns.  >>  
 >> Regards >>   >> Andrew Harrison >sC >you see Andrew, it's like this ... on VMS, I put you in a box, anddE >you can only move around and do the things I let you in your box ...sB >without vms privilidges, you can't do a thing, unless I use those@ >powerful acl's or proxies to extend your box ... I can give youG >whatever I need to give you, but not give you access to the system ...=F >in other words Andrew, a properly configured vms box is "UNHACKABLE"!  C Well, to be fair, the OS *is* vulnerable to whatever weaknesses areaD introduced via privileged software.  In my experience, the OS itself= is generally more secure.  And most times I think it would beOA difficult to induce the system to execute data as code via buffer=	 overruns.=  C However, buffer overruns could cause other behavior that would make F you unhappy - like system crashes.  If you get an ACCVIO caused from aC buffer overrun, and you're in certain priv'd modes, the system will-F err on the side of caution (so to speak) and you'll be rebooting.  TheC "ping of death" problems - circa UCX V4.1 or .2 iirc - was one such E animal.  Now, to be fair, this was pre-V5.x when we got the new code.oF Also, to be fair, most TCP/IP devices had the same problem - including8 the HP printers that we had where I worked at that time.  C But, there were CERT advisories to alert people of this weakness atM? that time.  And I don't think all TCP/IP stacks were vulnerable- either.   1 Not speaking for anyone, certainly not DEC/Compaqs- (get rid of the xxxz in my address to e-mail)n   ------------------------------  % Date: Tue, 09 Apr 2002 15:07:23 +0200f5 From: Martin Knoblauch <Martin.Knoblauch@TeraPort.de>h= Subject: Re: Why VMS is "unhackable" lesson 3 ... (follow-up)s+ Message-ID: <3CB2E78B.BFB59AE0@TeraPort.de>i   Larry Kilgallen wrote: > m > In article <N0ks8.2005$fL6.47616@news.cpqcorp.net>, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> writes: M > > It doesn't have to do it that way.  But yes, there are JIT compilers that P > > generate machine code on the fly.  They have to do more than simply generateJ > > the code and jump to it on Alpha.  They also need to get it out of thr > > dcache into the icache.  > C > Ok, I'll bite.  Why is it necessary to get it out of the dcache ?a  H  because you cannot execute intructions from d-cache? :-) Similar thingsB have to be done on MIPS if you want to execute stuff from the dataE segment. IIRC, it is enough (on MIPS) to just flush the dcache before- trying to execute.   Martin --  B ------------------------------------------------------------------B Martin Knoblauch         |    email:  Martin.Knoblauch@TeraPort.de7 TeraPort GmbH            |    Phone:  +49-89-510857-309C7 C+ITS                    |    Fax:    +49-89-510857-111e5 http://www.teraport.de   |    Mobile: +49-170-4904759i   ------------------------------  % Date: Tue, 09 Apr 2002 15:17:45 +0200o5 From: Martin Knoblauch <Martin.Knoblauch@TeraPort.de>r= Subject: Re: Why VMS is "unhackable" lesson 3 ... (follow-up)e+ Message-ID: <3CB2E9F9.50675092@TeraPort.de>l   Martin Knoblauch wrote:e >  > Larry Kilgallen wrote: > >to > > In article <N0ks8.2005$fL6.47616@news.cpqcorp.net>, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> writes: O > > > It doesn't have to do it that way.  But yes, there are JIT compilers thatyR > > > generate machine code on the fly.  They have to do more than simply generateL > > > the code and jump to it on Alpha.  They also need to get it out of thr > > > dcache into the icache.m > >rE > > Ok, I'll bite.  Why is it necessary to get it out of the dcache ?t > J >  because you cannot execute intructions from d-cache? :-) Similar thingsD > have to be done on MIPS if you want to execute stuff from the dataG > segment. IIRC, it is enough (on MIPS) to just flush the dcache beforeR > trying to execute. >   H  oops. Seems that it is (on MIPS at least) the other way round. You haveB to flush the i-cache before trying to execute stuff out of memory.  ) http://yarchive.net/comp/self_modify.htmls   Martin -- sB ------------------------------------------------------------------B Martin Knoblauch         |    email:  Martin.Knoblauch@TeraPort.de7 TeraPort GmbH            |    Phone:  +49-89-510857-309n7 C+ITS                    |    Fax:    +49-89-510857-111F5 http://www.teraport.de   |    Mobile: +49-170-4904759    ------------------------------  # Date: Tue, 09 Apr 2002 16:50:23 GMT % From: jlsue <jlsuexxxz@insightbb.com>i= Subject: Re: Why VMS is "unhackable" lesson 3 ... (follow-up)o8 Message-ID: <3s66buocdq2d4o1k5n2ie11iac4g0bk6t2@4ax.com>  4 On Tue, 09 Apr 2002 15:17:45 +0200, Martin Knoblauch% <Martin.Knoblauch@TeraPort.de> wrote:s   >Martin Knoblauch wrote: >> e >> Larry Kilgallen wrote:  >> >   >> >F >> > Ok, I'll bite.  Why is it necessary to get it out of the dcache ? >> nK >>  because you cannot execute intructions from d-cache? :-) Similar thingsIE >> have to be done on MIPS if you want to execute stuff from the datanH >> segment. IIRC, it is enough (on MIPS) to just flush the dcache before >> trying to execute.* >> * >*I > oops. Seems that it is (on MIPS at least) the other way round. You have C >to flush the i-cache before trying to execute stuff out of memory.* >*  B In my limited following of all this, the scenario of "flushing theC i-cache" sounds like a chicken-and-egg thing.  If you can't executeOF code in the data area without flushing the i-cache, how do you execute a command to flush that cache?  1 Not speaking for anyone, certainly not DEC/Compaq - (get rid of the xxxz in my address to e-mail)u   ------------------------------  # Date: Tue, 09 Apr 2002 11:55:15 GMTm# From: "John Smith" <a@nonymous.com>a Subject: Will IBM buy Sun?G Message-ID: <DEAs8.28470$cN1.7938@news01.bloor.is.net.cable.rogers.com>   > --------------------------------------------------------------(  This story was printed from Enterprise,,  located at http://www.zdnet.com/enterprise.?  --------------------------------------------------------------l   When will IBM buy Sun?: By David Berlind, Enterprise April 8, 2002 6:43 AM PT URL:  H When I interviewed IBM's software chief Steve Mills in February, he saidI something that has stuck in my head, and I couldn't figure out why. Now I J understand: Mills was throwing down a gauntlet that could ultimately wrest control of Java from Sun.T  G In response to a question about Microsoft's monopoly, Mills defined themG "Windows franchise" as consisting of Windows and Microsoft Office. This H franchise, he said, wasn't going to repeat itself. The tone in his voiceL when he said that was not just one of pragmatism. I had the distinct feeling> that it was of also one of commitment and personal conviction.  G Besides the Windows franchise, there are a lot of things that IBM would-L rather not have repeat themselves. Not buying MS-DOS when it had the chance.J Allowing the cloning of PCs. The 32-bit OS debacle (when then-OS/2 partnerF Bill Gates basically gave IBM its walking papers). Once Microsoft's NTK dispatched OS/2 as a viable option, IBM was cast adrift in the fast-growinggK software segment of Intel-based operating systems. Adding insult to injury,fH the "Wintel" platform became so heavily commoditized that it was now tooG easy for enterprise hardware competitors like Dell and the now possiblygA merging Hewlett-Packard and Compaq to carve out their own slices.-  I Having lost three golden opportunities (one hardware and two software) todA capture a "tax" on every Intel system sold, IBM found itself in a L marginalized position quite uncharacteristic of the world's largest computerI company, and beholden to the same company that once jilted it--Microsoft.s  I When Mills looked me straight in the eye and said that it wouldn't happendJ again, he may have been referring to more than just the Windows franchise.K IBM is a company that prefers to control its own destiny. It knows too welleI the risks of having its fate tied to the whims of Microsoft and Intel. InuI order to regain control of its destiny and prevent history from repeating1= itself, the company will have to do some marginalizing of itsFG own--particularly to Microsoft and at the expense of Sun. To accomplish>J that, IBM must do something it failed to do with OS/2. It must provide itsC customers with a clear path from Microsoft's platforms to extremelysJ compelling alternatives based on its own intellectual property, or, at theB very least, to some open-source IP over which it exerts tremendous
 influence.     Enter Java and Web serviceseL As a tool for marginalizing Microsoft, Java is everything OS/2 was not. JavaH is the heart and soul of IBM's WebSphere (the middleware for the rest ofK IBM's software portfolio). Java is the primary software vanguard that keeps K Microsoft from penetrating the datacenter. Even better, Java represents the'H first credible threat to Redmond on the client side. By all counts, someH version of Java now exists on many more client devices than does WindowsK (including desktop systems, phones, gaming consoles, set top boxes and evensJ credit cards). Java has the all-important buzz too. 3G networks may be theJ all-the-rage at this year's CeBIT, but more important are the applicationsJ they will enable. Judging by the number of Java-enabled phones (e.g., SonyJ Ericsson's Z700 and P800), Java is far ahead of the Windows alternative (aH variant of Windows CE). Finally, Java is winning the hearts and minds ofD developers. According to the Hurwitz Group, 35 percent of developersL surveyed plan to deploy Web services on Java vs. 15 percent for .Net, and 31 percent plan to use both.   G Java is the golden egg and Sun is the goose that laid it. To truly keepAJ history from repeating itself (especially on the server-side, where IBM isH somewhat beholden to Sun), IBM has few choices. It can take ownership ofI Java and, therefore, Sun as well. IBM can try to dislodge Java from Sun'seJ control and move it into the open source community where IBM might not ownK it, but could pretty much have its way with it. When asked about the desireeK to own Java, IBM's Director for eBusiness Standards Strategy Bob Sutor saidlJ "I don't know about owning it, but we'd sure like to see it open sourced."     Eclipse of the SunI If IBM did manage to bring Java into its intellectual property portfolio, G some of IBM and Steve Mill's biggest competitors--BEA, Oracle, ComputersD Associates--would be beholden to IBM in the same way that IBM is nowL beholden to Microsoft. In addition to new revenue streams from the licensingJ fees generated by Java, this is exactly the position in which IBM needs to be.d  L However, getting there (if IBM chooses to do so) would be a different story.H The words "hostile takeover" would be an understated way to describe anyK such move by IBM. The only thing golf lover and Sun CEO Scott McNealy wouldBC probably like to see on a tee more than golf balls and the heads ofrH Microsoft executives is the head of any IBM executive. Especially Mills.  K While Mills and his software division have probably done more for Java thaniH any other company besides Sun itself, IBM's Java-related activities haveK grown increasingly self-serving. Most notable is Big Blue's leadership rolelK behind Eclipse. Bearing a name that appears to foreshadow the coupe IBM may C have in mind, Eclipse is a specification for adapting heterogeneousoI application development tools into an integrated development environment.   F Prior to Eclipse, NetBeans was the specification for this sort of toolL integration. NetBeans continues to get support from most members of the JavaD constituency, including Sun. Now, Sun's VP for Java Richard Green isH claiming that Eclipse violates the tenants of the Java Community ProcessI (JCP) as well as the integrity of the write-once, run anywhere promise oftJ Java. IBM's top cheerleader for Eclipse, Scott Hebner, claims that EclipseJ does nothing of the sort. Experts I've spoken with understand the need forI both specifications, but do not disagree with Green's overall assessment.D  C And then there's the name: Eclipse. Choosing the word "eclipse" forFJ something that was bound to cause friction in the Java world isn't exactlyG an olive branch to a company named Sun. IBM's Sutor claims the name wastH derived from the code names of Eclipse's predecessors--Lunar Eclipse andH Solar Eclipse. "The name is purely coincidental," Sutor says. "I have noL idea why those names were picked." Even so, the fact that IBM is the drivingI force behind Eclipse now makes the name appear more than coincidental. ItcI could very well be the first of several visible moves to wrest control ofED Java away from Sun. Meanwhile, Sun claims that the JCP, and not Sun,- controls the fate of Java. But IBM disagrees.a     Migrating Microsoft customerseF Taking control of Java isn't the only thing IBM must do to marginalizeL Microsoft. IBM must make it easier for Microsoft's customers to migrate fromJ Windows to Java. If Eclipse opens that door, then IBM's collaboration withK Microsoft to form the Web Services Interoperability Organization (WS-I) andyJ the interoperability standards it supports knocks that door down. As ZDNetH Tech Update Eric Knorr points out, the circumstances under which Sun wasE issued a last minute invitation to join the WS-I as a contributor (asVK opposed to board member) reek of hardball politics. Should Sun join withouttK a board seat, the company's plans for Java could be dictated in part by theaL companies that license it: IBM, BEA, and Oracle. All are WS-I board members. Advantage: IBM.p  L Microsoft is probably happy to participate in anything that sticks it to SunJ the way the WS-I could. But I wonder if Redmond sees IBM's interest in theL WS-I the way I see it-as an interoperability standard that will allow IBM toH marginalize Microsoft by easily migrating Microsoft's customers to IBM'sJ Java-based solutions. Should control of Java, through purchase or donationK to the open-source community, ever be wrested from Sun, IBM would surely beo6 in the catbird seat. Talk about a deal with the devil.  J While IBM does what it can to keep Java's star shining brightly on its ownK terms, the company is also doing everything it can to dismantle the rest ofyL Sun. The two companies are engaged in a bloodbath on the hardware front thatG has resulted in an increased share of the server market for IBM, at thedI expense of Sun, HP, and Compaq. (Sun vociferously debates the accuracy ofiJ IDC's figures.) The now merging HP and Compaq have each announced plans toK discontinue their RISC technologies (PA-RISC and Alpha, respectively). IBM,<H however, is going full steam ahead with the PowerPC microprocessors thatG power the newest AIX-based enemy to Sun's Solaris and SPARC processors:  Regatta.  L Overall, the results for Sun, which is also embattled by Compaq, HP, and nowK Dell, are reflected in its stock price. As of this writing, Sun, which oncet= traded in the 60s, was dangerously close to a three-year low.p  K Should Sun's $28 billion market capitalization decline any further, there'soK only one company that could absorb Sun with pocket change: IBM. In additioneL to putting IBM in control of its own destiny and giving it the IP and marketK presence needed to combat Microsoft, an IBM takeover of Sun would also turn K the microprocessor battle against Intel into a two-horse race. If IBM makesf its move, when will that be?   ------------------------------  $ Date: Tue, 9 Apr 2002 16:20:55 +00002 From: John Eisenschmidt <jweisen@eisenschmidt.org> Subject: Re: Will IBM buy Sun?4 Message-ID: <20020409162055.B21225@eisenschmidt.org>   --SkvwRMAIpAhPCcCJ* Content-Type: text/plain; charset=us-ascii Content-Disposition: inline-+ Content-Transfer-Encoding: quoted-printable1  L You know, just because Ziff Davis became the media giant it is because of t=3 he PC doesn't mean the world revoles around PCs.=20g  L Yes IBM does well with the Thinkpad division, and yes I'm sure there are so=L ur grapes over OS/2, but do you think anyone is crying that they're not sel=L ling PCs at a profit of 6 cents per machine? They own Lexmark! They own Lot=L us! They make a fortune selling AS/400s and RS/6000s and Z/90s (if that's w=! hat they're called this week).=20r  L There is a small tug of war over Java, no denying that, but why would IBM b=L uy Sun other than for their customers? They are two completely different co=L mpanies in mindset and direction. You think HP and Compaq will be a difficu=
 lt merger?=20a  L There are also Sun's partners to consider. Larry Ellison is not going to li=L ke it if Sun buys IBM, since Oracle ties itself so closely to Java these da=L ys, and IBM just bought Informix. I would rather see Oracle and Sun merge a= nd split the software division.s  L Interesting conjecture on the part of the author, but I think it's pretty u= nlikely.  B Unless the Voices are Mistaken, John Smith (a@nonymous.com) Wrote:@ > --------------------------------------------------------------* >  This story was printed from Enterprise,. >  located at http://www.zdnet.com/enterprise.A >  --------------------------------------------------------------c >=20 > When will IBM buy Sun?< > By David Berlind, Enterprise April 8, 2002 6:43 AM PT URL: >=20J > When I interviewed IBM's software chief Steve Mills in February, he saidK > something that has stuck in my head, and I couldn't figure out why. Now IhL > understand: Mills was throwing down a gauntlet that could ultimately wrest > control of Java from Sun.c >=20I > In response to a question about Microsoft's monopoly, Mills defined therI > "Windows franchise" as consisting of Windows and Microsoft Office. ThisoJ > franchise, he said, wasn't going to repeat itself. The tone in his voiceL > when he said that was not just one of pragmatism. I had the distinct feel= ing @ > that it was of also one of commitment and personal conviction. >=20I > Besides the Windows franchise, there are a lot of things that IBM wouldVL > rather not have repeat themselves. Not buying MS-DOS when it had the chan= ce.rL > Allowing the cloning of PCs. The 32-bit OS debacle (when then-OS/2 partnerH > Bill Gates basically gave IBM its walking papers). Once Microsoft's NTL > dispatched OS/2 as a viable option, IBM was cast adrift in the fast-growi= ngL > software segment of Intel-based operating systems. Adding insult to injur= y,J > the "Wintel" platform became so heavily commoditized that it was now tooI > easy for enterprise hardware competitors like Dell and the now possiblyiC > merging Hewlett-Packard and Compaq to carve out their own slices.m >=20K > Having lost three golden opportunities (one hardware and two software) tofC > capture a "tax" on every Intel system sold, IBM found itself in a L > marginalized position quite uncharacteristic of the world's largest compu= ter-K > company, and beholden to the same company that once jilted it--Microsoft.t >=20K > When Mills looked me straight in the eye and said that it wouldn't happen L > again, he may have been referring to more than just the Windows franchise.L > IBM is a company that prefers to control its own destiny. It knows too we= llK > the risks of having its fate tied to the whims of Microsoft and Intel. InrK > order to regain control of its destiny and prevent history from repeatingg? > itself, the company will have to do some marginalizing of itseI > own--particularly to Microsoft and at the expense of Sun. To accomplishnL > that, IBM must do something it failed to do with OS/2. It must provide itsE > customers with a clear path from Microsoft's platforms to extremelysL > compelling alternatives based on its own intellectual property, or, at theD > very least, to some open-source IP over which it exerts tremendous > influence. >=20 >=20 > Enter Java and Web servicesrL > As a tool for marginalizing Microsoft, Java is everything OS/2 was not. J= avacJ > is the heart and soul of IBM's WebSphere (the middleware for the rest ofL > IBM's software portfolio). Java is the primary software vanguard that kee= psL > Microsoft from penetrating the datacenter. Even better, Java represents t= heJ > first credible threat to Redmond on the client side. By all counts, someJ > version of Java now exists on many more client devices than does WindowsL > (including desktop systems, phones, gaming consoles, set top boxes and ev= enL > credit cards). Java has the all-important buzz too. 3G networks may be theL > all-the-rage at this year's CeBIT, but more important are the applicationsL > they will enable. Judging by the number of Java-enabled phones (e.g., SonyL > Ericsson's Z700 and P800), Java is far ahead of the Windows alternative (aJ > variant of Windows CE). Finally, Java is winning the hearts and minds ofF > developers. According to the Hurwitz Group, 35 percent of developersL > surveyed plan to deploy Web services on Java vs. 15 percent for .Net, and=  31: > percent plan to use both.M >=20I > Java is the golden egg and Sun is the goose that laid it. To truly keepKL > history from repeating itself (especially on the server-side, where IBM isJ > somewhat beholden to Sun), IBM has few choices. It can take ownership ofK > Java and, therefore, Sun as well. IBM can try to dislodge Java from Sun'sxL > control and move it into the open source community where IBM might not ownL > it, but could pretty much have its way with it. When asked about the desi= reL > to own Java, IBM's Director for eBusiness Standards Strategy Bob Sutor sa= idL > "I don't know about owning it, but we'd sure like to see it open sourced." >=20 >=20 > Eclipse of the SunK > If IBM did manage to bring Java into its intellectual property portfolio,iI > some of IBM and Steve Mill's biggest competitors--BEA, Oracle, Computer F > Associates--would be beholden to IBM in the same way that IBM is nowL > beholden to Microsoft. In addition to new revenue streams from the licens= ing-L > fees generated by Java, this is exactly the position in which IBM needs to > be.c >=20L > However, getting there (if IBM chooses to do so) would be a different sto= ry.cJ > The words "hostile takeover" would be an understated way to describe anyL > such move by IBM. The only thing golf lover and Sun CEO Scott McNealy wou= ldE > probably like to see on a tee more than golf balls and the heads ofhJ > Microsoft executives is the head of any IBM executive. Especially Mills. >=20L > While Mills and his software division have probably done more for Java th= anJ > any other company besides Sun itself, IBM's Java-related activities haveL > grown increasingly self-serving. Most notable is Big Blue's leadership ro= leL > behind Eclipse. Bearing a name that appears to foreshadow the coupe IBM m= ayE > have in mind, Eclipse is a specification for adapting heterogeneousfK > application development tools into an integrated development environment.s >=20H > Prior to Eclipse, NetBeans was the specification for this sort of toolL > integration. NetBeans continues to get support from most members of the J= ava F > constituency, including Sun. Now, Sun's VP for Java Richard Green isJ > claiming that Eclipse violates the tenants of the Java Community ProcessK > (JCP) as well as the integrity of the write-once, run anywhere promise ofsL > Java. IBM's top cheerleader for Eclipse, Scott Hebner, claims that EclipseL > does nothing of the sort. Experts I've spoken with understand the need forK > both specifications, but do not disagree with Green's overall assessment.a >=20E > And then there's the name: Eclipse. Choosing the word "eclipse" for L > something that was bound to cause friction in the Java world isn't exactlyI > an olive branch to a company named Sun. IBM's Sutor claims the name wasrJ > derived from the code names of Eclipse's predecessors--Lunar Eclipse andJ > Solar Eclipse. "The name is purely coincidental," Sutor says. "I have noL > idea why those names were picked." Even so, the fact that IBM is the driv= inghK > force behind Eclipse now makes the name appear more than coincidental. ItdK > could very well be the first of several visible moves to wrest control ofnF > Java away from Sun. Meanwhile, Sun claims that the JCP, and not Sun,/ > controls the fate of Java. But IBM disagrees.h >=20 >=20 > Migrating Microsoft customers H > Taking control of Java isn't the only thing IBM must do to marginalizeL > Microsoft. IBM must make it easier for Microsoft's customers to migrate f= romwL > Windows to Java. If Eclipse opens that door, then IBM's collaboration withL > Microsoft to form the Web Services Interoperability Organization (WS-I) a= ndL > the interoperability standards it supports knocks that door down. As ZDNetJ > Tech Update Eric Knorr points out, the circumstances under which Sun wasG > issued a last minute invitation to join the WS-I as a contributor (asaL > opposed to board member) reek of hardball politics. Should Sun join witho= utL > a board seat, the company's plans for Java could be dictated in part by t= heL > companies that license it: IBM, BEA, and Oracle. All are WS-I board membe= rs.  > Advantage: IBM.n >=20L > Microsoft is probably happy to participate in anything that sticks it to = SunoL > the way the WS-I could. But I wonder if Redmond sees IBM's interest in theL > WS-I the way I see it-as an interoperability standard that will allow IBM=  tocJ > marginalize Microsoft by easily migrating Microsoft's customers to IBM'sL > Java-based solutions. Should control of Java, through purchase or donationL > to the open-source community, ever be wrested from Sun, IBM would surely = be8 > in the catbird seat. Talk about a deal with the devil. >=20L > While IBM does what it can to keep Java's star shining brightly on its ownL > terms, the company is also doing everything it can to dismantle the rest = ofL > Sun. The two companies are engaged in a bloodbath on the hardware front t= hatoI > has resulted in an increased share of the server market for IBM, at thenK > expense of Sun, HP, and Compaq. (Sun vociferously debates the accuracy ofpL > IDC's figures.) The now merging HP and Compaq have each announced plans toL > discontinue their RISC technologies (PA-RISC and Alpha, respectively). IB= M,J > however, is going full steam ahead with the PowerPC microprocessors thatI > power the newest AIX-based enemy to Sun's Solaris and SPARC processors: 
 > Regatta. >=20L > Overall, the results for Sun, which is also embattled by Compaq, HP, and = nowtL > Dell, are reflected in its stock price. As of this writing, Sun, which on= ce? > traded in the 60s, was dangerously close to a three-year low.i >=20L > Should Sun's $28 billion market capitalization decline any further, there= 'sL > only one company that could absorb Sun with pocket change: IBM. In additi= onL > to putting IBM in control of its own destiny and giving it the IP and mar= ketnL > presence needed to combat Microsoft, an IBM takeover of Sun would also tu= rnL > the microprocessor battle against Intel into a two-horse race. If IBM mak= es > its move, when will that be? >=20 >=20 >=20   --=20y/ John W. Eisenschmidt <jweisen@eisenschmidt.org>g6  Homepage URL    | http://www.eisenschmidt.org/jweisenL  PGP Public Key  | http://www.eisenschmidt.org/jweisen/misc/jeisenschmidt.a= scD  PGP Fingerprint | 5F9B F916 5AD1 3295 CF99 BC1E 1F97 E6A3 37E3 BEF2   --SkvwRMAIpAhPCcCJ' Content-Type: application/pgp-signaturen Content-Disposition: inlinet   -----BEGIN PGP SIGNATURE-----e Version: GnuPG v1.0.6 (OpenBSD)e* Comment: For info see http://www.gnupg.org  @ iD8DBQE8sxTnH5fmozfjvvIRAsK6AKDqn6IDEFJcav4sOHAmMXTpNBkk3wCeMOZp v1UYO1ArtVAUOw0Ipfvhbe8= =EsW4c -----END PGP SIGNATURE-----a   --SkvwRMAIpAhPCcCJ--   ------------------------------  % Date: Tue, 09 Apr 2002 08:47:36 +0200 ) From: Bart Zorn <B.Zorn@xs4all.nospam.nl>i Subject: Re: X/ Message-ID: <3CB28E88.5080303@xs4all.nospam.nl>g   Niclas Jansson wrote:eD > Hi I am new to vms and, have tried to install decwindows on my vaxI > 4000-300 and reach it from a pc with X -query foo. But when i install XuE > on the vaxen my nic dissaperad. Anyone know how to do this if it isw
 > possible >  > Niclas Jansson >   I Hey, this is a VMS newsgroup, so could you please explain what "X -query oF foo" is supposed to do? What is "X" that you installed on the VAX and E what is a "nic" which disappears? It helps if you also mention which eH version of OpenVMS you installed on the VAX (I assume it *is* VMS!) and  which version of TCP/IP.   Regards,  	 Bart Zorne   ------------------------------  % Date: Tue, 09 Apr 2002 09:03:06 +0100 - From: Roy Omond <Roy.Omond@BlueBubble.UK.Com>p- Subject: [Change topic] Re: Andrew's Literacy"1 Message-ID: <3CB2A03A.5B88E63F@BlueBubble.UK.Com>v   John Santos wrote:  ' > [... usual Andrew drivel snipped ...]B >sC > Andrew - Do you consistently and systematically misspell "misled"n > and "misgivings?"o  D Andrew consistently misspells a huge number of common English words.  . I think his "missled" is really "missiled" :-)  ( "Vunerable" is another favourite of his.  ? I somehow doubt very much if Andrew ever passed O-level EnglishgD (for non-UKans, that's one of the early exams taken around about theL age of 15 or 16;  failure at that level tends to put you on the scrap heap).% Hmmm... maybe he's really only 14 ...,  > In short, he's illiterate, but he's a damned fine SooperDooper0 System/Universe Architect of the First Order :-)  	 Roy Omondy Blue Bubble Ltd.   ------------------------------   Date: 9 Apr 2002 06:06:54 -0700u( From: bob@instantwhip.com (Bob Ceculski)1 Subject: Re: [Change topic] Re: Andrew's Literacy = Message-ID: <d7791aa1.0204090506.24b63874@posting.google.com>o  f Roy Omond <Roy.Omond@BlueBubble.UK.Com> wrote in message news:<3CB2A03A.5B88E63F@BlueBubble.UK.Com>... > John Santos wrote: > ) > > [... usual Andrew drivel snipped ...]o > >eE > > Andrew - Do you consistently and systematically misspell "misled"i > > and "misgivings?"i > F > Andrew consistently misspells a huge number of common English words. > 0 > I think his "missled" is really "missiled" :-) > * > "Vunerable" is another favourite of his. > A > I somehow doubt very much if Andrew ever passed O-level EnglishsF > (for non-UKans, that's one of the early exams taken around about theN > age of 15 or 16;  failure at that level tends to put you on the scrap heap).' > Hmmm... maybe he's really only 14 ...i > @ > In short, he's illiterate, but he's a damned fine SooperDooper2 > System/Universe Architect of the First Order :-) >  > Roy Omondm > Blue Bubble Ltd.  = that's what happens when you've been on slowaris too long ...    ------------------------------   End of INFO-VAX 2002.196 ************************