1 INFO-VAX	Sat, 26 Oct 2002	Volume 2002 : Issue 591       Contents: Re: "slow" CDE clock? 3 Re: 6 free RA81 disks near philly (for pickup only)  Re: Acrobat Reader?  Re: backup saveset Re: backup saveset0 Re: backup/image of shadowed volume misses files DS10 power consumption Re: DS10 power consumption( Re: Dual ISA ethernet adapter on VMS 7.2( Re: Dual ISA ethernet adapter on VMS 7.2 Re: eXcursion & Mozilla & RE: FC-AL support on VMS, coming soon?& RE: FC-AL support on VMS, coming soon?& Re: FC-AL support on VMS, coming soon?& Re: FC-AL support on VMS, coming soon?& Re: FC-AL support on VMS, coming soon? Re: File Attributes  Re: File Attributes  Re: File Attributes ( Re: Fiorina Gets Feisty at Gartner Event( Re: Fiorina Gets Feisty at Gartner Event( Re: Fiorina Gets Feisty at Gartner Event Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC Re: Immutable laws of the PC1 Re: It'll be interesting to see what happens here  Re: Monitor System Question , Now where's the OpenVMS OpenOffice software? Re: OpenVMS 7.3-1 and CSWING Re: OpenVMS 7.3-1 and CSWING Re: OpenVMS 7.3-1 and CSWING Re: page or pagelet  Re: page or pagelet  RE: page or pagelet  Re: page or pagelet 
 Re: PCL image $ Quorum disk appears to not be voting  Re: So I went to the HP IT forum  Re: So I went to the HP IT forum  Re: So I went to the HP IT forum  Re: So I went to the HP IT forum  Re: So I went to the HP IT forum$ Re: TCPIP$FTP_SERVER startup problem$ Re: TCPIP$FTP_SERVER startup problem Re: TPU port to Linux  TSM on the Freeware disk Re: TSM on the Freeware disk Re: TSM on the Freeware diskG Why did the VMS programmers confuse Halloween with Christmas this year? K Re: Why did the VMS programmers confuse Halloween with Christmas this year? K Re: Why did the VMS programmers confuse Halloween with Christmas this year?   F ----------------------------------------------------------------------  # Date: Sat, 26 Oct 2002 02:17:26 GMT . From: Keith Brown <kbrown2720@NOSPAMattbi.com> Subject: Re: "slow" CDE clock?- Message-ID: <WWmu9.78897$Hj7.40423@rwcrnsc53>    Phillip Helbig wrote:   J > My DECW$CLOCK seems to be in step with the system time.  The minute handH > on the CDE clock seems to move about 45 seconds too late.  Anyone else > seen this?   Mine does this too Phillip.    ------------------------------  % Date: Fri, 25 Oct 2002 20:06:12 -0400 * From: Brian Hechinger <wonko@4amlunch.net>< Subject: Re: 6 free RA81 disks near philly (for pickup only)9 Message-ID: <20021025200612.G1728947@marvin.4amlunch.net>   9 On Fri, Oct 25, 2002 at 06:23:28AM -0400, JF Mezei wrote:  > O > You do realise that you could heat your house with those , do you ? Just plug W > it in and turn on as many drives as is needed to keep it the house warm and toasty...   9 i could heat my house with the majority of what i own. ;)   M > I had a similar lot, and I gutted the drawers (much work, these things were N > built like tanks) and use one such unit as storage for all sorts of maps and4 > travel brochures etc. I have  TV set on top of it. > P > The cabinet also has a power controller which isolates some of the spikes from< > the power.  And of course, you get tons of fans in this !.  N that's an alternative of course, as you can never have too much storage space,N especially when dealing with trying to stock up as many spare parts as you canM since your favorite machines aren't made anymore, but i'd prefer they went to M someone who would use them for what they are meant to be used for.  i can buy  shelves. ;)    -brian --  ' Rich.  R-I-C-H.  No 'B', No 'T'.  -Rich    ------------------------------  % Date: Fri, 25 Oct 2002 23:51:55 -0500 ( From: Rich Jordan <duodec@speakeasy.net> Subject: Re: Acrobat Reader?1 Message-ID: <B8mdnXjyGY8zjCegXTWcpQ@giganews.com>    Chris Scheers wrote:j > Kilgallen@SpamCop.net (Larry Kilgallen) wrote in message news:<pcUNhcxcnaAw@eisner.encompasserve.org>... > [ >>In article <newscache$14x64h$f19$1@news.tiscali.fr>, DTL <Didier.Morandi@Free.fr> writes:  >>B >>Correct.  There was a supported product for VMS for a while, butD >>the (small) vendor dropped it, presumably due to lack of interest.E >>(I believe they had to make royalty payments independent of sales.)  >  >  > 7 > It's not exactly dropped, but it seems to be dormant.  > E > The problem is that there has been very little interest in actually A > paying for such a product.  The ONLY purchases were from non-US E > customers.  The US response has been "Well, if you give it away, we  > might try it."  E We tried the demo version and it worked OK, but I could not convince  I work to pay for something that they got free on peecees, and at the time  I I did not need it at home (using crappy NT on a nice Alpha to browse and  G email).  Now that my main net/browsing system is an Alpha/Mozilla box,  H it would be nice to have the option.  Depends on the price though.  I'm G not expecting everything to be "hobbyist" or dirt cheap, but it is for  ( home use and there is a budget for that.   Rich   ------------------------------  % Date: Fri, 25 Oct 2002 15:39:25 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>  Subject: Re: backup saveset $ Message-ID: <3db99dfe$1@news.si.com>  I >Thank you for your suggestion. I tried that but the size of the original / >save set is approx 303k blocks and when I copy : >it to disk it goes to over 1.8 mill. It does not stay the >same size.   - The default block size on tape is 8192 bytes.  --  A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Fri, 25 Oct 2002 20:33:15 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: backup saveset ' Message-ID: <3DB9F0DB.DBA9F86B@fsi.net>    Brian Tillman wrote: > K > >Thank you for your suggestion. I tried that but the size of the original 1 > >save set is approx 303k blocks and when I copy < > >it to disk it goes to over 1.8 mill. It does not stay the
 > >same size.  > / > The default block size on tape is 8192 bytes.   E Is that the default for BACKUP or is that stated as the default block . size in the ANSI labelled tape specifications?   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Fri, 25 Oct 2002 15:42:33 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> 9 Subject: Re: backup/image of shadowed volume misses files , Message-ID: <3DB99EA8.5108BF1C@videotron.ca>   Charlie Hammond wrote:I > My opinion is that /[NO]ALIAS is meaningless with /IMAGE and should not L > be used with /IMAGE.  I know of no cases with /IMAGE does not do the rightJ > thing if neither /ALIAS nor /NOALIEAS is included on the BACKUP command.7 > (Assuming the rest of the BACKUP command is correct!)   ( IF anyone confirm or deny the following:  4 BACKUP/IMAGE will have /ALIAS as default behaviour ?  H BACKUP/IMAGE  will backup to tape the VMS$COMMON tree twice (once as theB device:[vms$common...], and once as device:[sys0.syscommon...] ???  L BACKUP/IMAGE/NOALIAS will backup to tape the VMS$COMMON tree only once since6 one of the two is just a directory entry of the other.   ------------------------------  % Date: Fri, 25 Oct 2002 11:26:07 -0700 ' From: David Mathog <mathog@caltech.edu>  Subject: DS10 power consumption 8 Message-ID: <20021025112607.190b94ad.mathog@caltech.edu>  ' How much power does a 466 Mhz DS10 use?   C The upper limit must be less than 330 - 360W because we had 4 DS10s D plugged into a surge suppressor with a 12A breaker - and the breaker never blew.   @ These were simple DS10s, just the CPU, 512Mb RAM, floppy (idle), CDROM (idle), and IDE disk.    Thanks,    David Mathog mathog@caltech.edu> Manager, Sequence Analysis Facility, Biology Division, Caltech   ------------------------------  % Date: Fri, 25 Oct 2002 21:47:57 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)# Subject: Re: DS10 power consumption K Message-ID: <rdeininger-2510022147580001@1cust118.tnt2.nashua.nh.da.uu.net>   8 In article <20021025112607.190b94ad.mathog@caltech.edu>, mathog@caltech.edu wrote:   ( >How much power does a 466 Mhz DS10 use? > D >The upper limit must be less than 330 - 360W because we had 4 DS10sE >plugged into a surge suppressor with a 12A breaker - and the breaker  >never blew. > A >These were simple DS10s, just the CPU, 512Mb RAM, floppy (idle),  >CDROM (idle), and IDE disk.   The 466 MHz Quickspec is here:C     http://www.compaq.com/products/quickspecs/10398_na/10398_na.pdf    The 600 MHz Quickspec is here:C     http://www.compaq.com/products/quickspecs/10908_na/10908_na.pdf   I A "typical" DS10 600 MHz (or is it 466 MHz?) system is listed as having a J measured input power of 176 watts while running a system excerciser.  ThatC was with 1 GB RAM, 3 disk drives, CD, floppy, and 2 graphics cards.    ------------------------------  # Date: Sat, 26 Oct 2002 01:52:16 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network> 1 Subject: Re: Dual ISA ethernet adapter on VMS 7.2 : Message-ID: <kzmu9.1735$Zx.275561@news1.news.adelphia.net>   David J. Dachtera wrote:# > VAXman-, @SendSpamHere.ORG wrote:  > U >>In article <3DB8A3B5.A71ED8D6@sunset.net>, Tom Crabtree <tccrab@sunset.net> writes:  >> >>>TR: >> >><...snip...> >>N >>>>The reason why I ask is because ultimately I'd like to perform NAT routingO >>>>on the system to share my cable internet connection. Is this even possible? ? >>>>Has anyone ever done this and is there software to do this?  >>> E >>>OK, but routers are cheaper (and easier to set up).  You can buy a 2 >>>Linksys for $60-80.  Uses less electricity too. >>N >>...but Linksys suffers from poor firmware (many too many unprages) and lacks4 >>performance.  Try a Netopia router for such needs. > F > Netgear is good, too. I have a Netgear FR314 here. All I need now isE > broadband internet access (which is currently not available in this  > area).  G I am using a D-Link.  It usually works, occasionally I get some errors   reported by opcom.  < Something at least twice external to my network knocked the 9 router/firewall portion of it offline requiring a reboot.   H The D-Link has also sometimes required a reboot to accept a new network  configuration.  G Broadband Internet in my area is availble only from the Cable Company.  H However, if you do not need TELNET access, then Satellite access should  be available everywhere.  3 But the D-LINK unit sells for less than $50.00 U.S.   I I have also in the past used an old 486 DX 50 with two NIC cards running  H LINUX for NAT and a firewall, but then you need to pay attention to the I security issues and also to make sure you do not have an open SMTP relay.    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------  % Date: Fri, 25 Oct 2002 20:28:17 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> 1 Subject: Re: Dual ISA ethernet adapter on VMS 7.2 ' Message-ID: <3DB9EFB1.E93FC1D8@fsi.net>   ! VAXman-, @SendSpamHere.ORG wrote:  > U > In article <3DB8A3B5.A71ED8D6@sunset.net>, Tom Crabtree <tccrab@sunset.net> writes:  > >TR: > <...snip...>O > >> The reason why I ask is because ultimately I'd like to perform NAT routing P > >> on the system to share my cable internet connection. Is this even possible?@ > >> Has anyone ever done this and is there software to do this?E > >OK, but routers are cheaper (and easier to set up).  You can buy a 2 > >Linksys for $60-80.  Uses less electricity too. > N > ...but Linksys suffers from poor firmware (many too many unprages) and lacks4 > performance.  Try a Netopia router for such needs.  D Netgear is good, too. I have a Netgear FR314 here. All I need now isC broadband internet access (which is currently not available in this  area).   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Fri, 25 Oct 2002 19:50:59 GMT . From: peter@langstoeger.at (Peter LANGSTOEGER)  Subject: Re: eXcursion & Mozilla5 Message-ID: <Dghu9.154565$N_6.2299759@news.chello.at>   W In article <3DB8E3E9.1030907@tzora.co.il>, Mike Rechtman <rechtman@tzora.co.il> writes: K >> 10/16/2002 or thereabouts.  At least now it's on the download page with   >  ^^^^^^^^^^^M >I _do_ hope that means 16-Oct-2002 (16/10/2002) and not 10th of Nextober....  >for us non-USA types.  F That's why 2002/10/16, 2002-10-16 or 16-Oct-2002 is the preferred form5 for Non-US _and_ US fellows... (and is easy sortable)    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Fri, 25 Oct 2002 17:18:52 -0400 ' From: "Main, Kerry" <Kerry.Main@hp.com> / Subject: RE: FC-AL support on VMS, coming soon? T Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF402660ACF@kaoexc01.americas.cpqcorp.net>   Malcolm,  < Re: FC-AL on OpenVMS - check out latest roadmap slide 17 at:H http://www.openvms.compaq.com/openvms/roadmap/OPENVMS_ROADMAPS_files/ope nvms_roadmaps.ppt    In early 03:0 Lower cost FC External Storage Support (MSA1000)  - 2-node support Arbitrated Loop3 - Switched Support (using existing external Switch)   G Keep in mind FC-AL (on all platforms) is real, real low end performance H compared to switched fabric. Its like using 10MB thin wire ethernet withE shared wire vs 100MB switched Ethernet where each has dedicated port.   E But, for those looking at smaller low cost solutions with perhaps 1-2 $ nodes, it likely will be acceptable.   Regards   
 Kerry Main Senior Consultant  Hewlett-Packard (Canada) Co.! Consulting & Integration Services  Voice: 613-592-4660  Fax   : 613-591-4477 Email: kerryDOTmain@hpDOTcom-     (remove the DOT's and replace with "."'s)        -----Original Message-----; From: Malcolm Dunnett [mailto:nothome@spammers.are.scum]=20  Sent: October 25, 2002 12:27 PM  To: Info-VAX@Mvb.Saic.Com + Subject: FC-AL support on VMS, coming soon?       2   I've done some research since my earlier posting4 about FC support on VMS and have determined that VMS8 only supports switched fabric. This is a real impediment5 to my attempt to do FC on VMS on the cheap as I can't 4 find any cheap Compaq FC switches ( I got the RA80005 for $1000 and an HBA for $100 but the cheapest switch  I've seen is about $3000 ).   H   Searching the OpenVMS website I saw a "futures" slide suggesting FC-AL@ support coming "after Q4 2002" and a couple of surveys asking ifG customers are interested in FC-AL. I can't find any more details on the H plans though. Is there anything people are allowed to talk about yet? InC particular, would this support be available on existing hardware or E would it only to apply to some new type of adapter/controller not yet 
 announced?  1   At the moment it looks like these new toys will 2 only be of use to me on Tru64 or (ugh) Windows :-(  2   ps. To save Andrew the keystrokes I'll note that2 it appears that the RA8000/FC-AL is also supported on Solaris.    ------------------------------    Date: 25 Oct 2002 15:03:16 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett) / Subject: RE: FC-AL support on VMS, coming soon? - Message-ID: <NCumvQmdqSIr@malvm7.mala.bc.ca.>   U In article <BE56C50EA024184DAF48F0B9A47F5CF402660ACF@kaoexc01.americas.cpqcorp.net>,  -     "Main, Kerry" <Kerry.Main@hp.com> writes:  > > > Re: FC-AL on OpenVMS - check out latest roadmap slide 17 at:J > http://www.openvms.compaq.com/openvms/roadmap/OPENVMS_ROADMAPS_files/ope > nvms_roadmaps.ppt  >  > In early 03:2 > Lower cost FC External Storage Support (MSA1000)" > - 2-node support Arbitrated Loop5 > - Switched Support (using existing external Switch)  > J     Thanks for the pointer. It isn't clear from the slide though if that'sG happening in early 03 or juse sometime before 2005. It also isn't clear H if one will need an MSA1000 for FC-AL support or if it will be supported
 on HSG80s.  I > Keep in mind FC-AL (on all platforms) is real, real low end performance J > compared to switched fabric. Its like using 10MB thin wire ethernet withG > shared wire vs 100MB switched Ethernet where each has dedicated port.  >   G    I'm only looking for one ( or maybe 2 ) Alpha and one ( or maybe 2 ) D HSG80 maximum, so I wouldn't think performance would be that much ofI an issue. Currently I've got an RA7000, which tops out at around 16MB/sec 5 so anything better than that would be an improvement.    ------------------------------  # Date: Sat, 26 Oct 2002 00:04:32 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network> / Subject: Re: FC-AL support on VMS, coming soon? : Message-ID: <k_ku9.1540$Zx.262417@news1.news.adelphia.net>   Malcolm Dunnett wrote:W > In article <BE56C50EA024184DAF48F0B9A47F5CF402660ACF@kaoexc01.americas.cpqcorp.net>,  / >     "Main, Kerry" <Kerry.Main@hp.com> writes:  > > >>Re: FC-AL on OpenVMS - check out latest roadmap slide 17 at:J >>http://www.openvms.compaq.com/openvms/roadmap/OPENVMS_ROADMAPS_files/ope >>nvms_roadmaps.ppt  >> >>In early 03:2 >>Lower cost FC External Storage Support (MSA1000)" >>- 2-node support Arbitrated Loop5 >>- Switched Support (using existing external Switch)  >>L >     Thanks for the pointer. It isn't clear from the slide though if that'sI > happening in early 03 or juse sometime before 2005. It also isn't clear J > if one will need an MSA1000 for FC-AL support or if it will be supported > on HSG80s.  C The FC-AL driver and the MSA1000 is still in test and qualifcation  B phase.  If all goes well, it should be available well before 2005.  + What type of time frame are you looking at?   I I am not sure on the support issue for the HSG80s.  The HSG80s currently  & support FC-AL, so it should just work.  A The testing right now is concentrating on the MSA1000s.  However  I MSA1000s were not available at the time the driver was under development.   I >>Keep in mind FC-AL (on all platforms) is real, real low end performance J >>compared to switched fabric. Its like using 10MB thin wire ethernet withG >>shared wire vs 100MB switched Ethernet where each has dedicated port.  >>  D More like 100MB shared half duplex as opposed to a switch.  Low end 8 configurations probably would not notice the difference.  J The big limitation of FC-AL is that it does not scale as well as a Fabric.  I >    I'm only looking for one ( or maybe 2 ) Alpha and one ( or maybe 2 ) F > HSG80 maximum, so I wouldn't think performance would be that much ofK > an issue. Currently I've got an RA7000, which tops out at around 16MB/sec 7 > so anything better than that would be an improvement.   @ FC-AL support will not be available for OpenVMS 7.3 and earlier.  > And expect support to be limited to two node low end clusters.  ? The prices of FC switches also appears to be dropping, so that  B eventually there will no longer be as big of price difference for 0 arbitrated loop hubs and Fibre Channel switches.  H One of the options expected on the MSA1000 is a two node FC-AL hub that 7 will allow you to directly connect two OpenVMS systems.   2 And the MSA1000 will work with 2 Gb Fibre Channel.   -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------    Date: 25 Oct 2002 20:37:33 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett) / Subject: Re: FC-AL support on VMS, coming soon? - Message-ID: <BUzXaONnDWMf@malvm7.mala.bc.ca.>   ; In article <k_ku9.1540$Zx.262417@news1.news.adelphia.net>,  3     "John E. Malmberg" <wb8tyw@qsl.network> writes:   E > The FC-AL driver and the MSA1000 is still in test and qualifcation  D > phase.  If all goes well, it should be available well before 2005. > - > What type of time frame are you looking at?  >   7     We'll I'd like to start experimenting next week :-)   ?     I know that's not going to happen - for the moment I think  > I'll build a Tru64 system to play around with the fibrechannel gear.   @      We don't have any "production" needs for FC yet, but one ofD the objectives is to assess its suitability for our environment - to" do that properly I need it on VMS.  B      Kerry suggested the stuff might be available early in 03 - isC that realistic? ( I assume it would be a patch kit or point release  beyond VMS 7.3-1 ).    > F > More like 100MB shared half duplex as opposed to a switch.  Low end : > configurations probably would not notice the difference. > @     That was my sense of it. It's got to be significantly better than FWD (20MB/sec) SCSI.    > @ > And expect support to be limited to two node low end clusters. >   E    That would be fine - I wouldn't expect to initially build anything A bigger than a 2 node FC cluster. I envision using FC for disaster A tolerance, eg building a cluster with a DS20 and an RA8000 in one A building and another DS20 and RA8000 in another building. I could C then use HBVS ( or DRM on the RA8000 ) to replicate the data. There ? would also be some satellite nodes in the cluster communicating 
 via ethernet.   A > The prices of FC switches also appears to be dropping, so that  D > eventually there will no longer be as big of price difference for 2 > arbitrated loop hubs and Fibre Channel switches. > E     That would be good. Hubs are pretty cheap already on eBay though. C I often see 12 port Compaq hubs ( without GBICS ) for $200-300. The * best I've seen on a switch is about $3000.   ------------------------------  # Date: Sat, 26 Oct 2002 04:09:07 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network> / Subject: Re: FC-AL support on VMS, coming soon? : Message-ID: <Dzou9.1967$Zx.295649@news1.news.adelphia.net>   Malcolm Dunnett wrote:= > In article <k_ku9.1540$Zx.262417@news1.news.adelphia.net>,  5 >     "John E. Malmberg" <wb8tyw@qsl.network> writes:  > E >>The FC-AL driver and the MSA1000 is still in test and qualifcation  D >>phase.  If all goes well, it should be available well before 2005. >>- >>What type of time frame are you looking at?  >>9 >     We'll I'd like to start experimenting next week :-)    Taking the weekend off? :-)   A >     I know that's not going to happen - for the moment I think  @ > I'll build a Tru64 system to play around with the fibrechannel > gear.   H While official support is not there yet, if you want to experiment on a - test system, you can try it on OpenVMS 7.3-1.   B >      We don't have any "production" needs for FC yet, but one ofF > the objectives is to assess its suitability for our environment - to$ > do that properly I need it on VMS.  C FC-AL is "Latent" in 7.3-1.  That is not a support commitment, but   should allow you to do testing.   I Neither the SRM console or the driver will autonegotiate, so you need to  H follow the instructions for the wwidmgr instructions to put the adapter  into loop mode.   D >      Kerry suggested the stuff might be available early in 03 - isE > that realistic? ( I assume it would be a patch kit or point release  > beyond VMS 7.3-1 ).   C There will likely be a TIMA kit for various Fibre Channel and SCSI  2 issues before the official release of the MSA1000.  I The MSA1000 is the major driver for the FC-AL support, so I would expect  I that official support for both would be officially announced at the same   time.    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------    Date: 25 Oct 2002 11:03:40 -0700, From: colive@technologEase.com (Chris Olive) Subject: Re: File Attributes= Message-ID: <b10654c6.0210251003.21ea13c0@posting.google.com>   h Atlant Schmidt <atlantnospam@mindspring.com> wrote in message news:<3DB8134A.B784960B@mindspring.com>...! > briggs@encompasserve.org wrote:  > E > > Yes, we have the equivalent of system().  And like system(), that E > > call is limited by the ability of the caller to access the output . > > stream of the resulting forked subprocess. > > I > >  call lib$spawn ( 'WRITE SYS$OUTPUT F$FILE("your.file.name","RDT")' )  > > I > > Now all you need to do is harvest the revision date and time from the 1 > > output stream of that created subprocess. ...  > 0 > So there's no equivalent of system()'s brother. > (wrapper routine) popen(), which makes quick* > work of the output processing (and makes > `backticks` possible)? >  > Atlant  E Aside from calling the DEC C popen(), implementing this in FORTRAN (a # popen wrapper) would be as easy as:   , call SYS$CREMBX( ...with a mailbox name... )D call LIB$SPAWN( ...with SYS$INPUT = NL:, SYS$OUTPUT = mailbox above,3 with command-to-execute, and using NOWAIT flag... )   ? then open the mailbox for input using normal FORTRAN read/write . commands, and read from the mailbox until EOF.  F I'd opt for this over calling DEC C popen() from FORTRAN.  The routine& would take all of 10 minutes to write.  B And the fact that VMS 3GLs have no "popen()" or "system()" propersD doesn't speak to VMS's inferiority (if that was at all implied), but: rather the ability in very little code to produce the sameF functionality and in very little time speaks to it's SUPERIORITY.  8-)   Chris  -----  Chris Olive  colive(at)technologEase(dot)com    ------------------------------  % Date: Fri, 25 Oct 2002 14:30:27 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>  Subject: Re: File Attributes$ Message-ID: <3db98dd4$1@news.si.com>  D >I can get to the file attributes (i.e. "Revision Date") through theD >Lexical f$file_attributes within a command procedure.  But how do I7 >get to the same information through a Fortran program?   L Here's excerpts of some code I have.  Change ATR$S_EXPDATE and ATR$C_EXPDATEJ to ATR$S_REVDATE and ATR$C_REVDATE.  Of course, all status values from any calls should be checked.  K The excerpt comes from a program I have that checks for files that have not L been accessed for a while, using retention dates on a disk.  Here's a sample output:   % $ expired a301_disk:[tillman...]*.*;* I %EXPIRED-I-STATISTICS, expired files matching A301_DISK:[TILLMAN...]*.*;* I           1087 files (      270520 blocks) not referenced for at least 90  days. J            604 files (      232426 blocks) not referenced for at least 180 days. J            197 files (       42641 blocks) not referenced for at least 270 days. J            192 files (       31899 blocks) not referenced for at least 360 days. <             12 files (         966 blocks) never referenced.   Here's the excerpt:  ...          INCLUDE '($atrdef)'          INCLUDE '($fabdef)'          INCLUDE '($fibdef)'          INCLUDE '($namdef)'  ...           RECORD /atrdef/ atr( 3 )         RECORD /fabdef/ fab          RECORD /fibdef/ fib          RECORD /namdef/ nam  ... !         fab.fab$b_bid = FAB$C_BID !         fab.fab$b_bln = FAB$C_BLN !         fab.fab$l_fop = FAB$M_NAM #         fab.fab$l_nam = %loc( nam ) (         fab.fab$l_fna = %loc( filespec )         fab.fab$b_fns = speclen !         nam.nam$b_bid = NAM$C_BID !         nam.nam$b_bln = NAM$C_BLN &         nam.nam$b_rss = len( curfile )'         nam.nam$l_rsa = %loc( curfile ) (         nam.nam$b_ess = len( expanfile ))         nam.nam$l_esa = %loc( expanfile )   !         status = sys$parse( fab )  ... F                     fib.fib$l_acctl = FIB$M_NOLOCK .OR. FIB$M_NORECORD9                     fib.fib$w_fid_num = nam.nam$w_fid_num 9                     fib.fib$w_fid_seq = nam.nam$w_fid_seq 9                     fib.fib$w_fid_rvn = nam.nam$w_fid_rvn /                     fibdesc( 1 ) = FIB$C_LENGTH .                     fibdesc( 2 ) = %loc( fib )7                     atr( 1 ).atr$w_size = ATR$S_EXPDATE 7                     atr( 1 ).atr$w_type = ATR$C_EXPDATE 8                     atr( 1 ).atr$l_addr = %loc( expint )7                     atr( 2 ).atr$w_size = ATR$S_STATBLK 7                     atr( 2 ).atr$w_type = ATR$C_STATBLK 5                     atr( 2 ).atr$l_addr = %loc( sbk )   , C                   Obtain the desired info.  ?                     status = sys$qiow( , %val( channel ), %val( 
 IO$_ACCESS ), F         1                              iosb, , , fibdesc, , , , atr, )   --A Brian Tillman                   Internet: tillman_brian at si.com A Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------    Date: 25 Oct 2002 14:37:48 -0700, From: JimStrehlow@data911.com (Jim Strehlow) Subject: Re: File Attributes= Message-ID: <4b6ec350.0210251337.6fef44d9@posting.google.com>   = Atlant Schmidt <atlantnospam@mindspring.com> wrote in message  ... ( > Look at the FABs and RABs? Good golly!1 > No wonder you VMS folks are paid the big bucks! / > You work very, very hard for simple data! :-)  > 0 > If this question of file attributes isn't on a* > performance-critical path, maybe it'd be( > better to spawn something simple (DCL,% > Perl, whatever) and use *IT* to get  > the data for you.  >  ...  > Atlant    @ What is your business process that requires your need for such a fortran program?  A Most business applications have specifications for the desired or  required file types etc./ I usually just $ANALYZE/RMS/FDL  inputFileSpec,  see what type of file it is,9 design my application program to open the file as needed,  and write a new file as needed. C Third party applications and products need to handle FABs and RABs.   D Are you trying to write some utility program for the university that anyone else will use? A Are you writing the program for an educational challenge or class  assignment? E You might find $ANALYZE, $BACKUP, $COPY, $CONVERT, $SORT etc. OpenVMS A utility programs that can help solve a business problem without a ( special Fortran program "in the middle".  @ You may have a legitimate need for such a Fortran program; but I= thought I would just ask. I have often found people trying to C architect a bridge to cross a stream when all one needs is a sturdy 3 board longer than the stream is wide to get across. = Jim Strehlow, OpenVMS Systems Manager (17 years with OpenVMS)    ------------------------------  # Date: Fri, 25 Oct 2002 18:27:48 GMT - From: bdc@world.std.com (Brian 'Jarai' Chase) 1 Subject: Re: Fiorina Gets Feisty at Gartner Event & Message-ID: <H4JvAC.Kxr@world.std.com>  H In article <RZZt9.62510$Q3S.43404@news01.bloor.is.net.cable.rogers.com>," John Smith <a@nonymous.com> wrote:  : > http://www.baselinemag.com/article2/0,3959,605377,00.asp >  > October 8, 2002 & > Fiorina Gets Feisty at Gartner Event > By  Tom Steinert-Threlkeld > I > LAKE BUENA VISTA, Fla.-Merger accomplished, Carly Fiorina is coming out 	 > feisty.  >  > [...]  > I > In the long run, according to Fiorina, Sun could not keep up with HP by C > focusing on fast machinery, using a proprietary operating system, L > Solaris.*(2)  She said Sun's chief executive, Scott McNealy, was running aN > "stand-alone hot box company," as was EMC, the storage server market leader.K > Customers are no longer won over by speed of servers or desktop machines,  > she said.  > < > "In our judgment, those days are gone forever,'' she said.  H I find this statement very inconsistent with HP embracing the Itanium...E or at least with them embracing IA-64 as the next, great, fast CPU.   H And I really can't imagine HP trying to market their servers and desktopJ machines with an approach of downplaying the performance of those systems.  C Another thing I find amusing (and pathetic) is all this talk of the J competitors' proprietary operating systems.  Umm, the only non-proprietaryC OS that has any sort of market share is Linux.  And from what I can G see, IBM seems to be one of the single biggest commercial supporters of G Linux <http://www.ibm.com/linux/>. Sun is supporting and offering it on G their low-end servers <http://wwws.sun.com/software/linux/index.html>.  ( And then Dell is another Linux supporter< <http://www.dell.com/us/en/esg/topics/linux_linuxhome.htm>, ? as is HP <http://www.hp.com/wwsolutions/linux/about_linux_hp/>.e  C I guess I don't really see the differences between any of the majoriI players in the regard of all of them supporting non-proprietary operatingrE systems; though it's my opinion that IBM has been working harder thanh anyone else at it.  J And even beyond that--I see nothing at all wrong with proprietary hardwareG or software.  It's ludicrous that a company like HP would even consider J pointing to this as a flaw in their competitors, when they themselves rely$ so heavily on proprietary solutions.   -brian.e -- eF --- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----+ Font-o-Meter!      Proportional  Monospacedi'                                       ^o   ------------------------------  # Date: Fri, 25 Oct 2002 20:41:55 GMT $ From: winston@SSRL.SLAC.STANFORD.EDU1 Subject: Re: Fiorina Gets Feisty at Gartner Event 8 Message-ID: <00A15FB4.BF05080F@SSRL04.SLAC.STANFORD.EDU>  b In article <20021025084541.080f3273.mathog@caltech.edu>, David Mathog <mathog@caltech.edu> writes:! >On Thu, 24 Oct 2002 21:53:53 GMTo% >"John Smith" <a@nonymous.com> wrote:  > L >> "'Dude, you're getting a Lexmark' just doesn't have the same ring,'' saidI >> Fiorina, "which is maybe why Michael Dell has decided to private-label  >> printers.'' >nE >Which isn't all that different than what HP does for laser printers.e@ >According the gospel of Carly we should be dissing LaserJets as >rebadged Canon's.  M Last I looked (which was years ago) HP wasn't slapping an HP label on a wholefI printer made by Canon; they were adding a whole bunch of stuff to a CanoneG print engine.  That stuff included basically all the software/firmware,ID controls, readouts, etc.  So I don't think that's a fair comparison.  D Your points about printer cartridges (snipped below) are well-taken.   -- AlanS   ------------------------------   Date: 26 Oct 2002 05:00:22 GMT- From: djweath@attglobal.net (Dave Weatherall)M1 Subject: Re: Fiorina Gets Feisty at Gartner Event 5 Message-ID: <DTiotGxQ0bj6-pn2-w3qem4stEBdO@localhost>e  # Reading this thread  made me wonder-  F Has the downturn in the (IT) economy been a good thing or a bad thing  for the merger?:  7 Has it masked any fundamental flaws or made them worse?    Sugar! OT again - sorry.   -- s Cheers - Dave.   ------------------------------    Date: 25 Oct 2002 13:06:16 -0500+ From: young_r@encompasserve.org (Rob Young)T% Subject: Re: Immutable laws of the PCs3 Message-ID: <WGXDz2rSz9Yo@eisner.encompasserve.org>e  a In article <H9+J9jWyMU29@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:2c > In article <kGbNNgX+hCte@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:y >  > 	Correction... >  >>  @ >> 	So Opteron will be selling against whatever Xeon is shipping4 >> 	in June 2003 with 533 MHz FSB, that's one issue. >  > 	Make that a 667 MHz FSB:u > * > http://www.theinquirer.net/?article=5908 > Q > The monster with a huge 6 MB on-chip cache is expected to reach an unbelievable O > 477 million transistors, 8 times more than the first Pentium 4. Next year the,P > Pentium 4 and Xeon processors will be updated to reach the 0.09 micron processL > - the codenames are "Prescott" and "Nocona". The new processors are, amongO > other new features, expected to run at between 3.2 and 4 GHz in the first onecK > year of deployment, and use a sped-up 667 MHz front-side bus (you'll need 4 > dual-channel DDR333 memory to feed that one well). >   
 Correction...e  ! 	That 667 MHz FSB is a 2004 part:w  ( http://www.theinquirer.net/?article=4574  K Nocona will first be available on the E7501 chipset and have a 533MHz front'N side bus, but in the early part of 2004 that will change with the introductionK of the "Lindenhurst" chipset, which will have a 666MHz front side bus, come,3 with Dual Northway and fit into the Dobson chassis.    ---r  ? 	Now you know why the Armani clad suits get paid so much.  They. 	don't make mistakes.b   				Rob   M "The real point against pacifism is that is is not a cause at all, but only agO  weakening of all causes.  It does not announce any aim; it only announces thatcN  it will never use certain means in pursuing any aim.  It does not define its K  goal; it only defines a stopping place, beyond which nobody must go in ther  search for any goal."%                 		-- C.K. Chesterton ,   ------------------------------  % Date: Fri, 25 Oct 2002 18:22:01 +0100 ' From: Andrew Harrison SUNUK Consultancy % Subject: Re: Immutable laws of the PCs. Message-ID: <3DB97DB9.4010409@nospamn.sun.com>   Rob Young wrote:d > In article <20021025081223.1b64f794.mathog@caltech.edu>, David Mathog <mathog@caltech.edu> writes: >  > P >>Intel (and HP?) must be praying AMD runs out of money before Opteron launches. >> >  > C > 	Opteron is "shipping at the end of the first quarter 2003", this C > 	means that realistically systems show up May/June 2003 (assuming C > 	reasonable build time and qual and Opteron really doesn't slip).  > ? > 	So Opteron will be selling against whatever Xeon is shipping,? > 	in June 2003 with 533 MHz FSB, that's one issue.  The secondoC > 	issue is Intel Marketing.  A savvy Intel will be selling futuresb9 > 	in that timeframe for CPUs shipping in late 2003, i.e. 
 > 	Montecito:- > 6 > http://www.anandtech.com/cpu/showdoc.html?i=1677&p=2 > K > If Intel's desktop chips will get a 1MB L2 cache, what about their serverbI > solutions? It has been well known that Intel's Madison processor (third Q > generation derivative of Itanium) would have a massive 6MB on-die L3 cache on aoM > 0.13-micron process, but armed with Intel's 90nm process the Montecito coreeQ > (fourth generation Itanium) will boast an incredible 12MB on-die L3 cache usinga1 > close to 800 million of these 90nm transistors.o >  > ? > 	Opteron is a wonderful CPU, but certainly won't compete with27 > 	Itanium and a 800 million transistor/12 MByte cache.n >   < A reasonable rule of thumb is that every time you double you5 cache size you increase your SPEC performance by 10%.   < AMD's 32bit Intel Compiler based numbers are apparently 12025 uplift this by 20% for 64bit compilers and you end upi9 at 1440, the fastest Itanium II number is 810 with a 3 MB. cache.  8 You should expect at best 20% better performance because of the 3-12 MB hike.  7 This doesn't make Madison competitive and of course the04 numbers for Opteron are for the 2Ghz part not faster clocked units.  > Your assumptions seem to be based on AMD releasing Opteron and3 then giving up on trying to pump the clock rate up.K  = In reality this wont happen and in any case Opteron is likelyl: to be available in systems before Madison and it will also
 be faster.  D > 	AMD doesn't have to run out of cash (even though that is likely),E > 	Opteron won't have *nearly* the TPC marks of Monticeto/Madison nor @ > 	the SpecFp.  Opteron's niche is narrow compared to Itanium's. >   7 What makes you think that Itanium will out TPC Opteron,s5 concise technical response please not too much fluff.>   Regards  Andrew Harrison    ------------------------------  # Date: Fri, 25 Oct 2002 18:52:41 GMT # From: "John Smith" <a@nonymous.com>c% Subject: Re: Immutable laws of the PC I Message-ID: <Zpgu9.63042$mxk1.56177@news04.bloor.is.net.cable.rogers.com>   4 "David Mathog" <mathog@caltech.edu> wrote in message2 news:20021025081223.1b64f794.mathog@caltech.edu... > In this article: >A >sF > Intel (and HP?) must be praying AMD runs out of money before Opteron	 launches.     L You can bet that Intel will do anything to make that eventuality occur, evenE if it means that they are punished by the FTC *after* AMD goes out ofeL business. What are they going to do - shut down the only (sic.) manufacturer( of commodity microprocessors in the USA?  7 After all, aren't all the other processors proprietary?    ------------------------------  % Date: Fri, 25 Oct 2002 15:36:20 -0400t- From: JF Mezei <jfmezei.spamnot@videotron.ca>h% Subject: Re: Immutable laws of the PC , Message-ID: <3DB99D33.41862A94@videotron.ca>   Rob Young wrote:K >         AMD doesn't have to run out of cash (even though that is likely),sL >         Opteron won't have *nearly* the TPC marks of Monticeto/Madison norG >         the SpecFp.  Opteron's niche is narrow compared to Itanium's.n  N I think that Intel is smart enough not to allow AMD to sink too far. MicrosoftL rescued Apple when it was in dire straights, it knew that the death of AppleN would mean instant loss in its antitrust case with the various US governments.  ' Intel cannot afford to let AMD go down.t  G As far as Hammer vs IA64.  If, from day one 100% of the existing winteleN software can run on Hammer, and less than 100% can run on IA64, then IA64 will be at a big disadvantage.0  N The only way that IA64 can become "commodity" (or in Carly parlance: "industryN standard" is for it not only to surpass the 8086 by a large enough margin, butN also for Microsoft to start to provide certain software only on IA64 versions,9 making the 8086 architecture obsolete and killing Hammer.a  J But this assumes that Intel will be able to make IA64's for about the sameJ price as 8086s, and sell them at about the same price, AND will be able toS progress the architecture as fast as or faster than it could have the simpler 8086.u  M As long as IA64 remains a "high end" chip, it will have no more bigger marketnL share than Sparc or Alpha or PA-Risc. It will be relegated as a niche marketL chip that took billions and many many many years to get running with totallyK unimpressive performance.  It would seem to me that as a niche market chip,sH IA64 will have costed and continue to cost much much more than it costed; Digital and costs Sun to design/improve their architecture.,  N Intel may be able to eventually produce an IA64 with acceptable performance toN allow folks to migrate from the older (and by then slower) Alpha EV7. But willM it be a profitable venture for Intel ? With all the billions already sunk in,  can it ever be profitable  ?   ------------------------------    Date: 25 Oct 2002 15:52:03 -0500+ From: young_r@encompasserve.org (Rob Young)>% Subject: Re: Immutable laws of the PCc3 Message-ID: <qOCNF+iWM8We@eisner.encompasserve.org>w  X In article <3DB97DB9.4010409@nospamn.sun.com>, Andrew Harrison SUNUK Consultancy writes: >  > Rob Young wrote:e >> In article <20021025081223.1b64f794.mathog@caltech.edu>, David Mathog <mathog@caltech.edu> writes:o > E >> 	AMD doesn't have to run out of cash (even though that is likely),cF >> 	Opteron won't have *nearly* the TPC marks of Monticeto/Madison norA >> 	the SpecFp.  Opteron's niche is narrow compared to Itanium's.  >>   > 9 > What makes you think that Itanium will out TPC Opteron,e7 > concise technical response please not too much fluff.  >   7 	Cache size, cache latency.   Latency and L2 effects is C 	a tricky find.  However, as it was/is a definitive work or Richard @ 	Sites, of course a quick search is an indirect reference to it:  \ http://www-2.cs.cmu.edu/afs/cs.cmu.edu/academic/class/15740-f97/public/doc/21264-article.pdf  7 	Trying groups.google.com yields indirect results also:e  W http://groups.google.com/groups?selm=3prbid%249jr%40crl7.crl.com&oe=UTF-8&output=gplainM  ) 	But we do know better latency helps tpm.s  ) 	Regarding size... that is even trickier:-  ? http://research.compaq.com/wrl/projects/Database/isca99_rac.pdfa   	But a better result.i 	t4 	Itanium 2 with 1 cycle L1, 5 cycle L2, 12 cycle L3.  ? 	Montecito?  Maybe they keep the same cycles and associativity,e> 	don't know.  But I suspect L2 goes to 1.5 MBytes leaving 8 or< 	so for L3... all this to say we can see from slide 3 in theE 	isca99_rac.pdf showing a 50% increase in tpm when going from 4 MByte > 	to 8 Mbyte.  I do admit I may be comparing apples to oranges, 	but I made a stab.p    l 				RobH  aB Men with walkie-talkie                  I'm home again to you babeC Men with flashlights waving             You know it makes me wondersG Up upon the tower                       Sittin' in the quiet slipstreamo> The clock reads daylight savin'         Rollin' in the thunder  .                                 -- Neil Young    ------------------------------  # Date: Fri, 25 Oct 2002 20:50:21 GMT $ From: winston@SSRL.SLAC.STANFORD.EDU% Subject: Re: Immutable laws of the PCa8 Message-ID: <00A15FB5.EC80201B@SSRL04.SLAC.STANFORD.EDU>  \ In article <3DB99D33.41862A94@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: >Rob Young wrote: L >>         AMD doesn't have to run out of cash (even though that is likely),M >>         Opteron won't have *nearly* the TPC marks of Monticeto/Madison nor@H >>         the SpecFp.  Opteron's niche is narrow compared to Itanium's. ><O >I think that Intel is smart enough not to allow AMD to sink too far. MicrosoftoM >rescued Apple when it was in dire straights, it knew that the death of Apple O >would mean instant loss in its antitrust case with the various US governments.2  L If memory serves, Microsoft's "rescue" of Apple wasn't much of one. They putL $100 million or $150 million into Apple when Apple had a whole lot more cashL than that in the bank and didn't need the investment.  They came closer to aL real rescue by swearing to produce Office for Mac - but that just meant thatO they got to _sell_ Office for Mac, which I believe results in revenue for them.y   -- Alanm   ------------------------------    Date: 25 Oct 2002 17:09:53 -0500+ From: young_r@encompasserve.org (Rob Young)1% Subject: Re: Immutable laws of the PCe3 Message-ID: <eheRYc1Y35+c@eisner.encompasserve.org>7  _ In article <ZKmdnZDl74GrJySgXTWcog@metrocast.net>, "Bill Todd" <billtodd@metrocast.net> writes:    >>D >> AMD doesn't have to run out of cash (even though that is likely),A >> Opteron won't have *nearly* the TPC marks of Monticeto/Madison  > N > I think you'll find that Opteron flushes Madison down the toilet in TPC:  itM > will have better raw processor performance (TPC code really isn't very muchrJ > like FP code), it will have *far* better memory performance (whether youG > want to use latency or bandwidth as the criterion:  even 3-hop memoryiN > latency will be significantly better than Madison with the zx1 chipset), andG > its inter-processor communication is far better as well (not that TPC5> > requires a *lot* of coordination, but it does require some). >   3 	I say not a chance.  Here's two quick questions.  i  = 		1) Which OS and database will Opteron run when we see tpmC s
 			results?     4 		2) When will we see Opteron tpmC results posted to 			www.tpc.org?-  < 	Reasonable answers and I start to believe a tiny bit in the 	almighty Opteron.   				Robu   ------------------------------  % Date: Fri, 25 Oct 2002 17:40:24 -0400b* From: "Bill Todd" <billtodd@metrocast.net>% Subject: Re: Immutable laws of the PC'2 Message-ID: <ZKmdnZDl74GrJySgXTWcog@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:kGbNNgX+hCte@eisner.encompasserve.org...hG > In article <20021025081223.1b64f794.mathog@caltech.edu>, David Mathogg <mathog@caltech.edu> writes: >s > >sH > > Intel (and HP?) must be praying AMD runs out of money before Opteron	 launches.  > >  >uB > Opteron is "shipping at the end of the first quarter 2003", thisB > means that realistically systems show up May/June 2003 (assumingB > reasonable build time and qual and Opteron really doesn't slip).  J You mean kind of like Itanic2 systems weren't actually *obtainable* for atJ least a couple of months after the early July 'launch' (not that very manyJ seem to be being obtained even now, but that may be more a demand shortageA than a supply shortage)?  Possibly - but then again possibly not.0   >-> > So Opteron will be selling against whatever Xeon is shipping > in June 2003  I I see 'May/June' above just magically became just 'June' (even giving you @ the benefit of the doubt that it isn't actually March/April...).  % > with 533 MHz FSB, that's one issue.K  G Why?  Unlike K7 Athlons, at 5.3 GB/sec Opteron isn't at all starved foroI bandwidth.  Furthermore, unlike Xeon MP servers (which with a 533 MHz bus_K should, IIRC, have a *total* bandwidth of 4.3 GB/sec to share among however L many processors they contain), the aggregate available Opteron bandwidth forK both memory and I/O increases with each additional processor  - and becausetL they can feed each other over the HyperTransport links, at any given instantI a large portion of the total is available to any *single* processor if it 	 needs it.u     The secondB > issue is Intel Marketing.  A savvy Intel will be selling futures8 > in that timeframe for CPUs shipping in late 2003, i.e. > Montecito: >l6 > http://www.anandtech.com/cpu/showdoc.html?i=1677&p=2  I Except that  1) Montecito isn't scheduled to ship in 2003 at all, late oroG otherwise (it's been planned all along for 2004 - and still was as of 2,H weeks ago when Terry mentioned it in L'Inq) and  2) so far, Intel hasn'tI succeeded in selling *anything* about Itanic, whether futures or existingrJ hardware - except to some gullible vendors who unwisely scrapped their ownL platforms in anticipation (and in at least two cases then had to scramble to= resurrect them because the rosy promises never materialized).-   >kK > If Intel's desktop chips will get a 1MB L2 cache, what about their servera > solutions?  K What about them?  While extra cache is nice (Opteron will of course have atiH least 1 MB of its own on the chip, with 2 MB coming soon if not at firstK ship), fast main memory is far better - and at 48 - 57 ns. Opteron has Xeond% and P4 beat by a factor of 2 or more.r  =  It has been well known that Intel's Madison processor (thirdeL > generation derivative of Itanium) would have a massive 6MB on-die L3 cache on aH > 0.13-micron process, but armed with Intel's 90nm process the Montecito coreK > (fourth generation Itanium) will boast an incredible 12MB on-die L3 cache  usingt1 > close to 800 million of these 90nm transistors.o  E Well, that's not exactly news either.  Nor (as noted above) somethingvD Opteron will have to compete with in 2003, and since Opteron is alsoI scheduled to hit 90 nm. in (early) 2004 (the point where Fred Weber notedrE that the dual cores that the design already includes become eminentlyHK feasible, though he didn't announce that officially) I somehow suspect that E it will do just fine against Montecito - especially as Montecito wills@ apparently still be waiting 156 ns. (or longer, in > 4-processorH configurations) to reach main memory, vs. about 1/3 of that for Opteron.   >  >o> > Opteron is a wonderful CPU, but certainly won't compete with6 > Itanium and a 800 million transistor/12 MByte cache.  I Indeed it won't:  Opteron will still be out in front (whether you want to H use SPECint, memory latency, or bandwidth as the criterion), still usingF considerably less power in the process, and still a hell of a lot lessH expensive.  Since the initial 130 nm. Opteron is *already* demonstratingG performance equal the very top of the range (30% - 50% improvement overbG McKinley) that Intel predicts for the 130 nm. Madison (and without evenaG using 64-bit mode yet), just one of the two cores on the 90 nm. Opteront( should easily be the equal of Montecito.   >pC > AMD doesn't have to run out of cash (even though that is likely),e@ > Opteron won't have *nearly* the TPC marks of Monticeto/Madison  L I think you'll find that Opteron flushes Madison down the toilet in TPC:  itK will have better raw processor performance (TPC code really isn't very muchoH like FP code), it will have *far* better memory performance (whether youE want to use latency or bandwidth as the criterion:  even 3-hop memoryiL latency will be significantly better than Madison with the zx1 chipset), andE its inter-processor communication is far better as well (not that TPCr< requires a *lot* of coordination, but it does require some).    nor? > the SpecFp.  Opteron's niche is narrow compared to Itanium's.l  K You've got it exactly backward, Rob:  it's Itanic that's the niche product, H with FP-style code the only area in which it won't look like the fatter,L hotter, more expensive, and slower alternative that that Itanic1 established> as a precedent for the family.  Its only hope is that switchedH HyperTransport doesn't materialize in time to allow Hammer to invade theE 32-processor-and-up server range soon (since the arrival of dual-coreo2 Opterons will get them to the 16-processor level).   - bill   ------------------------------  % Date: Fri, 25 Oct 2002 18:39:27 -0400C- From: JF Mezei <jfmezei.spamnot@videotron.ca>t% Subject: Re: Immutable laws of the PC , Message-ID: <3DB9C812.531FF83A@videotron.ca>   Rob Young wrote:L >                 1) Which OS and database will Opteron run when we see tpmC" >                         results?  J Intel can only aim its IA64 at the high end "enterprise" market because itK needs high profit margins to recuperate the huge wad of money that was sunk  with Itanic.  K Hammer will serve both as a commodity wintel chip and potentially then growIJ into a serious enterprise chip. It is designed to compete in a price basis% with Intel's commodity 8086 products.t* (at least at the high end of the desktop).  M It is a lot easier to take a low cost, high volume commodity chip and grow itiM to enterprise markets, then to take a high cost, low volume incompatible chipA+ and concquer the low cost commodity market.H  H So, if Hammer initially is aimed at the desktop and wintel servers, thenL performance isn't that important since that market tends to look only at the Hz metric to gauge performance.   I From what I understood, AMD will initially release a low end "hammer" andd@ later on unleash the full 64 bit "enterprise version" of Hammer.   ------------------------------  # Date: Fri, 25 Oct 2002 23:31:11 GMTh- From: bdc@world.std.com (Brian 'Jarai' Chase)B% Subject: Re: Immutable laws of the PCn& Message-ID: <H4K9Bz.Fyp@world.std.com>  2 In article <ZKmdnZDl74GrJySgXTWcog@metrocast.net>,) Bill Todd <billtodd@metrocast.net> wrote:q: > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:kGbNNgX+hCte@eisner.encompasserve.org...a     > >eE > > AMD doesn't have to run out of cash (even though that is likely),nB > > Opteron won't have *nearly* the TPC marks of Monticeto/Madison > N > I think you'll find that Opteron flushes Madison down the toilet in TPC:  itM > will have better raw processor performance (TPC code really isn't very muchhJ > like FP code), it will have *far* better memory performance (whether youG > want to use latency or bandwidth as the criterion:  even 3-hop memory N > latency will be significantly better than Madison with the zx1 chipset), andG > its inter-processor communication is far better as well (not that TPCs> > requires a *lot* of coordination, but it does require some). >  >  norA > > the SpecFp.  Opteron's niche is narrow compared to Itanium's.n > M > You've got it exactly backward, Rob:  it's Itanic that's the niche product,uJ > with FP-style code the only area in which it won't look like the fatter,N > hotter, more expensive, and slower alternative that that Itanic1 established@ > as a precedent for the family.  Its only hope is that switchedJ > HyperTransport doesn't materialize in time to allow Hammer to invade theG > 32-processor-and-up server range soon (since the arrival of dual-coreg4 > Opterons will get them to the 16-processor level).  J In the realm of the exotic, it was recently announced that Cray and SandiaC National Laboratories will be using AMD Opterons in their Red Storm  supercomputer:  ?    http://biz.yahoo.com/rc/021021/tech_amd_supercomputer_2.htmle  B IBM's Blue Gene sounds even more interesting.  It's to be based on9 stripped down Power4 architecture CPUs with 32 cores/chi.   ,    http://news.com.com/2100-1001-233954.html  C The above are rather impractical and uncommon examples, but it does7D illustrate that intelligent people recognize Opteron (and Power4) asI viable in terms of their performance.  I other words, I don't think Intel I and HP have a sure win with the next generation(s) of Itanium.  They've aAJ fight ahead of them that's going to be waged on more than one front in theE server markets.  I do certainly think it's going to be interesting toa watch.  J Also, does anyone else find it as amusing as I do that Intel *brags* about< the ridiculously high transistor counts of their processors?   -brian.  -- EF --- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----+ Font-o-Meter!      Proportional  Monospacedu'                                       ^.   ------------------------------  # Date: Sat, 26 Oct 2002 00:10:56 GMTi- From: bdc@world.std.com (Brian 'Jarai' Chase)u% Subject: Re: Immutable laws of the PCs& Message-ID: <H4KB68.4rx@world.std.com>  3 In article <eheRYc1Y35+c@eisner.encompasserve.org>,N, Rob Young <young_r@encompasserve.org> wrote:@ > In article <ZKmdnZDl74GrJySgXTWcog@metrocast.net>, "Bill Todd"" > <billtodd@metrocast.net> writes:  P > > I think you'll find that Opteron flushes Madison down the toilet in TPC:  itO > > will have better raw processor performance (TPC code really isn't very muchaL > > like FP code), it will have *far* better memory performance (whether youI > > want to use latency or bandwidth as the criterion:  even 3-hop memoryMP > > latency will be significantly better than Madison with the zx1 chipset), andI > > its inter-processor communication is far better as well (not that TPCt@ > > requires a *lot* of coordination, but it does require some). > 5 > 	I say not a chance.  Here's two quick questions.  o > ? > 		1) Which OS and database will Opteron run when we see tpmC a > 			results?  0  J Well... Linux is already ported to the AMD x86-64 architecture.  RedHat isH committed to offering an x86-64 distribution of Linux.  IBM DB2 is beingE ported (or has been ported already?).  And Oracle has been very vocalLI about Linux for the past year or so.  I'd imagine Oracle won't want to be E too far behind IBM on this, and I doubt it'd be very hard for them tosC tweak their Linux port of Oracle to get it running on x86-64 Linux.V  8    http://www.redhat.com/partners/press_partner_amd.html    http://www.oracle.com/linux  6 > 		2) When will we see Opteron tpmC results posted to > 			www.tpc.org?   I I'd guess we'd see it for DB2/Linux around the time that Opteron systems   become available.e   -brian.: -- mF --- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----+ Font-o-Meter!      Proportional  Monospaced '                                       ^t   ------------------------------  % Date: Fri, 25 Oct 2002 21:11:12 -0400 * From: "Bill Todd" <billtodd@metrocast.net>% Subject: Re: Immutable laws of the PC 2 Message-ID: <l_ycnQhbOP0DdiSgXTWcrg@metrocast.net>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3DB9C812.531FF83A@videotron.ca...   ...n  K > From what I understood, AMD will initially release a low end "hammer" andDB > later on unleash the full 64 bit "enterprise version" of Hammer.  K That plan changed recently (perhaps because AMD realized that it would take L some time to ramp up production quantities and that selling high-end HammersL rather than low-end Hammers during that period would be significantly better for its financial position).  J While both variants will nominally release at the end of Q1 next year, theI initial emphasis will be on the server (Opteron) units (though I may have_J seen somewhere that they may not exceed 4-processor boards initially, just# as Itanic2 didn't at introduction).a   - bill   ------------------------------  % Date: Fri, 25 Oct 2002 21:26:49 -0400t* From: "Bill Todd" <billtodd@metrocast.net>% Subject: Re: Immutable laws of the PCR2 Message-ID: <JsSdnY_kQLHaciSgXTWcow@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:eheRYc1Y35+c@eisner.encompasserve.org...i@ > In article <ZKmdnZDl74GrJySgXTWcog@metrocast.net>, "Bill Todd"  <billtodd@metrocast.net> writes: >  > >>F > >> AMD doesn't have to run out of cash (even though that is likely),C > >> Opteron won't have *nearly* the TPC marks of Monticeto/Madison  > > L > > I think you'll find that Opteron flushes Madison down the toilet in TPC: itJ > > will have better raw processor performance (TPC code really isn't very muchL > > like FP code), it will have *far* better memory performance (whether youI > > want to use latency or bandwidth as the criterion:  even 3-hop memoryaL > > latency will be significantly better than Madison with the zx1 chipset), and,I > > its inter-processor communication is far better as well (not that TPCi@ > > requires a *lot* of coordination, but it does require some). > >a >t2 > I say not a chance.  Here's two quick questions. >e< > 1) Which OS and database will Opteron run when we see tpmC
 > results?  I Why do you care?  Oh - perhaps to obtain a direct comparison with Itanic?-L Too bad Itanic chose Windows and SQL to use in its TPC-C testing, since thatI makes it impossible to perform direct comparisons with serious enterprise6K database platforms (in an interestingly parallel situation, when Unisys ranmL its TPC-H tests on Itanic it chose the 300 GB test series - where no seriousK competition from the heavy hitters exists - rather than the 1000 GB or 3000rH GB test series, again making direct comparisons impossible, but it stillI touted the results as proof of its suitability for large, enterprise dataeI management, despite having run a database size that would fit on a singley) one of the new 320 GB Maxtor IDE drives).b  G But take heart:  Windows will be supporting Hammer before long.  In thetH meantime, should Itanic supporters be brave enough to run TPC-C tests on? Oracle, an earlier fairly direct comparison should be possible.    >h4 > 2) When will we see Opteron tpmC results posted to > www.tpc.org?  K That's actually a good question, because it highlights the possibility thatoF no system vendor may have the incentive to push Opteron (TPC-C being aL fairly expensive test to set up) that HP has to push (tow?) Itanic.  PerhapsC Oracle or IBM will step up to the plate, since the results (both in J tpmC/processor and $/tpmC) are likely to make their database products look good.s   > = > Reasonable answers and I start to believe a tiny bit in ther > almighty Opteron..  I I don't give a particular damn if you believe in the Tooth Fairy, Rob:  I I just wouldn't want your credulousness in all things Itanic to lead otherst astray.    - bill   ------------------------------  % Date: Fri, 25 Oct 2002 21:06:14 -0500c" From: xganon <remailer@xganon.com>% Subject: Re: Immutable laws of the PC-9 Message-ID: <b38d9a35885dca85c2aa1f044735a50a@xganon.com>7   >MC >>AMD doesn't have to run out of cash (even though that is likely),iD >>Opteron won't have *nearly* the TPC marks of Monticeto/Madison nor? >>the SpecFp.  Opteron's niche is narrow compared to Itanium's.. >> s > A >What makes you think that Itanium will out TPC Opteron, concise c. >technical response please not too much fluff. >   "         Cache size, cache latency.  C Big factors.  But like I said, branch predictions also.  Why do youe5 think a less powerful 4P Itanium 2 does better than a G 4P Power4 or a 4P Alpha?  Branch prediction is key to stronger Itanium t$ 2 transactions.  It isn't bandwidth.@ Couple improved BP with larger faster caches and zoom zoom zoom.  B hammer does better at specint.  That will sell a lot of systems in the server space.c   ------------------------------  % Date: Fri, 25 Oct 2002 23:29:07 -0400l- From: JF Mezei <jfmezei.spamnot@videotron.ca>a% Subject: Re: Immutable laws of the PCu, Message-ID: <3DBA0BFC.93C5320F@videotron.ca>   Bill Todd wrote:I > But take heart:  Windows will be supporting Hammer before long.  In thehJ > meantime, should Itanic supporters be brave enough to run TPC-C tests onA > Oracle, an earlier fairly direct comparison should be possible.     N And this is where Hammer stands a MUCH better chance than IA64. Because of itsN intrinsic support of 8086 code, Hammer gets instant access to the billions andH billions of 8086 programs already on the market, match or outperform the 8086-du-jour.   J Until those billions and billions of software titles run natively on IA64,' IA64 will have a serious disadvantage. b  N AMD will have a single architecture that will scale from desktop to enterpriseL systems. Intel will have to develop 2 separate architectures in paralel, andK so far, it looks like performance for those 2 architectures won't be so far  apart. /  N IA64 will have a niche with proprietary systems suchg as Tandem and HP-UX, butK I don't foresee it will have much future with Windows based systems. RemovesJ the mass market that windows brings, and you loose any hope of reaping theN rewards of large volumes and earning the coveted title of "industry standard".  M Although Alpha is dead, I think that IA64 may in fact end up breaking Intel'stG monopoly. The requirement to have separate binaries for IA64 means that L upgrading from 8086 to Ia64 will have the same pain factor as upgrading from 8086 to Power or Sparc.-  L And just like Compaq was able to originally produce a clone that ran "FlightN simulator" better than  IBM's own IBM-PCs, it is possible that AMD may be ableM to produce an 8086 that is better than Intel's own and this would marginalisei% both the Intel 8086 and Intel,s IA64.R   ------------------------------  # Date: Fri, 25 Oct 2002 20:32:30 GMTS$ From: winston@SSRL.SLAC.STANFORD.EDU: Subject: Re: It'll be interesting to see what happens here8 Message-ID: <00A15FB3.6E7E8BE1@SSRL04.SLAC.STANFORD.EDU>  c In article <cb2W$60U8avw@eisner.encompasserve.org>, koehler@encompasserve.org (Bob Koehler) writes:ao >In article <sNYt9.61917$Q3S.58334@news01.bloor.is.net.cable.rogers.com>, "John Smith" <a@nonymous.com> writes:.5 >> http://www.eweek.com/article2/0,3959,644757,00.aspe >>   >> http://www.openhack.com/t >> s5 >> http://www.eweek.com/article2/0,3959,643205,00.aspe >aI >   You'll note that HP is supporting this, but it's a MS and Oracle onlyl >   setup.   >nF >   No point in something else showing it can really be done without a6 >   large contigent of expert security patch appliers.  4 Not if you figure the real money is in services, no.   -- Alan    ------------------------------  % Date: Fri, 25 Oct 2002 20:26:05 -0500a1 From: "David J. Dachtera" <djesys.nospam@fsi.net>-$ Subject: Re: Monitor System Question' Message-ID: <3DB9EF2D.91416584@fsi.net>e   Alder wrote: >  > Phillip Helbig wrote:aE > >>Outlook Express uses Arial as its default font. You must copy theeH > >>screen or file to MS Word and use a Courier New or r-ansi font whichI > >>use by MS Outlook Express. You can copy-paste or attached the MS Word  > >>doc. to the email. > >  > >w > > Surely you jest! > > L > > Wouldn't a fixed-width font for the newsreader would be a better option? > F > He must have been, because I later found how to use fixed-width fromD > WITHIN Outlook Express.  Far too off-topic for this group, though.  ? Not really THAT far off-topic. The r_ansi font ships with WRQ'srA Reflection terminal programs, which are frequently used to accessG OpenVMS systems.   -- s David J. Dachtera@ dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Fri, 25 Oct 2002 20:49:54 GMT = From: "Colin Butcher" <colinDOT.butcherAT@xdeltaDOT.coDOT.uk>t5 Subject: Now where's the OpenVMS OpenOffice software?h= Message-ID: <S7iu9.1654$Cq6.10460410@news-text.cableinet.net>r  F OK, it's a little off topic, but I couldn't resist posting this UK job advert:eH ------------------------------------------------------------------------& Job Title: SENIOR/PRINCIPAL CONSULTANT  E Skills: One of our clients require a Senior/Principal Consultant. Thee  : main role will be to assess the use of alternative desktop  4 environments to that of using Microsoft products....   Reference: vnu nmy/dxl/c1e  H For more details and to apply for this job, please click the link below:  ( <http://www.newmonday.co.uk/Job/3705704>     Cheers, Colin.' (colinDOT.butcherAT@xdeltaDOT.coDOT.uk)    ------------------------------    Date: 25 Oct 2002 15:20:53 -0700- From: carlos.costa@datawest.ca (Carlos Costa)u% Subject: Re: OpenVMS 7.3-1 and CSWINGf= Message-ID: <7edc7395.0210251420.7caf99fe@posting.google.com>n  e "Dan Allen" <dallen@nist.gov> wrote in message news:<JFEPKAPBPMDFDBOIANGDCEJFCBAA.dallen@nist.gov>...k > > -----Original Message-----, > > From: Carlos Costa [mailto:calo@shaw.ca]- > > Sent: Thursday, October 24, 2002 12:13 AM  > > To: Info-VAX@Mvb.Saic.Comi) > > Subject: Re: OpenVMS 7.3-1 and CSWINGcN > > One of my points, though, that no one has really answered is how about the! > > assertion from HP/Compaq thato > > = > > "Compaq is confident that the new release has 100% binaryK> > > compatibility with the previous release. If an applicationI > > compatibility problem is discovered, Compaq will assign the problem acJ > > high priority and commit to providing a fix. There is no need for ISVsA > > to test on the new release or produce a new application kit."h > > K > > Certainly, CSWING is not the only application in the VMS world that has?L > > dubious coding. And just because an application is not freeware does not6 > > make it automatically good code (and vice versa!). > > N > > The above statement even says that "There is no need for ISVs to test". ToM > > me that means that if it works under 7.3-0, no matter what bugs it has inpM > > it, it should work under 7.3-1. By the way, my Gold Support TAM confirmed5O > > and duplicated the problem, so it's not just me. Apparently they use CSWING 0 > > too, but they haven't upgraded to 7.3-1 yet. > M >      	Are you serious! Do you really think that ANY operating system vendorhF > 	can or will commit to maintaining compatability with latent bugs in= > 	all of the application software ever developed for the OS!   A Excuse me, but yes, I am serious. If you read the above statementdD carefully you will see that that is exactly what HP/Compaq has done.F They have committed to maintaining 100% binary compatability. I didn'tE say it, THEY did. If they didn't mean it they shouldn't have said it.dF Binary compatibility is a pretty tight compatibility. It means that noD program needs to be re-compiled or re-linked. It will work as it is." They even say "no testing needed".  D Secondly, one of the reasons I pointed that statement out is becauseA that statement means, to me, that there is a bug in the VMS code. = HP/Compaq expected something like it because they said "If annF application compatibility problem is encountered...". Well, my friend,E an application compatibility problem *has* been encountered. So, now,uE instead of them crying out about all the the bugs in application code E and about how great, and solid, and bug-free VMS is, they ought to be B looking at *why* there is an application compatibility problem and honouring their word.e  ? There seems to be lots of scorn thrown the way of CSWING merelyeB because it is freeware. Non-free programs have bugs too, and lousyD code. The point is not whether there are bugs or not, but that thereE is a compatibility issue, and whether that issue is the result of bad > coding either in CSWING or in VMS is immaterial. HP/Compaq hasE "committed" to fixing it. Their word. They didn't say they would onlypC fix it "if it's a commercial program", or "if it is bug free code".T  E However, having said that, I don't expect HP/Compaq to admit that theeF CSWING issue is caused by a bug in VMS. This is why I appealed to this group of fine VMS users.     > While I don't<J > speak for VMS engineering or HP I'm confident they do not think that theO > assurances they gave with regard to compatability apply to buggy applicationsgL > i.e. they do not warrant 7.3-1 to be "bug compatible" with 7.3.  ObviouslyJ > fixing an OS bug that allowed an application bug to be masked or ignoredI > will cause the application to fail.  New releases often fix bugs - duh!o > 	h  > Yes, that is true. That is why they added the statement "If anC application compatibility problem is discovered, Compaq will assigneF the problem a high priority and commit to providing a fix.". They knowD that that may happen, and guess what? It did. No surprise there. The> only surprise is why people are opposed to finding a solution.    B >With regard to the "no need for ISV's to test" assertion they are	 obviously:C >silly and any ISV who believes that there is no need to test theirr product  >on a new release is silly too.o  D No, not silly. Merely confident. This is why this is an "EnhancementE release". The release notes go on to say "The release may contain new = hardware support, software support, software enhancements andm? maintenance, but the changes are isolated and have no impact onn applications."  F Their words are "no impact on applications". Seems pretty clear to me.E If they had done everything right then CSWING would still be working.sC If there is an impact in applications (as there clearly is) then itrE means that someone screwed up. So, let's be adult about it, 'fess up,t+ fix the problem, and get on with our lives.n  E If 7.3-1 is more than an Enhancment release, then it should have beentD called 7.4. as someone has already pointed out. But they didn't callD it 7.4 because they were expecting CSWING to work. Must have slipped their testing somehow.    B >It's been two days since I pointed out the weakness in the status checkingD >in the code snippet you posted and the likelyhood that it's failing to properlytD >detect and report a system service error.  Have you applied the one lineA >patch I offered and determined that my suspicions are or are not 
 confirmed?    Actually, I did. What I got was:  ! Searching, root = SHODAN:[000000] " %SYSTEM-W-NOSUCHFILE, no such file/ %TRACE-W-TRACEBACK, symbolic stack dump follows C   image    module    routine             line      rel PC          a abs PC>  CSWING                                     0 0000000000021310 0000000000031310>  CSWING                                     0 00000000000214A8 00000000000314A8>  CSWING                                     0 00000000000214A8 00000000000314A8>  CSWING                                     0 0000000000023064 0000000000033064>  CSWING                                     0 0000000000022D80 0000000000032D80>  CSWING                                     0 0000000000023C94 0000000000033C94>  CSWING                                     0 00000000000209F4 00000000000309F4>  CSWING                                     0 0000000000020070 0000000000030070>                                             0 FFFFFFFF8027B63C FFFFFFFF8027B63C$ %SYSTEM-W-NOMOREFILES, no more files/ %TRACE-W-TRACEBACK, symbolic stack dump follows:C   image    module    routine             line      rel PC          o abs PC>  CSWING                                     0 0000000000021310 0000000000031310>  CSWING                                     0 00000000000214A8 00000000000314A8>  CSWING                                     0 00000000000214A8 00000000000314A8>  CSWING                                     0 0000000000023064 0000000000033064>  CSWING                                     0 0000000000022D80 0000000000032D80>  CSWING                                     0 0000000000023C94 0000000000033C94>  CSWING                                     0 00000000000209F4 00000000000309F4>  CSWING                                     0 0000000000020070 0000000000030070>                                             0 FFFFFFFF8027B63C FFFFFFFF8027B63C" %SYSTEM-W-NOSUCHFILE, no such file/ %TRACE-W-TRACEBACK, symbolic stack dump followsTC   image    module    routine             line      rel PC          R abs PC>  CSWING                                     0 0000000000021310 0000000000031310>  CSWING                                     0 00000000000214A8 00000000000314A8>  CSWING                                     0 00000000000214A8 00000000000314A8>  CSWING                                     0 0000000000023064 0000000000033064>  CSWING                                     0 0000000000022D80 0000000000032D80>  CSWING                                     0 0000000000023C94 0000000000033C94>  CSWING                                     0 00000000000209F4 00000000000309F4>  CSWING                                     0 0000000000020070 0000000000030070>                                             0 FFFFFFFF8027B63C FFFFFFFF8027B63C$ %SYSTEM-W-NOMOREFILES, no more files/ %TRACE-W-TRACEBACK, symbolic stack dump followstC   image    module    routine             line      rel PC          - abs PC>  CSWING                                     0 0000000000021310 0000000000031310>  CSWING                                     0 00000000000214A8 00000000000314A8  	 etc, etc.A  E It means absolutely nothing to me. Told you I was weak in VMS systemsz programming.  F >If not I'd suggest you quit whining about what MIGHT be an OS bug and give usaF >some hard info on what's actually happening. If I had the environment to testmA >the damn thing myself I would have already done it for you! Thisi isn't rocket? >science, it's not even computer science, it's programming 101.o  E I don't know about that, the particular code is recursive and full of ? VMS specifics. Not for the faint of heart. And please don't gettC personal about this. You're hinting that I'm not as smart as you tofA not have fixed this myself, and you could very well be right. ButhB "Don't judge someone until you've walked a mile in his shoes." ButE you've only seen a small code fragment. There are many more thousands 	 of lines.c  C I admitted as such at the beginning of this thread. I don't know mytA way around the VMS programming world too well. Sir, this is why IeC appealed to this group. If I was able to fix it myself I assure you-C that I would have. And, perhaps, given another few months I'll knowvE enough about the VMS calls and idiosyncracies to figure it out. After F looking at this for a while, and getting nowhere, I thought I'd see ifE anyone more knowledgable and experienced may have already have looked @ at it. Please don't begrudge me that. I wouldn't exactly call itF whining, more like a feeble cry for help in the wilderness. Instead of3 a helping hand do I now get wolves on my trail? :-)d     Carlos   ------------------------------  # Date: Sat, 26 Oct 2002 00:29:24 GMTo- From: "John E. Malmberg" <wb8tyw@qsl.network> % Subject: Re: OpenVMS 7.3-1 and CSWINGc: Message-ID: <Ellu9.1582$Zx.265935@news1.news.adelphia.net>   Paul Repacholi wrote:r1 > carlos.costa@datawest.ca (Carlos Costa) writes:i > C >>I recently upgraded VMS to 7.3-1 (from VMS 7.3-0) on two separaterF >>clusters. I have also been using CSWING for years now on various VMS >>versions and I love it.  > F >>However, in 7.3-1 CSWING no longer works (it does under 7.3-0). WhenC >>it goes to search a new directory tree it goes off to never-neverW >>land.w > D >>Has anyone else experienced this, and more importantly, is there a
 >>fix out? >   D According to the Compaq C compiler, when you remove the /VAXC which 4 basically suppresses most warning messages, and add ! /WARN=ENABLE=(LEVEL4, QUESTCODE):t  G Several of the routines involved in doing the directory transversal do mA not always specify their return value.  This means that they are sG returning pseudo random numbers instead of either a logical value or a h pointer.  G C99 is required to diagnose this bug.  C89 and earlier is not required a to diagnose this bug.   F The latest Compaq C compiler can diagnose these bugs, but that is not  the default mode.   - > It has done this `forever' with ISO9660 CDsf  H It really depends on what is in the R0 register at the time the routine H exits.  That is not easily predictable.  A small change in the run-time 9 library, or even the compile options could change things.   F >>This problem is also disconcerting because 7.3-1 has, in HP/Compaq's@ >>own words, "100% binary compatibility with previous versions". >   H > I think you missunderstand the statment, and some of the moderatly susB > things CSWING does. But I too am curious as to what has changed.  E It will probably take me a few more lunch hours to get it to compile jD cleanly.  Currently I have up to swing_4.c compiling with no errors.  F But I have had to make several updates to the source, most of them by F adding prototypes and correcting the types of variables and structure  members.  E Most of the problems that I have seen are minor, but effectively the 3H directoy node transversal routines have always been sometimes returning  random numbers.a  H As pointed out on my earlier post, which seems to have been missed, the  following errors are present:i  I The directory transversal routines are sometimes returning psuedo random h  numbers instead of valid values.  H The status value in an IOSB is a 16 bit value, but is being declared in A the routine as a longword.  This bug is probably not causing the  H problem.  But there is no guarantee that the upper 16 bits will be zero 
 in this case.3  F It also appears that the smg*.h are missing #pragma settings that are  needed for Compaq/DEC C.   -John  wb8tyw@qsl.network Personal Opinion Onlye   ------------------------------  % Date: Fri, 25 Oct 2002 22:34:13 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)% Subject: Re: OpenVMS 7.3-1 and CSWINGsK Message-ID: <rdeininger-2510022234130001@1cust118.tnt2.nashua.nh.da.uu.net>p  = In article <7edc7395.0210251420.7caf99fe@posting.google.com>,o. carlos.costa@datawest.ca (Carlos Costa) wrote:   <snip>  @ >There seems to be lots of scorn thrown the way of CSWING merelyC >because it is freeware. Non-free programs have bugs too, and lousyh >code. t  J I think most folks hear have a lot of appreciation for freeware.  If thereC is any scorn, it is likely for the bugs that are being found.  Bugs - happen, but some of these are pretty blatant.a   <snip>  ? >Yes, that is true. That is why they added the statement "If anyD >application compatibility problem is discovered, Compaq will assignG >the problem a high priority and commit to providing a fix.". They knowsE >that that may happen, and guess what? It did. No surprise there. Then? >only surprise is why people are opposed to finding a solution.i  E I think several folks _here_ have offered pointers toward a solution.0  J The first part of the solution is to remove any latent bugs in CSWING thatJ might be the cause of the symptom.  If bugs remain that are due to V7.3-1,2 then it's time to beat HP over the head with them.  J IMHO, Compaq's compatibility statement is far too strong to be realistic. F It was probably written by a marketing person.  I can't picture a saneJ engineer making this statement with a straight face.  HP will have to sort it out.e    C >>It's been two days since I pointed out the weakness in the status 	 >checking E >>in the code snippet you posted and the likelyhood that it's failingf >to properlyE >>detect and report a system service error.  Have you applied the onet >line B >>patch I offered and determined that my suspicions are or are not >confirmed?o >i! >Actually, I did. What I got was:o >a" >Searching, root = SHODAN:[000000]# >%SYSTEM-W-NOSUCHFILE, no such filel  H Ah!  This is useful.  "No such file" means we asked for a file spec that isn't there.  G Might be worth looking at whether the program deals with this condition 	 properly.e  F The traceback is useful if you are trying to debug with a link map andH compiler listings.  It shows each level in the call chain of functions. J The addresses given here can be matched up with specific lines of code andH specific machine instructions.  The traceback answers the question, "how did we get here".   H But this isn't a very fun way to debug.  It's much easier to use the VMSJ debugger.  There's a fair learning curve for that, but the debugger manual is quite good.  0 >%TRACE-W-TRACEBACK, symbolic stack dump follows   <snip>  % >%SYSTEM-W-NOMOREFILES, no more filese  / This is a different condition, "no more files".w  # >%SYSTEM-W-NOSUCHFILE, no such filer   And another "no such file"...    etc.    F >It means absolutely nothing to me. Told you I was weak in VMS systems
 >programming.w  H I given one or two hints above.  I don't have the program in front of meJ to do more investigating.  See the C documentation and the debugger manualH if you are interested in gaining debugging skills, and feel free to post questions here.-  G >>If not I'd suggest you quit whining about what MIGHT be an OS bug ando >give usG >>some hard info on what's actually happening. If I had the environmentf >to testB >>the damn thing myself I would have already done it for you! This
 >isn't rocket @ >>science, it's not even computer science, it's programming 101.  C Well, that comment was uncalled for.  Old timers tend to forget howeE overwhelming a big program appears to someone with little experience.2  F >I don't know about that, the particular code is recursive and full of, >VMS specifics. Not for the faint of heart.   J Agreed.  If you don't have much VMS programming experience, CSWING doesn'tF look like an ideal self-study aid.  It has a number of errors, many ofG which have been dormant for years.  C lends itself to this problem in aeG number of ways which I won't get into at the moment.  And the old VAX C 9 was rather weak at pointing out bad programming practice.l   ------------------------------    Date: 25 Oct 2002 14:44:04 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones)n Subject: Re: page or pagelet= Message-ID: <8a646952.0210251344.705a9554@posting.google.com>t  d "Tom Linden" <tom@kednos.com> wrote in message news:<CIEJLCMNHNNDLLOOGNJIKEIAFPAA.tom@kednos.com>... > >-----Original Message-----,E > >From: VAXman- @SendSpamHere.ORG [mailto:VAXman- @SendSpamHere.ORG]k+ > >Sent: Thursday, October 24, 2002 2:21 PMh > >To: Info-VAX@Mvb.Saic.Com > >Subject: Re: page or pageletr > >s > >t? > >In article <8a646952.0210241256.296d1d1@posting.google.com>,n< > >jones.computer.srv@worldnet.att.net (Daryl Jones) writes:/ > >>VAXman-  @SendSpamHere.ORG wrote in message4/ >  news:<00A15D65.98CDBFDF@SendSpamHere.ORG>...-: > >>> In article <ap41pv$t0g1@rain.i-cable.com>, "Kenneth"& >  <yeung_kenneth@hotmail.com> writes:A > >>> >In VMS 7.2-1, Alpha 8400. Does any on know when I do a monB >  page, the free>C > >>> >page list and the modified page list shown is in the unit ofl >  page (8KB) or > >>> >pagelet (512K)? > >>> >r > >>> >s > >>>g > >>> Hmm.  Let's see. > >>>N' > >>> When I do it on my system, I see:  > >>>  > >>> $ MONITOR PAGE9 > >>>                             OpenVMS Monitor Utility < > >>>                            PAGE MANAGEMENT STATISTIICS	 > >>>   :'H >                                   CUR        AVE        MIN        MAX	 > >>>   :r; > >>>     Free List Size               237273.00  237273.59  >  237273.00  237276.00v; > >>>     Modified List Size              350.00     350.00  >  350.00     350.00 > >>>r > >>>: > >>> Then, if I issue:a > >>>i > >>> $ SHOW MEMORY/PHYSICALF > >>>               System Memory Resources on 22-OCT-2002 15:05:49.40 > >>>rB > >>> Physical Memory Usage (pages):     Total        Free      In >  Use    Modified: > >>>   Main Memory (2048.00Mb)         262144      237442 >  24353         349 > >>> > > >>> Of the physical pages in use, 5523 pages are permanently >  allocated to OpenVMS. > >>>  > >>>e > >>>r7 > >>> It would look to me to be a page of 8K.  (2GB/8K)e > >>>i > >>> + > >>> FYI, a pagelet is 512 bytes not 512K.h > >>F > >>A page on an Alpha VMS machine is 8192 bytes per page and on a VAXF > >>machine it is 512 bytes per page. This is due to RISC vs CISC chipF > >>technology where the image sizes grew to be 5-6 time larger on the > >>RISC chip. > >m# > >No it's not!  Who told you this?e > E > I think it is probably close to true.  When you write compilers younI > have occassion to to look at things like this and I think it is obviouseG > to anyone who has coded assembly on VAX and Alpha that it takes a lot D > more instructions to accomplish the same task.  In the case of ourJ > PL/I compiler, on Alpha it is more than 7 times larger.  Of course, thisH > uses the GEM backend versus VCG, so I think the 5-6 figure is probablyJ > fairly accurate.  If you compare the sizes of DECC you will find about aK > 9:1 ratio, but again, part of that may be due to the much larger backend.  >  > >n > >e? > >> A pagelet on an Alpha VMS processor is 512 bytes. It takes.H > >>16 pagelets (16*512=8192) to equal to a page on a Alpha VMS machine. > >n: > >Unless your particular Alpha has 16K, 32K or 64K pages. > >--t5 > >VAXman- OpenVMS APE certification number: AAA-0001a > >VAXman(at)TMESIS(dot)COM* > >*7 > >  "Well my son, life is like a beanstalk, isn't it?"p > >s > >---) > >Incoming mail is certified Virus Free..= > >Checked by AVG anti-virus system (http://www.grisoft.com).PD > >Version: 6.0.404 / Virus Database: 228 - Release Date: 10/15/2002 > >- > ----( > Outgoing mail is certified Virus Free.< > Checked by AVG anti-virus system (http://www.grisoft.com).C > Version: 6.0.404 / Virus Database: 228 - Release Date: 10/15/2002l   To Everyone:  D From Alpha Architecture Reference Manual, Digital Press, 1992, pages	 vii-viii.e  D "Alpha is designed to handle the largest computing problems of today< and tomorrow. When the Alpha architecture is compared to its< predecessor, the VAX architecture, two differences stand outC immediately. First, Alpha is a 64-bit architecture; VAX is a 32-bitsB architecture. This means that the Alpha virtual address extends to@ 64-bit linear range of bytes in memory. Supporting this extendedD virtual address space are an extended maximum physical address range@ (up to 48 bits) and larger pages (8KB to 64KB). Alpha's extendedA virtual address range allows direct manipulation of gigabytes andrE terabytes produced in electrical and mechanical design, database, andt# transaction processing, and imagings  B Second, Alpha is a RISC architecture; VAX is a CISC architecture."H ________________________________________________________________________  A From: Migrating an Application from OpenVMS VAX to OpenVMS Alpha,iC Software Version: OpenVMS Alpha Version 7.0 and OpenVMS VAX Version ' 7.0, DEC, 1995, Page 4-1, section 4.1.1i  E "There are several issues to consider when planning what hardware youaC will need for your migration. To begin, consider what resources area4 required in your normal VAX development environment: 	CPUx 	Disksh 	Memory  D To estimate the resources needed for an Alpha migration environment, consider the following: * 	Greater image size on the Alpha systems"  H ________________________________________________________________________    = Image size on the Alpha is bigger than on a VAX, again due to @ difference in RISC and CISC architecture. How big the image sizeE growth will be may be dependent upon whether the full VAX instruction B set used by the VAX 9000 or the Reduce instruction set used by the$ 6000-7000 VAXes are being converted.  F By increasing the page size, you will be able to map larger amounts ofD physical memory. With increase images size and memory, the increased4 page size will help load the images and data faster.  F The VAX/VMS was the first hardware and software design with each otherB in mind. Therefore, hardware and software memory mapping was based upon 512-byte page.H  < The disk system (ODS2, ODS5) still uses 512 bytes per block.   ------------------------------  % Date: Fri, 25 Oct 2002 21:01:10 -0400 * From: "Bill Todd" <billtodd@metrocast.net> Subject: Re: page or pagelet2 Message-ID: <n6ycnW6ZZaXZdCSgXTWcrg@metrocast.net>  D "Daryl Jones" <jones.computer.srv@worldnet.att.net> wrote in message7 news:8a646952.0210251344.705a9554@posting.google.com...e   ...i   > To Everyone: >_F > From Alpha Architecture Reference Manual, Digital Press, 1992, pages > vii-viii.i >dF > "Alpha is designed to handle the largest computing problems of today> > and tomorrow. When the Alpha architecture is compared to its> > predecessor, the VAX architecture, two differences stand outE > immediately. First, Alpha is a 64-bit architecture; VAX is a 32-biteD > architecture. This means that the Alpha virtual address extends toB > 64-bit linear range of bytes in memory. Supporting this extendedF > virtual address space are an extended maximum physical address rangeB > (up to 48 bits) and larger pages (8KB to 64KB). Alpha's extendedC > virtual address range allows direct manipulation of gigabytes and G > terabytes produced in electrical and mechanical design, database, andt% > transaction processing, and imagingr >mD > Second, Alpha is a RISC architecture; VAX is a CISC architecture."  E I'm not sure why you felt compelled to quote the above, since it addsr! exactly nothing to your argument.3  J > ________________________________________________________________________ >cC > From: Migrating an Application from OpenVMS VAX to OpenVMS Alpha,sE > Software Version: OpenVMS Alpha Version 7.0 and OpenVMS VAX Version ) > 7.0, DEC, 1995, Page 4-1, section 4.1.1n > G > "There are several issues to consider when planning what hardware you E > will need for your migration. To begin, consider what resources areb6 > required in your normal VAX development environment: >  CPUo	 >  Diskst
 >  Memory >CF > To estimate the resources needed for an Alpha migration environment, > consider the following: , >  Greater image size on the Alpha systems"  K No shit.  However, no one in any way suggested that image sizes didn't groweL somewhat in the migration to RISC:  we just pointed out that your contention0 that they grew by a factor of 5 - 6 was rubbish.   >iJ > ________________________________________________________________________ >i >s? > Image size on the Alpha is bigger than on a VAX, again due togB > difference in RISC and CISC architecture. How big the image sizeG > growth will be may be dependent upon whether the full VAX instructionaD > set used by the VAX 9000 or the Reduce instruction set used by the& > 6000-7000 VAXes are being converted. >eH > By increasing the page size, you will be able to map larger amounts ofF > physical memory. With increase images size and memory, the increased6 > page size will help load the images and data faster.  K Again, no shit.  For that matter, even *without* any increase in image size E or memory, an increased page size will help load images faster.  As IrF pointed out, the reason pages were small on VAX was because memory wasG expensive, not because in some way small pages were better for slightly  smaller images.   K Perhaps you should spend more time endeavoring to understand what you read.t   - bill   ------------------------------   Date: 26 Oct 2002 05:00:19 GMT- From: djweath@attglobal.net (Dave Weatherall)n Subject: RE: page or pagelet5 Message-ID: <DTiotGxQ0bj6-pn2-oje4SvoduuCO@localhost>-  F On Thu, 24 Oct 2002 22:39:40 UTC, "Tom Linden" <tom@kednos.com> wrote:    F > >>A page on an Alpha VMS machine is 8192 bytes per page and on a VAXF > >>machine it is 512 bytes per page. This is due to RISC vs CISC chipF > >>technology where the image sizes grew to be 5-6 time larger on the > >>RISC chip. > >M# > >No it's not!  Who told you this?n > E > I think it is probably close to true.  When you write compilers youtI > have occassion to to look at things like this and I think it is obviousoG > to anyone who has coded assembly on VAX and Alpha that it takes a lotuD > more instructions to accomplish the same task.  In the case of ourJ > PL/I compiler, on Alpha it is more than 7 times larger.  Of course, thisH > uses the GEM backend versus VCG, so I think the 5-6 figure is probablyJ > fairly accurate.  If you compare the sizes of DECC you will find about aK > 9:1 ratio, but again, part of that may be due to the much larger backend.e   It's chicken and egg time.  < RISC generally generates more instructions which means mem.  requirements go up.eB Memory prices come down because of demand/technology improvements.F Page sizes go up to reflect the 'standard' presence of more memory, to$ get the benefits that Bill spoke of.   or   mem prices come downB RISC makes more commercial sense in the context of overall system  price. r   -- i Cheers - Dave.   ------------------------------  % Date: Fri, 25 Oct 2002 22:41:48 -0400e2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: page or pageletK Message-ID: <rdeininger-2510022241490001@1cust118.tnt2.nashua.nh.da.uu.net>r  3 In article <lj1T2WDVTWnd@eisner.encompasserve.org>,h< kaplow_r@eisner.encompasserve.org.TRABoD (Bob Kaplow) wrote:    L >> For those people who aren't familiar with system calls, try the following >> lexical example in DCL: >> t* >> $write sys$output f$getsyi("PAGE_SIZE") >oG >Has DecQHP ever implemented an Alpha with a page size other than 8192?a  G EV7 is designed for 8 kB or 64 kB pages.  IIRC, EV8 was planned to havea only 64 kB pages.t  E I doubt VMS will use 64 kB pages on EV7.  Had EV8 been completed, VMSmJ would probably have used EV7 systems to verify operation with 64 kB pages,  to be ready when EV8 came along.   ------------------------------  % Date: Fri, 25 Oct 2002 23:47:03 -0500d( From: Rich Jordan <duodec@speakeasy.net> Subject: Re: PCL image1 Message-ID: <7bidnbF_pNEXjSegXTWcpQ@giganews.com>u  F Some wintel peecee print drivers have the ability to 'print to file'. H You might look at the drivers that HP provides for its PCL printers, or E the ones that come with windows; I know I've seen the option but I'm nE blessedly far away from the work peecee right now.  Get the image in sH whatever format, as long as you can display it on a peecee in a program F that can print it to the PCL printer, and tell the driver to print to D disk.  You will likely get an iniefficient (read large/bloated) PCL B file, but it should work.  If you use a HPIII/IIID driver it will , _probably_ work on nearly any later printer.       Lucas, Edward A (SAIC) wrote:F > Hello  >  aK > I have a site on NY, working on a NT system.  They scan a document (with b+ > Corel Draw) that I will use for printing.Q >  3C > My library on Alpha is a PCL library.  I really need a PCL image.m >  tJ > If I am not mistaken, when you print an image from a PC to a file a PCL  > image is created.  Hmmmmmm >  :H > Now If I am correct, what format does the site in NY need to scan the  > document in? >   J > I requested a PCL image and so far I received a JPG (will never work on J > an Alpha), a EPS format (will only work as a postscript file and a DCPS  > printer) and a .BMP.H > I cannot make any of these to work.  I bring them across as ASCII and # > Binary and have the same problem.  >  25 > Does anyone have any information that will help me?d >  .   ------------------------------    Date: 25 Oct 2002 15:01:58 -07007 From: stephen.mcelduff@transcore.com (Stephen McElduff) - Subject: Quorum disk appears to not be votinga= Message-ID: <1a941731.0210251401.2e518336@posting.google.com>S  D I am trying to configure a 2 Node DecNet based cluster with a quorumE disk. As best I can tell, I have set all of the sysgen parameters (inmD the modprarams.dat) files on each system such that DKA1 should be my quorum disk.  D I have configured it to give each node a single vote and to give the? quorum disk a single vote, so that if any single node dies, theeE cluster will transition but remain available (after a transition) forf my apps....r    E I believe that my quorum disk is not contributing a vote (see detailse below)    ; A cut/paste from my modparams.dat file follows: (the votes,JA expected_votes, disk_quorum values are the same on both machines)    VOTES=1  EXPECTED_VOTES=3: AGEN$INCLUDE_PARAMS SYS$MANAGER:AGEN$NEW_NODE_DEFAULTS.DAT SCSNODE="MCTMC1" SCSSYSTEMID=1536 NISCS_LOAD_PEA0=1l VAXCLUSTER=2 DISK_QUORUM="DKA1" QDSKVOTES=1N
 ALLOCLASS=200O INTERCONNECT="NI"o BOOTNODE="N"    > When I either shutdown or disconnect one of the nodes from theA network, the other system "hangs" and doesn't come back until thed other system gets booted again.-  F I use the Show Cluster Utility to show me the current contents of someF values relating to this. Most intriguing is the fact that the value of2 the QF_VOTE parameter is "NO" for the Quorum disk.  E I don't understand what I am missing or have wrong that's not lettingc the disk contribute a vote ?l  A Also strange is the value of the parameter named QF_SAME. For onelC system it is listed with a value of YES, for the other it is listedo with a value of NO.e  B When both systems are online the value of CL_VOTES is 2 (one wouldB expect it to be 3). The value of CL_EXP is 3, as one would expect.  u            > In each systartup_vms.com file, I mark the disk on as follows:    * mount/system/cluster $2000$DKA1: LABELNAME      ? I believe I followed the instructions exactly as in the OpenVMSnD Cluster Systems Manual relating to how to get a disk to have a vote.  F I checked for and found a quorum.dat file in the root directory of the disk drive in question (dka1).    0 Any help would be greatly appreciated..thanks...   ------------------------------  # Date: Fri, 25 Oct 2002 20:39:40 GMTs$ From: winston@SSRL.SLAC.STANFORD.EDU) Subject: Re: So I went to the HP IT forum 8 Message-ID: <00A15FB4.6EB8DC2D@SSRL04.SLAC.STANFORD.EDU>  b In article <apbkvn$5441$1@ID-46415.news.dfncis.de>, "James Gessling" <jgessling@yahoo.com> writes:A >Yesterday, October 24, Pleasanton, CA. (San Francisco Bay area).  >sE >I came away quite pleased.  Maybe this VMS stuff has a future.  Mary ? >Jane Vasquez, who works for Mark Gorham as VMS product managerwG >gave a very upbeat talk and we had plenty of time for discussion.  ShetB >seemed very open to comments and suggestions about what direction( >VMS should take.  That was after lunch.  L David Mathog may or may not be interested to know that I brought up what areM basically his objections about the EDU license (vs. the OPENVMSEDU FAQ, whichrI says the license is to be interpreted in ways not supported by the actualpM words) and the default speed of stuff in RMS (coupled with a suggestion abouteF changing to less-safe defaults on a per-process basis, although other K system managers in the room said I should just be using XFC).  Anyway, MarycO Jane (maryjane.vazquez@hp.you-know-what) says the EDU thing is now her baby andeG she'd love to hear any suggestions for getting VMS back in the schools.    >mD >Lowlight was the Intel guy before lunch.  Trying to make the 8 year8 >process of getting Itanium out the door a good thing.    / That was very obvious spin; I wasn't impressed.t  
 >And slammingu< >AMD's hammer chip with snide comments about itanium being a1 >whole new thing vs. an evolutionary development.   I Yeah.  His main point was "Intel has a lot of money and we intend to keepe; spending until this is good enough to take over the world."n   > : >My main comment was that HP needs to get the whole tcp/ip< >software thing figured out.  Having three competing stacks,A >with different features is just a mess.  Especially when two arer7 >owned by the same company, process software.  Maybe HPo? >should give their tcp services to process and let them work itt& >out, they seem to do a very good job.  M I agree Process does a great job.  I'd hate to see HP drop TCP/IP developmentk* now that their stack doesn't suck, though.   -- Alanl   ------------------------------    Date: 25 Oct 2002 17:09:08 -0700+ From: seanobanion@attbi.com (Sean O'Banion)r) Subject: Re: So I went to the HP IT forumd= Message-ID: <f883d5a4.0210251609.704cfd51@posting.google.com>u  2 I have to agree about May Jane, and the Intel Guy.  F I congratulated them on moving the COE work that was talked about lastD year at the IT Forum into UNIX Portability.  I pointed out that UNIXF Portability needs a "killer app", something major that would show whatE it took to port an application, and I proposed Oracle Applications asmD a possible example: Mary Jane responded that they are always talkingB to Oracle.  Since she didn't even write this comment down (she didA write down the other comments), I took this to mean that this hash> occurred to them (well, duh), but she could not comment on it.  B Can anybody comment about the UNIX presentation that Jan Mark did?  E Stephen De Luco, who set up this IT Forum, hoped to repeat this ForumnD next year, which everyone seemed to agree with.  With budgets tight,E so that travel to HP ETS might be difficult, this kind of event givesu@ customers a similar opportunity to talk with directly someone in$ OpenVMS Engineering, like Mary Jane.  C I believe that both this IT Forum and HP ETS are essential parts of.@ listening to us customers, and I hope they keep them both going.     Sean  g "James Gessling" <jgessling@yahoo.com> wrote in message news:<apbkvn$5441$1@ID-46415.news.dfncis.de>...rB > Yesterday, October 24, Pleasanton, CA. (San Francisco Bay area). > F > I came away quite pleased.  Maybe this VMS stuff has a future.  Mary@ > Jane Vasquez, who works for Mark Gorham as VMS product managerH > gave a very upbeat talk and we had plenty of time for discussion.  SheC > seemed very open to comments and suggestions about what directionh) > VMS should take.  That was after lunch.s > E > Lowlight was the Intel guy before lunch.  Trying to make the 8 year E > process of getting Itanium out the door a good thing.  And slamminge= > AMD's hammer chip with snide comments about itanium being a 2 > whole new thing vs. an evolutionary development. > ; > My main comment was that HP needs to get the whole tcp/ip = > software thing figured out.  Having three competing stacks,7B > with different features is just a mess.  Especially when two are8 > owned by the same company, process software.  Maybe HP@ > should give their tcp services to process and let them work it' > out, they seem to do a very good job.  >  > Jimn   ------------------------------   Date: 26 Oct 2002 00:42:55 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)) Subject: Re: So I went to the HP IT forumd4 Message-ID: <apcoef$fkvk$1@ID-135708.news.dfncis.de>  8 In article <00A15FB4.6EB8DC2D@ssrl04.slac.stanford.edu>,' 	winston@SSRL.SLAC.STANFORD.EDU writes:  > N > David Mathog may or may not be interested to know that I brought up what areO > basically his objections about the EDU license (vs. the OPENVMSEDU FAQ, which K > says the license is to be interpreted in ways not supported by the actual O > words) and the default speed of stuff in RMS (coupled with a suggestion about H > changing to less-safe defaults on a per-process basis, although other M > system managers in the room said I should just be using XFC).  Anyway, Mary8Q > Jane (maryjane.vazquez@hp.you-know-what) says the EDU thing is now her baby and I > she'd love to hear any suggestions for getting VMS back in the schools.3 >   N Well, I'm all for keeping it as simple as possible (for both sides concerned).  E Should I email her suggesting that they write up something that says 0E you can use Hobbyist PAKS and use the systems for anything related to0E academics,  Of course, it would be better if a paralel system was setFB up so they could track how much academic use (or at least how manyD academic licenses were applied for) there actually was.  It would beI useful if they actually saw this number going up ove some period of time. I But, I imagine the more this costs the less likely it is to actually come  about.   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>       ------------------------------    Date: 25 Oct 2002 19:32:39 -07001 From: susan_skonetski@hotmail.com (Sue Skonetski)0) Subject: Re: So I went to the HP IT forum = Message-ID: <857e9e41.0210251832.16f24298@posting.google.com>    Jim,  L Thank you for your feedback on this event.  I have forwarded to Mark and MJ.  
 Warm Regards,  sue     g "James Gessling" <jgessling@yahoo.com> wrote in message news:<apbkvn$5441$1@ID-46415.news.dfncis.de>...mB > Yesterday, October 24, Pleasanton, CA. (San Francisco Bay area). > F > I came away quite pleased.  Maybe this VMS stuff has a future.  Mary@ > Jane Vasquez, who works for Mark Gorham as VMS product managerH > gave a very upbeat talk and we had plenty of time for discussion.  SheC > seemed very open to comments and suggestions about what direction ) > VMS should take.  That was after lunch.4 > E > Lowlight was the Intel guy before lunch.  Trying to make the 8 year0E > process of getting Itanium out the door a good thing.  And slamming = > AMD's hammer chip with snide comments about itanium being a 2 > whole new thing vs. an evolutionary development. > ; > My main comment was that HP needs to get the whole tcp/ip = > software thing figured out.  Having three competing stacks, B > with different features is just a mess.  Especially when two are8 > owned by the same company, process software.  Maybe HP@ > should give their tcp services to process and let them work it' > out, they seem to do a very good job.  >  > Jim    ------------------------------  % Date: Fri, 25 Oct 2002 22:48:45 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)) Subject: Re: So I went to the HP IT forumTK Message-ID: <rdeininger-2510022248450001@1cust118.tnt2.nashua.nh.da.uu.net>u  L In article <apcoef$fkvk$1@ID-135708.news.dfncis.de>, bill@cs.uofs.edu wrote:  9 >In article <00A15FB4.6EB8DC2D@ssrl04.slac.stanford.edu>,e/ >        winston@SSRL.SLAC.STANFORD.EDU writes:o >> !O >> David Mathog may or may not be interested to know that I brought up what are1P >> basically his objections about the EDU license (vs. the OPENVMSEDU FAQ, whichL >> says the license is to be interpreted in ways not supported by the actualP >> words) and the default speed of stuff in RMS (coupled with a suggestion aboutI >> changing to less-safe defaults on a per-process basis, although other  N >> system managers in the room said I should just be using XFC).  Anyway, MaryI >> Jane (maryjane.vazquez@hp.you-know-what) says the EDU thing is now her  baby andJ >> she'd love to hear any suggestions for getting VMS back in the schools. >> i >hO >Well, I'm all for keeping it as simple as possible (for both sides concerned).I >uF >Should I email her suggesting that they write up something that says F >you can use Hobbyist PAKS and use the systems for anything related to >academics,   I Yes, it sounds like a good time to explain your objections to the currentpG scheme, and toss in any suggestions you have for improving the program.h  H IIRC, you had your campus lawyer look at the license terms, and he had aJ number of objections.  I you still have that discussion at hand, you mightH as well send it along.  Or maybe my memory is playing tricks, and you're someone else.  :-)   ------------------------------  # Date: Fri, 25 Oct 2002 20:17:35 GMTa( From: Alder <PGDEHMKOKIMD@spammotel.com>- Subject: Re: TCPIP$FTP_SERVER startup problemw* Message-ID: <3DB9A68F.90800@spammotel.com>   Carl Karcher wrote: A > In a previous article, Alder <PGDEHMKOKIMD@spammotel.com> wrote>H > ->I'm having a similar problem trying to configure Anonymous FTP with M > ->TCP/IP Services for OpenVMS 5.0 on OpenVMS Alpha 7.2 (the hobbyist kit).  J > ->   Anonymous logins are refused because of a privilege issue, I think. > -> eM > ->I configured it from the menus in TCPIP$CONFIG and start the server from s > ->the SYSTEM account with: > ->  ! > ->$ @TCPIP$SYSTEM:TCPIP$FTP_RUNr > J > That's what the INETACP process runs in response to a connection on port* > 21. This is _not_ how to start manually. > H > If you do TCPIP$SHUTDOWN followed by TCPIP$STARTUP that should installH > the necessary images with the proper privs. As previously mentioned inI > this thread, it's TCPIP$SERVICE_SETUP.COM (called by TCPIP$STARTUP.COM)rH > that sets up the services for TCPIP V5.0. In TCPIP V5.1 and V5.3, eachF > service has a separate setup procedure (e.g. TCPIP$FTP_STARTUP.COM). >   G Thanks, Carl.  I think it's finally percolated through the concrete :-)6  H I was confused (obviously) by the fact that anonymous logins were being G rejected for password violations.  I assume then that passwords should nI be disabled for this account(?)  If that's not a good idea, or there are sF any other problems lurking in this account setup, I'd appreciate your  comments.  Cheers.  < Username: ANONYMOUS                        Owner:  ANONYMOUS< Account:  ANONY                            UIC:    [3375,1]  ([ANONY,ANONYMOUS]) < CLI:      DCL                              Tables: DCLTABLES# Default:  SYS$SYSDEVICE:[ANONYMOUS]  LGICMD:   _NL:G Flags:  DisCtlY DefCLI LockPwd Restricted DisWelcome DisNewMail DisMail G          DisReport DisReconnect DisForce_Pwd_Change DisPwdDic DisPwdHisg# Primary days:   Mon Tue Wed Thu Frio+ Secondary days:                     Sat SunoF Primary   000000000011111111112222  Secondary 000000000011111111112222F Day Hours 012345678901234567890123  Day Hours 012345678901234567890123F Network:  ##### Full access ######            ##### Full access ######F Batch:    -----  No access  ------            -----  No access  ------F Local:    -----  No access  ------            -----  No access  ------F Dialup:   -----  No access  ------            -----  No access  ------F Remote:   -----  No access  ------            -----  No access  ------; Expiration:   (none)    Pwdminimum:  6   Login Fails:    12e5 Pwdlifetime:  (none)    Pwdchange:  25-OCT-2002 13:02 G Last Login:   (none) (interactive), 25-OCT-2002 12:43 (non-interactive)r9 Maxjobs:         0  Fillm:        50  Bytlm:        52200i9 Maxacctjobs:     0  Shrfillm:      0  Pbytlm:           0 9 Maxdetach:       0  BIOlm:        18  JTquota:       4096R9 Prclm:           8  DIOlm:        18  WSdef:          350e9 Prio:            8  ASTlm:       100  WSquo:          512.9 Queprio:         4  TQElm:        15  WSextent:       512p9 CPU:        (none)  Enqlm:       100  Pgflquo:      10240a Authorized Privileges:    NETMBX       TMPMBX Default Privileges:     NETMBX       TMPMBX    3 $ SHOW SECURITY SYS$SYSDEVICE:[000000]ANONYMOUS.DIRo  : SYS$SYSDEVICE:[000000]ANONYMOUS.DIR;1 object of class FILE       Owner: [SYSTEM]f=       Protection: (System: RWE, Owner: RWE, Group: RE, World)f       Access Control List:  oC (IDENTIFIER=[ANONY,ANONYMOUS],OPTIONS=PROTECTED+NOPROPAGATE,ACCESS=             READ+EXECUTE)   ------------------------------   Date: 26 OCT 2002 00:34:37 GMT2 From: karcher@kort.waisman.wisc.edu (Carl Karcher)- Subject: Re: TCPIP$FTP_SERVER startup problemm2 Message-ID: <26OCT02.00343751@kort.waisman.wisc.e>  @ In a previous article, Alder <PGDEHMKOKIMD@spammotel.com> wrote: ->  J ->I was confused (obviously) by the fact that anonymous logins were being I ->rejected for password violations.  I assume then that passwords should  K ->be disabled for this account(?)  If that's not a good idea, or there are hH ->any other problems lurking in this account setup, I'd appreciate your  ->comments.  Cheers.  ; Yes, TCPIP$CONFIG does setup the account with /nopassword. r   --G -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madisont5 --              karcher.nomorespam@waisman.wisc.edu  s   ------------------------------  % Date: Fri, 25 Oct 2002 21:05:48 -0000 - From: wspencer@ap.nospam.org (Warren Spencer)h Subject: Re: TPU port to Linux5 Message-ID: <92B2A064Cwarrenspencer1977@216.168.3.30>   7 Kesav_Tadimeti@KeaneIndia.com (Kesav Tadimeti) wrote inn? <8EA11405E59BD611BA7100104B93C2606E167A@exdel01.del.mgsl.com>: ,   >Hi all,/ >Is there a port of EVE or EDT for Unix/Linux. 0I >Using Emacs / Vi is very cumbersome.  BTW, do Digital unix users have tof >use Vi or do they use EVE?e >e >Thanks & regards, >Tadimeti Kesavh >KEANE INDIA Ltd.  >E9 - E12, SDF >NEPZt >NOIDA - 201 305 >U.P, INDIAh >r  >Telefon: +91-120-456 8210 (211)& >e-mail: kesav_tadimeti@keaneindia.com >s  H I'm using emacs on Linux, but I specify "edt-emulation-on" in my .emacs I file.  This provides EDT keypad emulation, but little else in the way of iD compatibility.  For my needs though, it's close enough to be usable.   ws   -- u   Warren Spencer' Senior Software Engineer (not a writer)u The Associated Press  < ** Time flies like an arrow.  Fruit flies like a bananna. **   ------------------------------  % Date: Fri, 25 Oct 2002 16:19:48 -0400i- From: "kenrbnsn1@rcn.com" <kenrbnsn1@rcn.com> ! Subject: TSM on the Freeware diskm< Message-ID: <168270-2200210525201948789@M2W041.mail2web.com>  F I'm doing a short term contract installing a VAX 3100/85 to eventuallyK replace a VAX 3100/20=2E They have a decserver 700 currently attached=2E I=s t'seE been a long time since I've played with a decserver (any model), so ISG thought I would try to use the TSM from the Freeware (CD & Web area)=2EgF Downloading from the Web or using the files off the CD, gives the same results:   $ r tsm021=2Edcx_vaxexeu  K                 FTSV DCX auto-extractible compressed file for OpenVMS (VAX=a )n8                 FTSV V3=2E0 -- FTSV$DCX_VAX_AUTO_EXTRACT<                 Copyright (c) Digital Equipment Corp=2E 1993  ? Options: [output_file_specification [input_file_specification]]0  ? The decompressor  needs to know  the filename to use for the=20e> decompressed file=2E If you don't specify any, it will use the< original name  of the  file before it  was  compressed,  and> create  it in  the  current  directory=2E  If  you  specify  a= directory name, the file will be created in that directory=2Ee  % Decompress into (file specification):l5         Opening and checking compressed file=2E=2E=2Eh< Decompressing (press Ctrl-T to watch the evolution)=2E=2E=2E+         Creating decompressed file=2E=2E=2EtG         Original file specification: SYS$SYSROOT:[SYSMGR]TSM021=2ESAV;1 (         Decompressed file specification:) DKA100:[KSROPR=2EDOWNLOADS]TSM021=2ESAV;140 %FTSV-W-COMP_DECOMPERR, fatal decompressor errorK -DCX-E-INVDATA, invalid data presented for expansion or bounded compressio=4 n5    H Is there a different place I could get a good copy? (or hints on getting into the decserver manually?)A   Thanks Kenm  D --------------------------------------------------------------------+ mail2web - Check your email from the web at: http://mail2web=2Ecom/ =2E   ------------------------------  % Date: Fri, 25 Oct 2002 17:05:17 -0400 & From: David M Smith <dsmit115@csc.com>% Subject: Re: TSM on the Freeware diskj8 Message-ID: <eacjru8d2kiunvc7aqdif13a45rs1up1cg@4ax.com>  K On Fri, 25 Oct 2002 16:19:48 -0400, "kenrbnsn1@rcn.com" <kenrbnsn1@rcn.com>n wrote:  G >Downloading from the Web or using the files off the CD, gives the same<	 >results:t  1 >%FTSV-W-COMP_DECOMPERR, fatal decompressor error0L >-DCX-E-INVDATA, invalid data presented for expansion or bounded compression  I >Is there a different place I could get a good copy? (or hints on gettingg >into the decserver manually?)   Try the copy at:  D 	http://www.compaq.com/support/digital_networks_archive/servers/tsm/  P The one named TSMECO07021_A-DCX_VAX.EXE worked for me. (I was able to decompress0 it without error, I did not try to install it!):  ! $ run TSMECO07021_A-DCX_VAX.EXE;1   K                 FTSV DCX auto-extractible compressed file for OpenVMS (VAX) 6                 FTSV V3.0 -- FTSV$DCX_VAX_AUTO_EXTRACT:                 Copyright (c) Digital Equipment Corp. 1993 	...6 Decompressing (press Ctrl-T to watch the evolution)...%         Creating decompressed file... G         Original file specification: DISK13:[P_MCSKEANE]TSMECO07021.A;1-B         Decompressed file specification: disk:[dir]TSMECO07021.A;17 Successful decompression, decompression report follows: A         File Size: 14266.03 Blocks, 7133.01 Kbytes, 7304207 bytes >         Decompression ratio is 1 to 1.83 ( 82.83 % expansion )*         Elapsed CPU time:    0 00:00:26.87*         Elapsed time    :    0 00:00:29.00M         Speed : 53961.92 Blocks/min, 26980.96 Kbytes/min, 460475.03 bytes/secy
 $ dir tsm*   Directory disk:[dir]  L TSMECO07021.A;1        26082/26082    14-AUG-1998 14:19:52.05  (RWED,RWED,,) TSMECO07021_A-DCX_VAX.EXE;1oL                        14267/14274    25-OCT-2002 17:00:54.16  (RWED,RWED,,)  % Total of 2 files, 40349/40356 blocks. I ------------------------------------------------------------------------- I David M. Smith 302.391.8533                       dsmit115 at csc dot comtI Computer Sciences Corporation     (Opinions are those of the writer only)nI -------------------------------------------------------------------------s   ------------------------------  # Date: Sat, 26 Oct 2002 01:04:57 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network>d% Subject: Re: TSM on the Freeware diski: Message-ID: <ZSlu9.1649$Zx.269730@news1.news.adelphia.net>   kenrbnsn1@rcn.com wrote:H > I'm doing a short term contract installing a VAX 3100/85 to eventuallyK > replace a VAX 3100/20. They have a decserver 700 currently attached. It's G > been a long time since I've played with a decserver (any model), so IaG > thought I would try to use the TSM from the Freeware (CD & Web area). H > Downloading from the Web or using the files off the CD, gives the same
 > results:  F Downloading from the HP Freeware link at http:/www.openvms.compaq.com ' will work, I recently tried it my self.   H The kit on the CD-ROMs, and any site that is mirroring the CD-ROM as-is 
 is broken.  B For the license keys, you will need to go to the Digital Networks  Product Group web site.r  3 A search for TSM on their web site shows a link to:(  * http://digitalnetworks.net/dr/npg/tsm.html  D But the link to the License Keys, and readme files is getting a 404  error from their server.  C You will need to contact them to get a license key for the product.d  G There is also on the freeware cd-rom a directory of tools I had a hand AD in writing that can manipulate the TSM product to do semi-automatic ! backups of your Terminal Servers.i  H I do not know if they support a DS700, but they do support DS 200, 300, B 90, and some Lantronix models.  It's all DCL, so you can fix them.  G I no longer have resonsibilities for DecServers, so I can not test any  < updates, but anyone can submit stuff to the freeware CD-ROM.  G As one alternative, there is a program in the DECUS archives that will lF also remotely manage terminal servers.  I do have experience using it.  G The NCP CONNECT NODE command in phase IV decnet allows you to access a gE terminal server over the network.  Any terminal emulator that allows eB scripting can be used on it, or directly to the console port of a % terminal server to do the management.i   -Johnp wb8tyw@qsl.netowrk Personal Opinion Onlyi   ------------------------------  % Date: Fri, 25 Oct 2002 13:42:15 -0400e From: Ray <lists@atc.edu>eP Subject: Why did the VMS programmers confuse Halloween with Christmas this year?& Message-ID: <3DB98277.1F4574D@atc.edu>   Because:   OCT 31 = DEC 25r   ------------------------------   Date: 25 Oct 2002 18:58:04 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)T Subject: Re: Why did the VMS programmers confuse Halloween with Christmas this year?4 Message-ID: <apc47s$9sht$1@ID-135708.news.dfncis.de>  & In article <3DB98277.1F4574D@atc.edu>, 	Ray <lists@atc.edu> writes:
 > Because: >  > OCT 31 = DEC 25r  @ I thouught that was the reason why old time programmers gave out@ presents on Halloween.  And why would this year be any different than any other??  :-)o   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>   e   ------------------------------  % Date: Fri, 25 Oct 2002 20:52:14 +0000n2 From: John Eisenschmidt <jweisen@eisenschmidt.org>T Subject: Re: Why did the VMS programmers confuse Halloween with Christmas this year?4 Message-ID: <20021025205214.A22931@eisenschmidt.org>  : Unless the Voices are Mistaken, Ray (lists@atc.edu) Wrote:
 > Because: >  > OCT 31 = DEC 25>   ::deep breath::    BOO!  ? And I thought "Only COBOL programmers understand why women haten periods" was bad.    -- l/ John W. Eisenschmidt <jweisen@eisenschmidt.org>0C    Public Key   |  http://www.eisenschmidt.org/jweisen/misc/pgp.asc D    Fingerprint  |  5F9B F916 5AD1 3295 CF99 BC1E 1F97 E6A3 37E3 BEF2J Is this mail an attachment? http://www.jensbenecke.de/misc/outlook.en.html  C "Don't you want a Fanta?" --  http://www.fantanas.com/fantanas.html    ------------------------------   End of INFO-VAX 2002.591 ************************