1 INFO-VAX	Fri, 22 Feb 2002	Volume 2002 : Issue 104       Contents:5 Re: 500au ethernet apparently won't operate at 100bps 5 Re: 500au ethernet apparently won't operate at 100bps " Backup/Restore Verification Errors& Re: Backup/Restore Verification Errors& Re: Backup/Restore Verification Errors& Re: Backup/Restore Verification Errors& Re: Backup/Restore Verification Errors& Re: Backup/Restore Verification Errors. Re: Cisco 2948G-L3 layer three switch with VMSL Re: Create an Open Mail Relay (GASP!) and Turn it Into a Relay Spam HoneypotP Create an Open Mail Relay (GASP!) and Turn it Into a Relay Spam Honeypot (Yay!)  Re: DCL procedure's / Re: DECWindows and the MIT-SHM server extension  Re: Einstein Re: Einstein Re: einstein RE: Einstein, Re: EV8 and McKinley analysis by Paul DeMone, Re: EV8 and McKinley analysis by Paul DeMone, Re: EV8 and McKinley analysis by Paul DeMone, Re: EV8 and McKinley analysis by Paul DeMone, Re: EV8 and McKinley analysis by Paul DeMone, Re: EV8 and McKinley analysis by Paul DeMoneJ Re: FreeVMS: what are CPQ's feelings likely to be ?, was: Re: MPE/iX users= Re: Help on creating a detached process using SYS$CREPRC(...) = Re: Help on creating a detached process using SYS$CREPRC(...) = Re: Help on creating a detached process using SYS$CREPRC(...)  Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles Re: Itanium troubles' Monitor lock and monitor io questions ! 8 Re: Monitor lock and monitor io questions ! - correction OpenVMS IPF presentation Re: OpenVMS IPF presentation Re: OpenVMS IPF presentation Re: OpenVMS IPF presentation OT - Programmer war stories 0 re: OT gassless cars Was Re: "Crashless Windows"0 Re: OT gassless cars Was Re: "Crashless Windows"0 Re: OT gassless cars Was Re: "Crashless Windows"0 Re: OT gassless cars Was Re: "Crashless Windows"0 Re: OT gassless cars Was Re: "Crashless Windows" Re: PGP for OpenVMS? Re: PGP for OpenVMS? Re: PGP for OpenVMS?= Re: Poor Performance with OpenVMS Advanced Server and TCPWare = Re: Poor Performance with OpenVMS Advanced Server and TCPWare = Re: Poor Performance with OpenVMS Advanced Server and TCPWare = Re: Poor Performance with OpenVMS Advanced Server and TCPWare = Re: Poor Performance with OpenVMS Advanced Server and TCPWare 4 Re: POSIX for OpenVMS - is there something similar ?4 Re: POSIX for OpenVMS - is there something similar ? Postscript printing  Re: Postscript printing  Re: Postscript printing ) Re: Profitability and the survival of VMS ) Re: Profitability and the survival of VMS A Re: Proof that EV8 was fantastic ... and Capellas "gave" it away! A Re: Proof that EV8 was fantastic ... and Capellas "gave" it away!  Re: Screen Saver Problem (?) Re: Screen Saver Problem (?) Re: Shannon on the merger  Re: Shannon on the merger  Re: Singular they  Re: Singular they  RE: Singular they L Re: Singular they was Re: Sv: Younger recruits versus experienced veterans (L Re: Singular they was Re: Sv: Younger recruits versus experienced veterans (P Re: Singular they was Re: Sv: Younger recruits versus experienced veterans ( was SOAP on OpenVMS? Re: SOAP on OpenVMS?7 Re: Strange status value - when can I expect to get it? 7 Re: Strange status value - when can I expect to get it? 7 Re: Strange status value - when can I expect to get it? 7 Re: Strange status value - when can I expect to get it? & Re: SYSPRV & $GETUAI - Dumb question ?$ Re: Using Edit inside a procedure...- Re: What is the best NFS package for OpenVMS? - Re: What is the best NFS package for OpenVMS?  Re: World Class Sportfishing!   F ----------------------------------------------------------------------    Date: 22 Feb 2002 06:48:49 -0800( From: bob@instantwhip.com (Bob Ceculski)> Subject: Re: 500au ethernet apparently won't operate at 100bps= Message-ID: <d7791aa1.0202220648.6798a2ed@posting.google.com>    David Spencer <spencer@pageweavers.spaamfree.com> wrote in message news:<210220021255190269%spencer@pageweavers.spaamfree.com>... N > I've got a 500au (running VMS 7.2-1). I've also got two network hubs. One isK > a 10baseT and the other is a 10/100 autosensing hub. The 500au works fine J > when attached to the 10baseT hub. When I connect it to the 10/100 hub itH > runs only at 10mbs and has lots of problems with traffic. I've got the@ > ewa0_mode set to Auto-Negotiate but it doesn't seem to matter. > L > I've tried switching ports and cables but it makes no difference. I've gotJ > another 500au that's similarly configured and it works at 100bps without > problems.  > M > Do I have a bad part someplace? I've compared the ewa0 settings between the K > two machines and there appears to be no difference between them. Is there L > are hardware jumper or something that I can check? I'd like to get maximum- > performance out of this thing if I could...  > $ > Thanks for any help and/or advice, >  > -- Dave Spencer, PageWeavers  H your problem can happen on any alpha ... first of all your nic should beF a DE500-BA as earlier versions may have issues and do not use a DE600!G Next, compaq support told us their may be problems using auto-negotiate J and you should always hard set ewa0_mode at the console level, this allowsF you to run 10 or 100 but not 100 full duplex as to do that you need toH use a managed switch and turn off spanning tree as auto switching is not	 reliable!    ------------------------------    Date: 22 Feb 2002 11:12:22 -06001 From: Graham Allan <allan@physics.umn.edu.nospam> > Subject: Re: 500au ethernet apparently won't operate at 100bps/ Message-ID: <w53k7t5igp5.fsf@crail.spa.umn.edu>   * bob@instantwhip.com (Bob Ceculski) writes:   > David Spencer <spencer@pageweavers.spaamfree.com> wrote in message news:<210220021255190269%spencer@pageweavers.spaamfree.com>... P > > I've got a 500au (running VMS 7.2-1). I've also got two network hubs. One isM > > a 10baseT and the other is a 10/100 autosensing hub. The 500au works fine L > > when attached to the 10baseT hub. When I connect it to the 10/100 hub itJ > > runs only at 10mbs and has lots of problems with traffic. I've got theB > > ewa0_mode set to Auto-Negotiate but it doesn't seem to matter. > J > your problem can happen on any alpha ... first of all your nic should beH > a DE500-BA as earlier versions may have issues and do not use a DE600!I > Next, compaq support told us their may be problems using auto-negotiate L > and you should always hard set ewa0_mode at the console level, this allowsH > you to run 10 or 100 but not 100 full duplex as to do that you need toJ > use a managed switch and turn off spanning tree as auto switching is not > reliable!   B As a matter of priciple I never trusted auto-negotiate, but I haveE also observed that while it never used to work correctly in the past, F in recent times it's become quite reliable.... whether due to firmware; updates in the workstations, or in the switches themselves.    Graham --  I -------------------------------------------------------------------------  Graham Allan9 School of Physics and Astronomy - University of Minnesota I -------------------------------------------------------------------------    ------------------------------  # Date: Fri, 22 Feb 2002 15:50:51 GMT ' From: Rick Dyson <Rick-Dyson@UIowa.EDU> + Subject: Backup/Restore Verification Errors ) Message-ID: <3C7668DB.5EF6573C@UIowa.EDU>   H I've been testing doing image backups to DLT, restoring to a spare disk,B and then comparing the original disk to the spare disk to test the integretity of the backup.  O I'm bothered by what appears to be hundreds and hundreds of verification errors   when I perform this exercise. :(  D I am using Alphas running OpenVMS v7.2-1 with all know ECOs applied.  E The initial backup command used to a TZ877 tape library (Compac III):   C $ Mount /Foreign /Cache = Tape_Data /BlockSize = 65535 MKC600: Blah E $ Backup /NoInitialize /Image /CRC /BlockSize = 65535 /Label = Blah - >          /Ignore = Interlock /List = BackUp_Logs:Foo-bar.log -!          DSA1: MKC600:FOO-Bar.bck   L The original disk is an RZ29B-VA dual shadow set between two nodes each with0 an HSZ40 controller where the disks are shelved.  , The restore backup and compare commands are:   $ Mount /For $1$DGA117: E $ Backup /Log /Image /Initialize /BlockSize = 65535 /Ignore = Label -           MKC600:Foo-Bar.bck - (          $1$DGA117: /By_Owner = Original $  $ Dismount $1$DGA117: - $ Mount /Override = Identification $1$DGA117: 7 $ Backup /Compare DSA1:[Blah...] $1$DGA117:[Blah...]      E The spare disk is a new 1" Compac 9 GB drive in an EMA12000 SAN off a 9 HSG80 controller via a Fiber Channel switch interconnect.   D I then proceed to get from 500-1500 of these kinds of error messages( (with different block numbers):         	           ; %BACKUP-E-VERIFYERR, verification error for block 237034 of  $1$DGA117:[Blah]DATA.DAT;1  D Is this OK?  Does it indicate that if I lost the original disks     G completely and had to restore from tape, that I would not have the same K data set on the restored volume?  Could it be due to the difference between L the original and target scratch drives (i.e., RZ29B-VA and Compaq 142671-B22 9 Gb Ulta3 10k)?  I I have tried this with several different savesets from different tapes (I N have always used the same scratch disk as the restore target), always with theN same kinds of negative results.  The amount of verification errors varies, butM is always at least 500 instances per 4 GB disk. (note: these source disks are ) typically 99% full of data in 1-5 files.)   H I would greatly appreciate any suggestions or insight you can provide onJ this situation!  I would realy like to be more confident of my backups. :)   Thanks! 
 Rick Dyson --  J Richard L. Dyson                                      rick-dyson@uiowa.eduJ  _   _  _____                      http://www-pi.physics.uiowa.edu/~dyson/J | | | ||_   _|  Senior Systems Analyst   --   INFORMM-Cerner Systems Group2 | | | |  | |    The University of Iowa Health CareJ | \_/ | _| |_   Health Care Information Systems BT1000 GH  O: 319/384-7016J  \___/ |_____|  Iowa City, IA 52242-1052                 FAX: 319/384-7020   ------------------------------  # Date: Fri, 22 Feb 2002 16:52:57 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>/ Subject: Re: Backup/Restore Verification Errors 0 Message-ID: <3C76758B.FFE69E9B@blueyonder.co.uk>   Rick Dyson wrote:  > J > I've been testing doing image backups to DLT, restoring to a spare disk,D > and then comparing the original disk to the spare disk to test the > integretity of the backup. > Q > I'm bothered by what appears to be hundreds and hundreds of verification errors " > when I perform this exercise. :( > F > I am using Alphas running OpenVMS v7.2-1 with all know ECOs applied. > G > The initial backup command used to a TZ877 tape library (Compac III):  > E > $ Mount /Foreign /Cache = Tape_Data /BlockSize = 65535 MKC600: Blah G > $ Backup /NoInitialize /Image /CRC /BlockSize = 65535 /Label = Blah - @ >          /Ignore = Interlock /List = BackUp_Logs:Foo-bar.log -# >          DSA1: MKC600:FOO-Bar.bck    Why no /VERIFY?    > N > The original disk is an RZ29B-VA dual shadow set between two nodes each with2 > an HSZ40 controller where the disks are shelved. > . > The restore backup and compare commands are: >  > $ Mount /For $1$DGA117: G > $ Backup /Log /Image /Initialize /BlockSize = 65535 /Ignore = Label -  >          MKC600:Foo-Bar.bck - * >          $1$DGA117: /By_Owner = Original   Again, why no /VERIFY?   > $  > $ Dismount $1$DGA117: / > $ Mount /Override = Identification $1$DGA117: 6 > $ Backup /Compare DSA1:[Blah...] $1$DGA117:[Blah...] > G > The spare disk is a new 1" Compac 9 GB drive in an EMA12000 SAN off a ; > HSG80 controller via a Fiber Channel switch interconnect.  > F > I then proceed to get from 500-1500 of these kinds of error messages! > (with different block numbers):  > = > %BACKUP-E-VERIFYERR, verification error for block 237034 of  > $1$DGA117:[Blah]DATA.DAT;1 >  > Is this OK?      No.   3 >Does it indicate that if I lost the original disks I > completely and had to restore from tape, that I would not have the same $ > data set on the restored volume?     Quite possibly.   * >Could it be due to the difference betweenN > the original and target scratch drives (i.e., RZ29B-VA and Compaq 142671-B22 > 9 Gb Ulta3 10k)?   I doubt it.    > K > I have tried this with several different savesets from different tapes (I P > have always used the same scratch disk as the restore target), always with theP > same kinds of negative results.  The amount of verification errors varies, butO > is always at least 500 instances per 4 GB disk. (note: these source disks are + > typically 99% full of data in 1-5 files.)   L Are you SURE nothing else is writing to DSA1 between writing of the tape andK comparing to the other disk? Is DSA1 mounted in a cluster? If so, check all K nodes to ensure no access. Best, dismout DSA1 and mount it readonly before  J doing the initial backup and only remount it writeable once the compare is completed.       > J > I would greatly appreciate any suggestions or insight you can provide onL > this situation!  I would realy like to be more confident of my backups. :)  O First, always use /VERIFY. You did not verify your disk to tape or tape to disk Q backups. It is possible that errors are being introduced at that stage. Are there 5 any disk or tape errors logged during the operations?   S Second, check that nothing else is writing to DSA1: during the process (see above).    regards    > 	 > Thanks!  > Rick Dyson > --L > Richard L. Dyson                                      rick-dyson@uiowa.eduL >  _   _  _____                      http://www-pi.physics.uiowa.edu/~dyson/L > | | | ||_   _|  Senior Systems Analyst   --   INFORMM-Cerner Systems Group4 > | | | |  | |    The University of Iowa Health CareL > | \_/ | _| |_   Health Care Information Systems BT1000 GH  O: 319/384-7016L >  \___/ |_____|  Iowa City, IA 52242-1052                 FAX: 319/384-7020   --   Tim.Llewellyn@cableinet.co.uk     C Standard disclaimer applies. My views in no way represent those of  ! my employers or service provider.    ------------------------------  % Date: Fri, 22 Feb 2002 12:36:33 -0500 - From: "Peter Weaver" <peter.weaver@stelco.ca> / Subject: Re: Backup/Restore Verification Errors 2 Message-ID: <Jkvd8.18384$X2.204015@nnrp1.uunet.ca>  4 "Rick Dyson" <Rick-Dyson@UIowa.EDU> wrote in message# news:3C7668DB.5EF6573C@UIowa.EDU...  >...J > I'm bothered by what appears to be hundreds and hundreds of verification errors" > when I perform this exercise. :(  G As others have pointed out, this should only happen if some process has @ written to the original disk between the backup and the restore.   >...6 > $ Backup /Compare DSA1:[Blah...] $1$DGA117:[Blah...] >...  H AFAIK you should not use [Blah...] when using /COMPARE. You have to walkH through directory by directory and compare them one by one. You probablyK knew that, but I'm just adding this for anyone else looking at your example  and thinking of trying it.  K Below is a .COM I hacked together for a disk replacement we are doing soon. K We will do a BACKUP/IGNORE=INTERLOCK/RECORD from one disk to another during J the night, then do a mimum boot and BACKUP/SINCE=BACKUP/REPLACE to get allL files modified during the morning. Then I'm going to run the procedure belowK to compare the two disks and see how far we can get through the comparision  before our time limit is hit.      --E AOL CD Use #3415; As a handy prop when trying to explain the sport of % Curling to your Mother and your Wife.      $!  $ disk1 = f$parse(p1,,,"DEVICE")  $ disk2 = f$parse(p2,,,"DEVICE") $! $ final_status == 0  $! $ call do_compare 'disk1 'disk2  $ call do_compare 'disk2 'disk1  $!0 $ if final_status .ne. 0 then write sys$output -7 "There was at least one error, please check the output"  $! $ exit $! $do_compare: subroutine  $!- $ write sys$output "Comparing ''p1' to ''p2'"  $!? $! Note that even though this will not pickup empty directories B $! it is OK since the directory itself will not compare correctly.A $! i.e. If DKA100:[TEMP.EMPTY] is empty then we will see an error E $! in the file DKA100:[TEMP]EMPTY.DIR even though DKA100:[TEMP.EMPTY]  $! itself was not tested.X $!= $ directory 'p1[000000...]/head/notrail/total/out=compare.lsta $ open/read infile compare.lst $! $read_loop:e) $ read infile inline/end=end_of_read_loopt- $ if "''inline'" .eqs. "" then goto read_loope# $ dir_name1 = inline - "Directory " : $ dir_name2 = f$parse(p2,dir_name1,"*.*;*",,"SYNTAX_ONLY")J $ write sys$output "Comparing ''dir_name1' to ''p2' ''f$cvtime(,,"time")'" $ on error then continue& $ backup/compare 'dir_name1 'dir_name23 $ if $status .ne. %X10000001 then final_status == 1i $ goto read_loop $! $end_of_read_loop: $! $ close infile% $ delete/noconfirm/nolog compare.lst.t $endsubroutine $!   ------------------------------  % Date: Fri, 22 Feb 2002 17:19:13 +0000 ( From: Nic Clews <sendspamhere@127.0.0.1>/ Subject: Re: Backup/Restore Verification Errors ) Message-ID: <3C767D91.877CE27B@127.0.0.1>e   Rick Dyson wrote:a > J > I've been testing doing image backups to DLT, restoring to a spare disk,D > and then comparing the original disk to the spare disk to test the > integretity of the backup. > Q > I'm bothered by what appears to be hundreds and hundreds of verification errors " > when I perform this exercise. :(   Rick,r  F I've responded AFTER reading Tim's response, that I'm not sure I fully agree with, but here goes.  ? I believe you're seeing what I would call 'expected behaviour'.   E You're doing an online backup, backing up all the files also open fornG write, that could change while backup up that file (/IGNORE=INTERLOCK),cF restoring this historical perspective of the disk, then getting a long= list of all the differences where a WRITE did actually occur.e  F Take the case of a single very large file open for write. BACKUP couldE copy the data, and right after it's copied block 123 of the file, you"H application could change it, and a data block not yet copied to tape. Is? file integrity affected? Depends on your application. If it's a?G relational database, then almost certainly you'll have a 'corrupt' copyp
 of your data.S  H Is there any point performing a verification as part of the backup of anC open-for-write disk? Well all it proves (and what your exercise has C proved) is the integrity of the backup saveset, you can recover the @ data, even if it is a corrupt version, the saveset itself is notF corrupt. It's a good idea to ensure the saveset is readable, but close6 to doubling the backup duration may not be acceptable.  B If you're not taking a queiscesed (sp?)/ database shutdown type ofG backup of your data, and your data is sensitive to inter and extra fileoD relationships, then you need to do this. That would be the saveset I/ would worry about if I saw verification errors.    HTH  -- c( Regards, Nic Clews CSC Computer Sciences nclews at csc dot com    ------------------------------  % Date: Fri, 22 Feb 2002 17:27:38 +0000r( From: Nic Clews <sendspamhere@127.0.0.1>/ Subject: Re: Backup/Restore Verification Errorse) Message-ID: <3C767F8A.45D33705@127.0.0.1>    Nic Clews wrote: > D > If you're not taking a queiscesed (sp?)/ database shutdown type ofI > backup of your data, and your data is sensitive to inter and extra file F > relationships, then you need to do this. That would be the saveset I1 > would worry about if I saw verification errors.a  @ Bad practice I know, but I forgot to say, you can do the compareF directly from tape. INITing a new disk may give you slightly different9 structured system files and a 'false' verification error.h  0 You get all this BACKUP functionality for FREE*.    What do you get with Windoesn't?   Nowt, nothing, zero.  + *Well in the base operating system license.- -- -. Regards, Nic "Bob" Clews CSC Computer Sciences nclews at csc dot comt   ------------------------------  # Date: Fri, 22 Feb 2002 17:30:04 GMTo' From: Rick Dyson <Rick-Dyson@UIowa.EDU>2/ Subject: Re: Backup/Restore Verification Errorsp) Message-ID: <3C76801C.1B9F2678@UIowa.EDU>r   Tim Llewellyn wrote: > > $ Mount /For $1$DGA117:eI > > $ Backup /Log /Image /Initialize /BlockSize = 65535 /Ignore = Label - ! > >          MKC600:Foo-Bar.bck -0, > >          $1$DGA117: /By_Owner = Original >  > Again, why no /VERIFY?  I Because I had found in the past that the DLT with it's error checking wassM reliable and all tests with /verify were always (apparently!) successful.  It ) also more than doubles the backup window.o  G Note, in this testing, I have done "Backup/verify" from the tape to ther0 scratch disk and it did not report any problems.  M > > I have tried this with several different savesets from different tapes (ItR > > have always used the same scratch disk as the restore target), always with theR > > same kinds of negative results.  The amount of verification errors varies, butQ > > is always at least 500 instances per 4 GB disk. (note: these source disks arec- > > typically 99% full of data in 1-5 files.)c > N > Are you SURE nothing else is writing to DSA1 between writing of the tape andM > comparing to the other disk? Is DSA1 mounted in a cluster? If so, check allaL > nodes to ensure no access. Best, dismout DSA1 and mount it readonly beforeL > doing the initial backup and only remount it writeable once the compare is > completed.  K 	Yes. But unless the tape/saveset is actually bad maybe that is the source. J The disks are archive data (hence the reason the volumes are full) and are
 considered static by us and applications.  N If the files are read/accessed for restoring some data, could the time tags inI the header be changing?  The disks are dismounted and mounted for variousDI reasons between the backup and my testing right now.  The shadow sets are)J broken, re-joined etc.  Maybe some of these make little marks on the disk?  L I will investigate the idea of mounting these disks with "/NoWrite".  I haveM never used that and it did not occur to me, duh!.  Especially since these are0M to be filled and then never deleted and we have to keep them forever...  ThatnL is why I want a reliable backup set so when the hardware finally failes and I we have to replace disks and repopulate them with the old data, I will be0+ confident that the files are not corrupted!e  L > > I would greatly appreciate any suggestions or insight you can provide onN > > this situation!  I would realy like to be more confident of my backups. :) > Q > First, always use /VERIFY. You did not verify your disk to tape or tape to disknS > backups. It is possible that errors are being introduced at that stage. Are theren7 > any disk or tape errors logged during the operations?   L None.  I have had long discussions on the merit of /Verify and /CRC with DLTO drives and the guy who wrote Backup for OpenVMS once suggested it was not worthe+ the time and so I have heeded his words. :)o  F Thanks for the comments and ideas.  It has stirred some good thoughts.  N I will try some virgin backups of the disks to tape (with and without /Verify) and P then restores, etc. with the disk write-locked (and without) to investigate this further.   rick   ------------------------------    Date: 22 Feb 2002 02:32:41 -08005 From: mortimer_mouse@mailandnews.com (Mortimer Mouse)-7 Subject: Re: Cisco 2948G-L3 layer three switch with VMS < Message-ID: <b05aa881.0202220232.cab4cf5@posting.google.com>  e Alan Greig <a.greig@virgin.net> wrote in message news:<79p97u0tjtgamql40aaq143dsn7lja9gnj@4ax.com>...'D > What about the suggestion from "joe" to "put all the ports on that6 > bridge group in one port channel  group (like 7500'sH > do) and on the port channel group, put the IP address. I do it here."? >   D I believe he's talking about etherchannel between devices, which youA can certainly do on the 2948G-L3.  I couldn't possibly comment onEF whether your VMS cluster could be configured for etherchannel, though.   ------------------------------    Date: 22 Feb 2002 10:34:19 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) U Subject: Re: Create an Open Mail Relay (GASP!) and Turn it Into a Relay Spam Honeypoto3 Message-ID: <fgW1mLWL5hlB@eisner.encompasserve.org>   _ In article <3C7653C9.A232B76E@mail.tds.net>, "brad.madison" <brad.madison@mail.tds.net> writes:  > Create an open mail relay:   > Voila: an Open Mail Relaye > % > Turn it into a relay spam honeypot:  > $ > $ STOP/QUEUE  UCX$SMTP_NODENAME_01  E > A Linux honeypot in Moscow stopped (STOPPED!) over 1.7 million spam$J > deliveries in about a month.  That may be an atypical number but it sureI > is nice.  As a bonus a bunch of spammer dialup accounts were wiped out.l  B More sophisticated spammers, however, run some test messages to an; address they control to ensure the mail is getting through.n  D I have no general objection to SMTP honeypots, but keep in mind thatB they do not relieve all the extra load on the Internet due to spam (or spam attempts).v   ------------------------------  # Date: Fri, 22 Feb 2002 14:20:57 GMTt0 From: "brad.madison" <brad.madison@mail.tds.net>Y Subject: Create an Open Mail Relay (GASP!) and Turn it Into a Relay Spam Honeypot (Yay!) a, Message-ID: <3C7653C9.A232B76E@mail.tds.net>   Create an open mail relay:  E Grab an old Vax/Alpha out of a closet, configure it with UCX SMTP (noBG anti-relaying features), hook it to the net and start it.  Don't bother H giving it a DNS entry, don't create any MX records pointing to it.  Just
 let it be.   Voila: an Open Mail Relays  # Turn it into a relay spam honeypot:r  " $ STOP/QUEUE  UCX$SMTP_NODENAME_01  E Or maybe it's 00, or both.  You get the idea: stop the delivery queueyB and you have a honeypot.  Obviously you don't want to do this on aE regular system, but thiss is supposed to be that old clunker from theuG closet.  Ask yourself: who will send mail to or throught this otherwises unknown IP number?    G If you are located on a net segment likely to be tested by spammers for_H open relay you should soon see in the queue an open relay test message.  It may look something like:     F Received: from 209.19.209.202 by xx.xx.xx.xx; Wed, 16 Jan 02 01:26 CSTA Message-Id: <iiipppiiipppiiipppiiipppiiipppiiipppiii@ip.ip.ip.ip>G% Date: Tue, 15 Jan 2002 23:18:08 -2000a From: netpro77@pacbell.net Subject: Re: Re: To: <Undisclosed Recipients> MIME-Version: 1.0e0 Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: 7bitu
 X-Priority: 3: X-MSMail-Priority: Normaln2 X-Mailer: Microsoft Outlook Express 5.00.3018.13008 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3018.1300   050048057046049057046050048057046050048050058050048057045049057045050048057045050048050045115100045048049046099118120046097108103120$ 046110101116058049050057055058049058  E (This is from an old free PMDF queue, so the UCX version version willI differ.)  H Save a copy and force delivery and soon you should receive relay spam.     Simple, eh?.  G Oh.  That numeric message is just decimal ASCII, and encodes the source)B of the message and a few other things.  I munged the fields in theF Message-Id because the first is the target IP address in decimal ascii/ and the second is the IP address in plain text.t  C A Linux honeypot in Moscow stopped (STOPPED!) over 1.7 million spamhH deliveries in about a month.  That may be an atypical number but it sureG is nice.  As a bonus a bunch of spammer dialup accounts were wiped out.l  ' <http://www.corpit.ru/cgi-bin/h0n5yp0t>c   --> "Our problems are mostly behind us.  What we have to do now is. fight the solutions."  ---Stult's Report (from" http://www.reznor.com/~aj/quotes/)  2 See: http://frs.axisline.net/ for my honeypot page   ------------------------------  % Date: Fri, 22 Feb 2002 11:52:30 -0500n0 From: "Syltrem" <syltrem@videotron.spammenot.ca> Subject: Re: DCL procedure's2 Message-ID: <uGud8.478$a04.2606@tor-nn1.netcom.ca>   OK.My turn then.F This is a link to my program library page. I only have a few things inB there. Maybe one day I'll ckean up my directory and add some more.< In there you`ll find 2 of the most useful procedures I have:D 1 = SEARCH_REPLACE.COM which I use to replace a string by another in multiple files.iK If you look at the code you will see it was written a very long time ago inoC an emergency situation and I didn't have time to give symbols fancy E meaningful names. But it works just great so you don't care. Can also & replace apostrophes and double quotes.K 2 = RESUBMIT.COM which resubmits a job that`s already on a queue, after you  modified the .COM ' Will resubmit with same parameters etc.h   And there are others, too.1 http://pages.infinit.net/syltrem/prog_library.htm      --   Syltrem I http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)o> To reply to myself directly, remove .spammenot from my address  E "Brian Tillman" <tillman_brian@notnoone.notnohow.com> a crit dans le * message de news: 3c72a0ab$1@news.si.com...0 > >i am looking for some usefull DCL-procedure's& > >maybe you wanna share some programs >?I > Here's my entry.  We populate our VMS Mail Profile programmatically and.J > sometimes get eight bit characters in it.  Using REMOVE from within MAILL > won't let you delete those names (can't type them on the keyboard) so I'veH > produced the following that lets you step through VMSMAIL_PROFILE.DATA fromF > any given starting point until you find the name you want to delete. > --C > Brian Tillman                   Internet: tillman_brian at si.comdC > Smiths Aerospace                          tillman at swdev.si.comt? > 3290 Patterson Ave. SE, MS      Addresses modified to preventr> > Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@": >        This opinion doesn't represent that of my company >  >i > $! CLEANMAIL.COME > $   profile = f$parse( "vmsmail_profile", "sys$system:.data", , , )u. > $   open/read/write/share mailfile 'profile' > $!
 > $ again:I > $   read/end=done/prompt="What is the name to be deleted? " sys$commandr name > $   name := 'name' > $!1 > $   read/key="''name'"/match=ge mailfile record  > $!	 > $ redo:o< > $   address = f$edit( f$extract( 0, 31, record ), "trim" )L > $   read/end=done/prompt="Is ''address' the name you want? " sys$command - >         response > $   if responsel
 > $   then > $!6 > $       read/key="''address'"/delete mailfile record@ > $       read/end=done/prompt="Any more? " sys$command response > $       if response0 > $       then goto againe > $       else goto done > $       endif/ > $!
 > $   else > $! > $       read mailfile record > $       goto redoc > $! > $   endif< > $!	 > $ done:E > $   close mailfile
 > $   exit >E >7   ------------------------------  % Date: Fri, 22 Feb 2002 12:25:36 -0500u5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>x8 Subject: Re: DECWindows and the MIT-SHM server extension0 Message-ID: <2fvd8.85$fL6.1053@news.cpqcorp.net>  J Look at the sources to x11perf.  It does shm putimage operations using the standard sources.7  C Frankly, I think that it is the only thing that has ever tested it.A      B Mark Berryman wrote in message <3C752C43.4BBF57A0@Mvb.Saic.Com>...H >Does anyone have any experience with the shared-memory extension in theI >X-windows server on VMS who can give me some pointers?  So far, my everya? >attempt has aborted with the following error when I attach the  >shared-memory segment:O6 >BadAccess (attempt to access private resource denied)1 >  Major opcode of failed request:  131 (MIT-SHM) 3 >  Minor opcode of failed request:  1 (X_ShmAttach)' >oH >(Much) more detail is available if needed but I wanted to see if anyone >has actually used this first. >  >Mark Berryman >Mark.Berryman@Mvb.Saic.ComD   ------------------------------  + Date: Fri, 22 Feb 2002 08:46:37 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>I Subject: Re: Einstein(; Message-ID: <01KEKHAOZH828Y7KQF@sysdev.deutsche-boerse.com>r  F > > Is there another NG to move this and the other Einstein thread to? >  0/ > sci.physics.relativity looks like a good bet.   H Actually, it's an unmoderated, relatively high-traffic group with a lot G of threads "disproving" relativity etc.  Perhaps sci.physics.research;  G it's a moderated group (I used to be one of the moderators back when I eF was in academia), low-traffic, high-quality.  The question of whether F Einstein knew about the Michelson-Morley result and to what extent it G played a role in his thinking would be perhaps barely on-topic for the iI group.  A definitive answer with references (perhaps to Pais's excellent i- SUBTLE IS THE LORD) would probably come soon.    ------------------------------    Date: 22 Feb 2002 11:30:25 -0600- From: koehler@encompasserve.org (Bob Koehler)  Subject: Re: Einsteind3 Message-ID: <yHGBemJ1oPyY@eisner.encompasserve.org>"  R In article <3C7567BC.C5C06874@spam.net>, Brass Christof <welcome@spam.net> writes: > "Alan E. Feldman" wrote: >  > [SNIP] > E > There must be a better place to discuss this - or is it in any way o > related to VMS?   H    Of course, it's related to VMS.  I was a grad student in physics when!    I first got my hands on a VAX.a  G    What do you want us to discuss the plural of everyone, or the length.    of Carly's hair?u   ------------------------------    Date: 22 Feb 2002 11:33:16 -0600- From: koehler@encompasserve.org (Bob Koehler)t Subject: Re: einstein 3 Message-ID: <ItJFP9g340ZJ@eisner.encompasserve.org>i  R In article <3C756A1E.3AD44797@spam.net>, Brass Christof <welcome@spam.net> writes: >>  I >> In other words, the history of science is in some sense independent of J >> science; an extraterrestrial civilisation would agree with us about theK >> facts of science, but would have an entirely different history.  Art and 6 >> art history, by contrast, are more closely related. >  > Well spoken! > J > But in history of physics the observations played a much more important > > role than theoretical considerations ahead of experimenting. >   G    Any extraterrestrial civilisation that gets Einstine's theory before 9    Newton's and Maxwell's you might want to be afraid of.c   ------------------------------  % Date: Fri, 22 Feb 2002 10:47:46 -0800,# From: "Tom Linden" <tom@kednos.com>3 Subject: RE: Einsteinl9 Message-ID: <CIEJLCMNHNNDLLOOGNJIAEELEDAA.tom@kednos.com>    > -----Original Message-----7 > From: Alan E. Feldman [mailto:spamsink2001@yahoo.com]t+ > Sent: Thursday, February 21, 2002 6:39 PMo > To: Info-VAX@Mvb.Saic.Com  > Subject: Re: Einsteino >n > 4 > Brass Christof <welcome@spam.net> wrote in message& > news:<3C7567BC.C5C06874@spam.net>... > > "Alan E. Feldman" wrote: > >t
 > > [SNIP] > >kF > > There must be a better place to discuss this - or is it in any way > > related to VMS?s > > F > > > However, Maxwell's equations were derived from experiment, so ifI > > > Einstein used that he was still using experimental results. You areiJ > > > right about the problem of the force depending on the speed of lightI > > > (speed of the electrodynamic field as you put it, which is the samedJ > > > thing). This was a great mystery until Einstein came up with Special > > > Relativity.T > >-G > > I think your conclusion isn't correct that a theoretical hypothesis-G > > which partly uses facts that are experimentally supported is itselfrG > > to be regarded as based on experimental results. What really countsi >iC > If Nobel prizes are based on "invention and discovery", certainlyr) > Special Relativity counts as discovery!e  A Boy, Plato certainly left his legacy, theories are inventions nota discoveries.   >iI > > is the part that isn't based on experimental evidence. AFAIK Einstein K > > is in that respect unique that he proposed several theories long beforerI > > any expermimental data could be gathered. Especially the influence ofa > E > What several theories? His explanation/theory for the photoelectriciF > effect explained it whereas it could not be explained satisfactorilyH > in the framework of classical physics. He explained Brownian motion inD > terms of statistical mechanics. He developed special relativity toH > reconcile the contradictions between Newtonian mechanics and classicalD > electromagnetic theory. He developed General Relativity to explainA > gravitation in a way that was consistent with the principles ofiD > Special Relativity. These theories had predicted various amount of > other things.  >oH > Many theories predict new things. Dirac predicted the existence of theE > positiron before (maybe not long before, I'd have to look it up) it G > was known. Many other particles were predicted by physicists based on < > quantum mechanics and known properties of known particles. >sK > > gravity upon photons couldn't be measured when he predicted the effect.I > C > This is exactly what I was talking about when I said the theoriesIG > predict new results that can be tested by experiment. Einstein is not A > unique in this respect. He is unique in that he did much of hiseE > greatest work outside of academia and he was truly a great thinker.c >cB > And the only thing that delayed cheking the effect of gravity onB > photons was WWI. When the war was over, Eddington (IIRC) went onF > expedition to the next solar eclipse and measured stars near the sun) > to be deflected as given by the theory.- >iE > Some theories do not predict anything that can be currently tested,i > e.g. string theory.o >r > >oI > > > Also, Richard P. Feynman and two other physicists (I can't rememberoK > > > how to spell their names) were awarded a Nobel prize for developing ahA > > > method called "renormalization" which was a theoretical (or K > > > mathematical, if you prefer) trick to almost literally sweep infiniteiI > > > terms from a series under the carpet. Their work was strictly abouteK > > > calculating quantites in quantum mechanics, so that is as theoreticaleK > > > or mathematical as you can get in physics. But it is still related to I > > > the real world, or at least what experiments tell us about the reale< > > > world. And more importantly, they still got the prize. > >nG > > There is a nice anecdote about the relation between theoretical and G > > experimental physics that is too long to tell it here. The moral iseF > > that normally the theory is built based upon the observations. ButA > > there are indeed exceptions which have to be regarded as veryh? > > precious. There is also the very unpleasant phenomenon that @ > > observations are dropped that don't fit into the theoreticalB > > picture. Only when the theory advances those experimental data0 > > are sometimes revived and newly interpreted. >eC > Sometimes the initial experiments are wrong. Distances to distantID > stars, nebula, and especially galaxies were initially way off. TheG > discovery of a second type of Cepheid variables (which was previouslynH > taken to be a "standard candle") cleared up much of this. So it's okay" > to ignore the incorrect results! >  > >dF > > > Of course the criteria for a Nobel prize may have changed in the5 > > > intervening years between Einstein and Feynman.o >o > Disclaimer: JMHO > Alan E. Feldman # > afeldman ;d;d;dnonon gfigroup.come >e   ------------------------------  # Date: Fri, 22 Feb 2002 08:27:54 GMTs* From: "Bill Todd" <billtodd@metrocast.net>5 Subject: Re: EV8 and McKinley analysis by Paul DeMone)C Message-ID: <eind8.152252$d34.11330328@bin8.nnrp.aus1.giganews.com>t  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:ABgYa0TnSmMp@eisner.encompasserve.org...    ...o  @ > that the architecture isn't totally brain dead.  After all, atA > 800 SpecInt2000 that would put it more than 40% better than the2H > best UltraSparc integer number out there today... and we know how wellB > Sun is doing with their weak UltraSparc.  Imagine how well Intel? > will do with a much stronger McKinley in an Industry Standard  > box!  K Imagine indeed, given that Sun manages to compete quite effectively alreadyrF against servers using processors with McKinley-level processor SPECint performance.  K And Sun won't be competing with 'an Industry Standard box' any more than itoL is competing today with industry-standard IA32 servers.  Sun makes its moneyA on higher-end servers, which contain a great deal of system-level J engineering (partitionability being one obvious example) that doesn't seemJ likely to be appearing any time soon in something like a Dell box (even ifJ they eventually support enough processors to make things like partitioningI worthwhile).  So Sun will be competing with proprietary higher-end server I designs from the likes of Compaq and/or HP (as will the IBM 690s that youiK mention below - and in that context you might remember that Compaq high-end 3 server RAM also costs a lot more than Dell RAM...).7   - bill   > H > Price performance?  Ever priced an IBM 690?  Bet that McKinley serversC > are considerably cheaper and yeah... the sales angle for McKinleyd> > will be price performance angles, especially if you get themA > with 16, 32 or 64 GBytes of RAM.  IBM RAM costs a lot more thanr > Dell RAM.  ;-) >e > Robb >t >    ------------------------------  # Date: Fri, 22 Feb 2002 07:50:09 GMTr* From: "Bill Todd" <billtodd@metrocast.net>5 Subject: Re: EV8 and McKinley analysis by Paul DeMone C Message-ID: <RKmd8.149599$Aw2.10648035@bin7.nnrp.aus1.giganews.com>i  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:ABgYa0TnSmMp@eisner.encompasserve.org...s   ...    > Earlier you wrote:   ...>  H > Or, to put it another way, Itanic is about 2 years behind the industryL > leaders in integer performance today, and appears to be 3 - 4 years behindD > in being able to implement the next couple of rounds of technology advances >n > [snip] >tF > A document that suggests that this situation will completely reverse	 itself in J > 3 or 4 years without *major* analysis to back it up is - well, laughable is the@ > most charitable characterization I can come up with right now. >o@ > It may seem laughable ... but if they roll at 750 integer thatC > will be a great achievement and will go from being 2 years behindm9 > to nearly caught up... and completely reversing itself.   J Utter crap, Rob.  If they roll in at 750 *this* year that will be a numberJ which, if Merced had hit it *last* year would have been 'nearly caught up'B *then*.  So indeed McKinley will have improved things (as everyoneK expected), but only to the point of being about 1 year behind rather than 2i in SPECint.i  D And in other performance areas such as bandwidth (and, at least whenL compared with EV7 and Hammer, memory latency) it will remain 3 years behind:H POWER4 has on-chip switched MP facilities today, EV7 will ship with themI soon, Hammer will ship with them before year's end - and Itanic will ship E with them sometime after Madison/Deerfield (2005 seems a good guess).3  L This is nothing like a complete reversal of Itanic's inferiority:  it is, asJ I said, just a tad better than the worst expectations suggested.  You even$ highlight these inferiorities below:   ...n  < > Marvel based systems ought to have about 7-10 times higherC > STREAM numbers than McKinley based systems... that will translatesF > into much higher numbers for bandwidth sensitve applications... thatB > and 30%+ greater local memory latency, 64 processor systems withD > better average memory latency than competing 32 processor systems.B > EV7 will make for much stronger high-end systems.  A nice niche. > L > > McKinley's main problem will be that even at 800 SPECint (if it actuallyG > > attains that) it will look sickly compared with the very definitelyeF > > still-robust POWER4 competition, which has not only about the same figureK > > for just *one* of the 2 cores on its die (which makes for roughly twice  theWK > > performance both per package and per Watt) but dramatically more MP andpI > > memory bandwidth available as well.  In the higher-end server marketsyK > > McKinley is aimed at, both are significant disadvantages - while in the K > > lower-end server markets it might otherwise find refuge in its own IA32oI > > brethren will make it look expensive and slow (as will Hammer when itt
 > > arrives).  > >t >y@ > Yeah but.... that 1 cycle access L1 and 64 GB/sec L2 bandwidth7 > should make for some very nice McKinley tpmC numbers.t  K It would surprise me somewhat if tpmC executes anything but its code (whichMK most architectures manage to pre-fetch reasonably efficiently regardless of K their cache speed) significantly out of cache:  do you have a reference fory this assumption?   > J > > Just not being quite as bad as some had feared does not seem likely to beJ > > enough to assure McKinley's success:  it'll still have to earn it, and fromG > > any reasonable price/performance viewpoint shows little prospect ofe doing  > > so.b > A > Just have to wait and see some numbers.  The fact that McKinleygA > has upside on a known weak spot of the Itanic (Integer), proves/1 > that the architecture isn't totally brain dead.e  G No, it only proves that a good implementation can get even a brain-deaddA architecture off the ground (see Paul's comments to that effect).iK Unfortunately, once you've *got* a good implementation further gains becomeoL a great deal harder to come by in the next generation:  I've always concededH that POWER/EV7/Hammer-style on-chip glue could be grafted on by 2005 andK Alpha-style internals substituted by 2006, but that's a long time to remain'H uncompetitive - and while the Alpha team is occupied transforming ItanicJ internals into what they should have been already, the competition will be= busy forging ahead into the next generations of improvements.o   ....  H > Price performance?  Ever priced an IBM 690?  Bet that McKinley serversC > are considerably cheaper and yeah... the sales angle for McKinleyt> > will be price performance angles, especially if you get themA > with 16, 32 or 64 GBytes of RAM.  IBM RAM costs a lot more than  > Dell RAM.  ;-)  D Most people who purchase IBM 690s likely do so because they need theJ performance they provide, since IBM has far less expensive POWER boxes forA those who need less.  And it's far from clear that *any* McKinleyeL configuration will ever approach that level of performance (and since POWER4K can ride the same performance curve that Madison does, the same is true for H Madison as well).  So on the high end, for at least the next 3 - 4 yearsJ Itanic seems unlikely to be able to compete on performance, period - which$ makes price/performance a non-issue.  K On the lower end, of course, Itanic will get killed in price/performance bydG its IA32 siblings.  And in the mid-range by Hammer.  The result is that I Itanic's competitive position at the back of the pack hasn't changed:  it K just isn't quite as far behind the front-runners as it might have been, but I until there's some credible evidence that it's about to catch up adoption I rates will not be anything like those required to drive its volume up and2L its price down into 'commodity' territory - which is what its domination has always been predicated upon.   - bill   ------------------------------   Date: 22 Feb 2002 14:00:06 GMT) From: leslie@clio.rice.edu (Jerry Leslie)p5 Subject: Re: EV8 and McKinley analysis by Paul DeMonen' Message-ID: <a55it6$eov$1@joe.rice.edu>a  ) Bill Todd (billtodd@metrocast.net) wrote:r : J : And Sun won't be competing with 'an Industry Standard box' any more than> : it is competing today with industry-standard IA32 servers.    % Sun has abandoned its IA32 version...   6    http://www.nwfusion.com/news/2002/0108sunintel.html3    Sun to drop Intel support in Solaris 9, 01/08/02d      By Ashlee Vance    IDG News Service, 01/08/02   E   "Sun Microsystems Tuesday said it will not support chips from InteleG    with the release of its Solaris 9 operating system this year, as thedA    hardware maker looks to cut costs in a tough economic climate.   E    Most corporate Solaris users run the operating system on Sun's own-G    64-bit UltraSparc processors, but a number of users have also workedbI    with the software on servers with Intel's 32-bit chips. Companies suchDH    as Dell Computer typically load Solaris onto Intel-based servers uponH    customer request, even though Dell does not directly offer Solaris as
    an option.l  D    This tradition, however, will come to a close with the release ofE    Solaris 9 in the first half of this year. Sun decided the costs ofsD    support, such as bug tracking and software patches, for Solaris 9=    running on Intel was not worth the cost to the company..."-  $ and Sun will be selling Linux boxes:  I    http://www.sun.com/2002-0206/linux/;$sessionid$C30G3BRUN5ZKPAMTA1LU4GQr    Sun technologies - Products  F   "SUN BROADENS SUPPORT FOR LINUX - Aggressive new program expands the(    role of Linux on entry-level servers.  G    07.Feb.02--Sun Microsystems has embraced the Linux operating system, F    rolling out a multipart program that will significantly broaden theH    offerings of Linux on low-end Sun servers and commit new resources toB    the ongoing development of the Open Source operating system..."    4 --Jerry Leslie     (my opinions are strictly my own)   ------------------------------  # Date: Fri, 22 Feb 2002 17:48:57 GMTs* From: "Bill Todd" <billtodd@metrocast.net>5 Subject: Re: EV8 and McKinley analysis by Paul DeMone C Message-ID: <dwvd8.156144$d34.11684915@bin8.nnrp.aus1.giganews.com>n  6 "Jerry Leslie" <leslie@clio.rice.edu> wrote in message! news:a55it6$eov$1@joe.rice.edu...-+ > Bill Todd (billtodd@metrocast.net) wrote:6 > :0L > : And Sun won't be competing with 'an Industry Standard box' any more than> > : it is competing today with industry-standard IA32 servers. >5' > Sun has abandoned its IA32 version...w  F We know, John, but that has absolutely nothing to do with what we were talking about.   - bill   ------------------------------  % Date: Fri, 22 Feb 2002 12:47:39 -0500o5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> 5 Subject: Re: EV8 and McKinley analysis by Paul DeMones0 Message-ID: <Jzvd8.86$fL6.1066@news.cpqcorp.net>  J I don't get it Bill.  While at least conceding relative performance parityL between Sparc and McKinley (probably a combination of using the artificiallyA inflated Sun spec numbers, and the low side estimates of McKinley K performance) - you then say it doesn't matter - because the devil is in the-H system-level engineering done by Sun - and somehow an "Industry StandardK Box" (which I don't know what that means, but will take a wild ass guess itJJ means an Intel CPU in it that can run Windows) -- somehow won't measure up* to the system engineering that Sun can do.  J I've looked at some of the near term, and long term IPF "enterprise" classE boxes being designed and built here at Compaq, and they are in no way H inferior to what Sun is building (or they are likely to build).  I'm notC sure how the comment got turned into a Dell desktop box comparison.i  H I at *least* expected you to shrug it off as "so what, Sparc sucks too",J ignoring the fact that those boxes of junk Sun sells make them the #1 UNIX vendor.   J Or to say "yeah, well it will never be as fast as an Alpha that might have been built".  L If a IPF box ever shows up that is as fast as the then-current PowerN chip -9 what will be the next reason why IPF sucks and is doomed?m       Bill Todd wrote in message ... >l9 >"Rob Young" <young_r@encompasserve.org> wrote in message'. >news:ABgYa0TnSmMp@eisner.encompasserve.org... >  >... >eA >> that the architecture isn't totally brain dead.  After all, ataB >> 800 SpecInt2000 that would put it more than 40% better than theI >> best UltraSparc integer number out there today... and we know how well C >> Sun is doing with their weak UltraSparc.  Imagine how well Intela@ >> will do with a much stronger McKinley in an Industry Standard >> box!i >tL >Imagine indeed, given that Sun manages to compete quite effectively alreadyG >against servers using processors with McKinley-level processor SPECinta
 >performance.h >dL >And Sun won't be competing with 'an Industry Standard box' any more than itG >is competing today with industry-standard IA32 servers.  Sun makes itsb moneyiB >on higher-end servers, which contain a great deal of system-levelK >engineering (partitionability being one obvious example) that doesn't seemiK >likely to be appearing any time soon in something like a Dell box (even ifaK >they eventually support enough processors to make things like partitioningpJ >worthwhile).  So Sun will be competing with proprietary higher-end serverJ >designs from the likes of Compaq and/or HP (as will the IBM 690s that youL >mention below - and in that context you might remember that Compaq high-end4 >server RAM also costs a lot more than Dell RAM...). >h >- billo >t >>I >> Price performance?  Ever priced an IBM 690?  Bet that McKinley serversoD >> are considerably cheaper and yeah... the sales angle for McKinley? >> will be price performance angles, especially if you get themhB >> with 16, 32 or 64 GBytes of RAM.  IBM RAM costs a lot more than >> Dell RAM.  ;-)  >> >> Rob >> >> >t >n   ------------------------------    Date: 22 Feb 2002 12:47:02 -0600+ From: young_r@encompasserve.org (Rob Young)r5 Subject: Re: EV8 and McKinley analysis by Paul DeMonet3 Message-ID: <0k$edNglCJpf@eisner.encompasserve.org>k  p In article <fbwd8.153867$Aw2.11063336@bin7.nnrp.aus1.giganews.com>, "Bill Todd" <billtodd@metrocast.net> writes:  H > 2.  To avoid complete redundancy it seems reasonable to assume that anL > 'industry standard box' is something more than simply any box that happens- > to contain an industry standard processor.      B 	One wrinkle.. Industry Standard Server.  Since AMD really doesn'tE 	do much with IA32 servers (their share of the server space is almostrC 	non-existent ... currently).  So Industry Standard Server has comeoC 	to mean a server with an Intel processor in it and most observers,rC 	writers and analysts would tend to agree with that.  By leveragingiA 	that built in marketing, the next step of course is to refer to m= 	an IPF box as an Industry Standard Server and of course youryA 	corporation wants to participate in using Industry Standards andw@ 	leverage all the cost savings and other benefits of being on an 	Industry Standard platform.   				Robo   ------------------------------  + Date: Fri, 22 Feb 2002 11:27:05 +0000 (UTC)e9 From: Roar =?iso-8859-1?Q?Thron=E6s?= <roart@nvg.ntnu.no>lS Subject: Re: FreeVMS: what are CPQ's feelings likely to be ?, was: Re: MPE/iX usersa- Message-ID: <a559u9$ntj$1@tyfon.itea.ntnu.no>c  H Simon Clubley <simon_clubley@remove_me.altavista.co.uk-earth.ufp> wrote:, : bob@instantwhip.com (Bob Ceculski) writes: :> tF :> if Compaq/HP doesn't want "paying" vms customers, and try to do theH :> same to vms, then I hope FreeVMS comes out eventually and we will not* :> buy any Compaq/HP product again "ever"!  N : The FreeVMS concept does raise some questions about if CPQ will try to block : it.a  J : Imagine for a moment that the FreeVMS project does actually does produce> : a viable subset of VMS. How will Compaq/HP feel about this ?  M It will be 5.x-ish while we are today at 7.3 (a bit of a feature difference).u  F I doubt any much money would be lost, since it can not replace current$ VMS systems due to lack of software.  J : Will they decide that they will try and block it for some reason or will : they welcome it ?-  E Are you implying that they actually have a long-time VMS strategy? :)l, (Maybe some of them have heard about VMS...)  & Mixed feelings and practises, perhaps?K Just like the VMS developers who do not want opensourcing, but on the otherbJ hand the sourcelistings are sold for almost free to subscribing customers.I (Almost free in the comparative sense; compared to software licensing andr hardware prices)  H : Will CPQ/HP demand that the word "VMS" not be used anywhere within theK : resulting product as they may feel that it could impact on the reputationu
 : of CPQ/HP ?   / VMS and OpenVMS are trademarks, FreeVMS is not.i  < The general use of the word VMS is another thing completely.A It is used here on comp.os.vms, and also in the IBM clone of DLM.e6 (Grep (grep -i vms) for VMS in the source available at4 http://www-124.ibm.com/developerworks/projects/dlm/)   -- '
 -Roar Thronsw   ------------------------------    Date: 22 Feb 2002 07:44:35 -0600 From: briggs@encompasserve.orgF Subject: Re: Help on creating a detached process using SYS$CREPRC(...)3 Message-ID: <$7QTmL9awb4f@eisner.encompasserve.org>e  [ In article <2c0966c2.0202211510.55f509d8@posting.google.com>, hchen3@uwo.ca (Sammy) writes:i > Hi:l >   H >   Thanks for the help. I re-wrote the program based on your suggestion8 > and I am still having a bit trouble. (detail follows).  4 I'll be sprinkling my comments in your code below...   >  > int main() > {  > 	int status; > 	unsigned int pid;9 > 	unsigned int base = 4; /* priority for subprocesses */n > 1 > 	$DESCRIPTOR(image, "SYS$SYSTEM:LOGINOUT.EXE");7# > 	$DESCRIPTOR(output, "test.log");? > 	$DESCRIPTOR(proc, "Test");t > + > 	// attribute required to build a mailboxa" > 	$DESCRIPTOR(boxName,"IPC_BOX"); > $ > 	unsigned short int writeChan = 1;: > 	unsigned int maxmsg = 4096;  // each message is 4k longB > 	unsigned int bufquo = 40960; // up to 10 messages can be stored >  > 	struct _iosb iosb;- > 8 > 	// create a mailbox and assign a channel to access it; > 	status = sys$crembx(0, &writeChan, maxmsg, bufquo, 0, 0,e+ >                             &boxName, 0);8 >  > 	if (status != SS$_NORMAL) > 	{5 > 	   printf ( "SYS$CREMBX has returned abnormally - r+ >                   status = %d\n",status);m > 	   exit(status);. > 	} >    Things look good so far.  % > 	// write java -help to the mailboxo > 	char msg[] = "java -help"; A > 	status = sys$qio(0, writeChan, IO$_WRITEVBLK|IO$M_NOW, &iosb, r? >                          0, 0, msg, strlen(msg), 0, 0, 0, 0);o > 	if (status != SS$_NORMAL) > 	{: > 	   printf ( "SYS$QIO has returned abnormally - status = > %d\n",status); > 	   exit(status);y > 	}  
 Two problems.a  B 1.  You used SYS$QIO.  You should have used SYS$QIOW.  There is noB reason not to wait for I/O completion when I/O completion is going to be immediate.  > 2.  You aren't checking the I/O completion status in the IOSB.  C Both of these are cosmetic.  I don't expect that the QIO is failingl; and it doesn't matter whether it completes synchronously or7 asynchronously.a  ! > 	// write logout to the mailboxh > 	char log[] = "logout"; A > 	status = sys$qio(0, writeChan, IO$_WRITEVBLK|IO$M_NOW, &iosb,  > >                          0,0, log, strlen(log), 0, 0, 0, 0); > 	if (status != SS$_NORMAL) > 	{: > 	   printf ( "SYS$QIO has returned abnormally - status = > %d\n",status); > 	   exit(status);r > 	}   Same objections apply here.y >  > 	// attribute to getdvi: > 	char buffer[1024];i# > 	$DESCRIPTOR(deviceName, buffer);h > 	struct item_list_3m > 	{ > 		unsigned short buflen; > 		unsigned short itmcod; > 		void * bufadr; > 		unsigned short *reslen;0 > 	} itemList[] =s > 	{> > 		{1023, DVI$_FULLDEVNAM, buffer, &deviceName.dsc$w_length}, > 		{0, 0, 0, 0} > 	};u > A > 	// use SYS$GETDVI to get the actual mailbox device informationiA > 	status = sys$getdvi(0, 0, &boxName, itemList, &iosb, 0, 0, 0);g >  > 	if (status != SS$_NORMAL) > 	{3 > 	   printf ( "SYS$GETDVI has returned abnormally  , >                    status = %d\n",status); > 	   exit(status);c > 	} >   F Same problems as with your $QIO calls.  You are using the asynchronousB form of the system service.  And you're not checking the IOSB.  So2 you have no guarantee that it is actually working.  * > 	buffer[deviceName.dsc$w_length] = '\0';A > 	cout << "The real device name of mailbox: " << buffer << endl;h  E On the other hand, you do have debugging code that does indicate thatuD the $GETDVI completes promptly and successfully in spite of the lack of that guarantee.  A But your debugging output indicates that the device name does notrD end with a colon.  And my debugging output indicates that the deviceG name does end with a colon.  Very interesting.  Are you sure you didn't  edit that out? > / > 	// create a detach process to run java -helpi; > 	status = sys$creprc(&pid, &image, &deviceName, &output, rH >                             0, 0, 0, &proc, base, 0, 0, PRC$M_DETACH); >  > 	if (status != SS$_NORMAL) > 	{3 > 	   printf ( "SYS$CREPRC has returned abnormally  , >                    status = %d\n",status); > 	   exit(status);p > 	} >s The $CREPRC call looks good.   >         return 0;a  C And you have another possible problem here.  You created a mailbox.a@ And you created a process to read from that mailbox.  Now you'reA exitting.  But you have no guarantee that the created process has D opened the mailbox yet.  If you exit (and implicitly close your openE channel to the mailbox) before the detached process opens its channeleE to the mailbox then the mailbox goes *poof*.  Temporary mailboxes ares> automatically deleted when their reference count goes to zero.  B This is probably why the other person who sanity-checked your code found that it worked for him.I  A > the log file was created and I got the following error message:  > , > Error opening primary input file SYS$INPUTA > Error in device name or inappropriate device type for operationo  D From this, we can conclude that the device name that you passed intoD $CREPRC did have a trailing colon, despite the debugging output thatC you present below.  I'm betting that the above diagnosis is correctnB and the creating program exitted before the created process opened its input file.,  + Try calling lib$put_output ( &device_name )v  G > I suspect that I did not use the function SYS$GETDVI properly becauseoE > it seems like SYS$CREPRC has trouble reading from the mailbox giveniC > the device name. Can any one give me a hand? Thanks for the help.c > F > By the way, just in case anyone wants to know. here is the output of > the program: > 
 > $run sam1 > The real device name of mailbox: _ALPHA2$MBA144r   	John Briggs   ------------------------------    Date: 22 Feb 2002 08:18:59 -0600 From: briggs@encompasserve.orgF Subject: Re: Help on creating a detached process using SYS$CREPRC(...)3 Message-ID: <Db2BaNFF6YpL@eisner.encompasserve.org>   P In article <00A09EA1.AA7E1862@SendSpamHere.ORG>, system@SendSpamHere.ORG writes:] > In article <2c0966c2.0202211510.55f509d8@posting.google.com>, hchen3@uwo.ca (Sammy) writes:a >>Hi:t >>  H >>  Thanks for the help. I re-wrote the program based on your suggestion8 >>and I am still having a bit trouble. (detail follows). [...]n, >>Error opening primary input file SYS$INPUTA >>Error in device name or inappropriate device type for operationn >> > F > Works for me (well, I did change the "java -help" to something a bit& > more accessible like "SHOW SYSTEM").  < Failed for me with his reported symptoms until I tacked on a     system ( "wait 0:0:10" );g  G at the end to keep the creator from exitting before the created process 5 got going.  Presumably the debugger saved your bacon.t  F > You might consider using the VMS debugger.  A wonderful tool.  Also,C > the problem may not be the program but the fact that "java -help"r= > does not want to run with its input being a mailbox device.d  " Agree 200% about the VMS debugger.  H > You might also want to include the #includes and the struct definitionG > of _iosb so that other readers here might be able to compile and testeF > your program without having to make all of the necessary amendments.   Heh.  I had the same problem.    	John Briggs   ------------------------------  # Date: Fri, 22 Feb 2002 16:57:32 GMTi From: system@SendSpamHere.ORG F Subject: Re: Help on creating a detached process using SYS$CREPRC(...)0 Message-ID: <00A09F20.92CF6079@SendSpamHere.ORG>  T In article <Db2BaNFF6YpL@eisner.encompasserve.org>, briggs@encompasserve.org writes:Q >In article <00A09EA1.AA7E1862@SendSpamHere.ORG>, system@SendSpamHere.ORG writes:h^ >> In article <2c0966c2.0202211510.55f509d8@posting.google.com>, hchen3@uwo.ca (Sammy) writes: >>>Hi: >>>   I >>>  Thanks for the help. I re-wrote the program based on your suggestiont9 >>>and I am still having a bit trouble. (detail follows).m >[...]- >>>Error opening primary input file SYS$INPUT B >>>Error in device name or inappropriate device type for operation >>>y >> lG >> Works for me (well, I did change the "java -help" to something a bith' >> more accessible like "SHOW SYSTEM").  > = >Failed for me with his reported symptoms until I tacked on ao >w >  system ( "wait 0:0:10" );  E Well, as you've pointed out in another follow-up, his code is riddled D with problems -- $QIO without $SYNCH, no check of the IOSB, etc.  HeD needs to correct some of the fundamentals and also look at insuring C that his program will live long enough for process creation to openo a channel to the mbx.U   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMn            pJ   "And of course, I'm a genius, so people are naturally drawn to my fiery I   intellect.  Their admiration overwhelms their envy!" -- Calvin & Hobbesa   ------------------------------  # Date: Fri, 22 Feb 2002 07:12:26 GMTl. From: "Duane Sand" <Duane.Sand@mindspring.com> Subject: Re: Itanium troublesh- Message-ID: <ubmd8.12876$2Q1.53962@rwcrnsc54>    "Terry C. Shannon"  wrotet, > "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote3 > > JF Mezei  <jfmezei.spamnot@videotron.ca> wrote:l > > >Nick Maclaren wrote:.C > > >> But, to forestall Alexis Cousein, the CURRENT committment to-L > > >> MIPS is pretty solid.  There was a (say) 18 month glitch, but that isG > > >> long past and the MIPS line is slowly but steadily making up the. > > >> lost ground.t > > >rL > > >Does this mean that Tandem could remain on MIPS if it were allowed to ? > >  > > Yes. > J > Yes, of course. There's a MIPs refresh coming out in May, and another inK > about 18-24 months. Separately, the revived IPF port is just about a donee > deal.g  J SGI's stock price is now $3/share, "recovered" from a non-tradeable low of $0.50.I No one outside of SGI sees much hope for SGI ever gaining the volumes andgE profits it would need to ongoingly fund high-end chip development forM
 Unix-scaleL systems for 5 years ahead, let alone 10 years.  (Mips Technology's chips forH embedded stuff are very successful, but none of those designs or profits1 help the high end chips used by SGI and NonStop.)l  N My recollection is that the Mips chip line has fallen further and further from its K initial industry leadership position rather continuously, throughout the 12g yearsyH since DEC switched from Mips Unix boxes to Alpha and others then learnedO to do fast clockrate designs too.  The recent "18 month glitch" when SGI had nou- design teams at all was not the only problem.s  F Getting some long-overdue shrink speedups now on old Mips chip designsA is a godsend gap-filler right now, until NonStop finally ships oni something/anythingB more permanent.  But the Mips line is not that permanent solution.  3 -- Duane Sand, not speaking for Compaq or whatever.t   ------------------------------  % Date: Fri, 22 Feb 2002 02:47:03 -0500s- From: JF Mezei <jfmezei.spamnot@videotron.ca>o Subject: Re: Itanium troubleso, Message-ID: <3C75F775.77C9649D@videotron.ca>   Duane Sand wrote:,H > Getting some long-overdue shrink speedups now on old Mips chip designsC > is a godsend gap-filler right now, until NonStop finally ships onr > something/anythingD > more permanent.  But the Mips line is not that permanent solution.  G Would tandem customers prefer to stick with MIPS a while longer until agH permanent owner of NSK/Tandem, can be found and a permanent architecture settled once and for all ?  E How do Tandem customers react to the constant change of ownership and  architecture flip flops ?>   ------------------------------   Date: 22 Feb 2002 12:37:59 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren) Subject: Re: Itanium troublesa0 Message-ID: <a55e37$q2d$1@pegasus.csx.cam.ac.uk>  ' In article <3C75A5BB.A76E9DD4@igs.net>,d% Paul DeMone <pdemone@igs.net> writes:q |> Nick Maclaren wrote: O |> > In article <Pine.SGI.4.44.0202211007410.210522-100000@wendigo.bio.dfo.ca>,m9 |> > "George N. White III" <WhiteG@dfo-mpo.gc.ca> writes:t |> > |>xR |> > |> Certainly SGI and MIPS have not made the advances that one would expect ifH |> > |> MIPS development efforts had not waivered. If SGI had a strongerQ |> > |> commitment to MIPS would we now have machines similar to POWER4 from SGI?s |> > lF |> > Unlikely.  SGI are taking a different path.  However, my guess isG |> > that we would now have 1 GHz MIPS chips in the Origin 3000 series,i' |> > and they might well be quad issue.p |>  E |> The MIPS R1x000 series ARE quad issue now (actually four fetch andn |> five issue).t  ? Thanks for the correction - I had misremembered, and it was thet= number of floating-point pipelines that I was thinking of.  Ir= believe that one of the reasons that the MIPS chips deliver ae@ much higher proportion of their peak than most is that they have9 a higher fetch issue rate than floating-point issue rate.      Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679-   ------------------------------   Date: 22 Feb 2002 12:48:26 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren) Subject: Re: Itanium troubles 0 Message-ID: <a55emq$qlc$1@pegasus.csx.cam.ac.uk>  ^ In article <ubmd8.12876$2Q1.53962@rwcrnsc54>, "Duane Sand" <Duane.Sand@mindspring.com> writes: |> "Terry C. Shannon"  wrote/ |> > "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrotet6 |> > > JF Mezei  <jfmezei.spamnot@videotron.ca> wrote: |> > > >Nick Maclaren wrote:F |> > > >> But, to forestall Alexis Cousein, the CURRENT committment toO |> > > >> MIPS is pretty solid.  There was a (say) 18 month glitch, but that iseJ |> > > >> long past and the MIPS line is slowly but steadily making up the |> > > >> lost ground. |> > > >O |> > > >Does this mean that Tandem could remain on MIPS if it were allowed to ?i |> > > |> > > Yes.l |> >M |> > Yes, of course. There's a MIPs refresh coming out in May, and another infN |> > about 18-24 months. Separately, the revived IPF port is just about a done
 |> > deal. |> aM |> SGI's stock price is now $3/share, "recovered" from a non-tradeable low ofo	 |> $0.50.aL |> No one outside of SGI sees much hope for SGI ever gaining the volumes andH |> profits it would need to ongoingly fund high-end chip development for
 |> Unix-scalerO |> systems for 5 years ahead, let alone 10 years.  (Mips Technology's chips forCK |> embedded stuff are very successful, but none of those designs or profits$4 |> help the high end chips used by SGI and NonStop.)  D Are you sure?  Chip development is not as expensive as is often made@ out, and there are positive advantages in doing it on the cheap,D such as you are not tempted to throw armies of peasants at a problemA to maintain schedules.  Obviously, SGI will not start running itsa8 own fabs.  I agree that SGI is not out of the woods yet.  Q |> My recollection is that the Mips chip line has fallen further and further from  |> itsN |> initial industry leadership position rather continuously, throughout the 12 |> yearsK |> since DEC switched from Mips Unix boxes to Alpha and others then learnedcR |> to do fast clockrate designs too.  The recent "18 month glitch" when SGI had no0 |> design teams at all was not the only problem.  D As a users of them, I disagree.  While they HAVE lost ground in peakC MFlops, that is largely a figure for the amusement of marketdroids.SA In terms of delivered performance, they are pretty good and (as Ie% said) are even making up lost ground.t  I |> Getting some long-overdue shrink speedups now on old Mips chip designsdD |> is a godsend gap-filler right now, until NonStop finally ships on |> something/anythingoE |> more permanent.  But the Mips line is not that permanent solution.   A That may be true, but current evidence is that the IA-64 line mayn well not be, either :-)      Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679m   ------------------------------    Date: 22 Feb 2002 09:19:13 -0600- From: koehler@encompasserve.org (Bob Koehler)i Subject: Re: Itanium troublesd3 Message-ID: <63izk9weVF4X@eisner.encompasserve.org>-  V In article <ZC5d8.69$lj4.1925@news.get2net.dk>, "Dr. Dweeb" <Dweeb@NoSpam.com> writes: > L > The jury is out on this and I remain unconvinced that these people (CarleyM > and Curly) are not just thinking with their collective "big swinging dicks"i4 > a process not at all related to rationmal thought. >   E    Ah, so you think that's an area where they come up a little short?i   ------------------------------  + Date: Fri, 22 Feb 2002 05:32:23 -0800 (PST)g. From: Fabio Cardoso <fabiopenvms@yahoo.com.br>0 Subject: Monitor lock and monitor io questions !@ Message-ID: <20020222133223.94304.qmail@web20207.mail.yahoo.com>  1 I upgraded my Alphaserver 4100 memory from 4GB to " 6GB. My software configuration is:  , OpenVMS is 7.2 / RDB 7.0-31 and Gembase 6.0.    . I have some questions because I think for some0 specific applciations I didnt see an improvment.. My application are much more io / memory bound0 I have 4 CPUs EV56 and they allways stay between 200-300 % at the monitor sys.   . I hope this goes formatet below.  It is my mon, lock/curr and I believe the values are high.  3                             OpenVMS Monitor Utility 5             +-----+        LOCK MANAGEMENT STATISTICS /             | CUR |              on node C16000f3             +-----+         22-FEB-2002 10:23:53.49   6                                      0         5         10        15        206                                      + - - - - + - - - - + - - - - + - - - - -+"  New ENQ Rate                    |"  Converted ENQ Rate              |6                                      |         |         |         |          |"  DEQ Rate                        |$  Blocking AST Rate                 |6                                      |         |         |         |          |$  ENQs Forced To Wait Rate          |#  ENQs Not Queued Rate             |t6                                      |         |         |         |          |&  Deadlock Search Rate                |&  Deadlock Find Rate                  |6                                      |         |         |         |          |   Total Locks                   |!  Total Resources                | 6                                      |         |         |         |          |6                                      + - - - - + - - - - + - - - - + - - - - -+      ) Below are my monitor io/curr informationsr  3                             OpenVMS Monitor Utilityd2             +-----+          I/O SYSTEM STATISTICS/             | CUR |              on node C16000v3             +-----+         22-FEB-2002 10:30:47.63,  6                                      0         20        40        60        806                                      + - - - - + - - - - + - - - - + - - - - -+"  Direct I/O Rate                 |#  Buffered I/O Rate                | %  Mailbox Write Rate                 |Y&  Split Transfer Rate                 |"  Log Name Translation Rate       |$  File Open Rate                    |6                                      |         |         |         |          |"  Page Fault Rate                 |#  Page Read Rate                   | $  Page Read I/O Rate                |&  Page Write Rate                     |&  Page Write I/O Rate                 |&  Inswap Rate                         |   Free List Size                |!  Modified List Size             |f6                                      + - - - - + - - - - + - - - - + - - - - -+6                                                                                 n           =====  ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazili fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?2 Yahoo! Sports - Coverage of the 2002 Olympic Games http://sports.yahoo.comt   ------------------------------  + Date: Fri, 22 Feb 2002 05:44:00 -0800 (PST)r. From: Fabio Cardoso <fabiopenvms@yahoo.com.br>A Subject: Re: Monitor lock and monitor io questions ! - correction;@ Message-ID: <20020222134400.11481.qmail@web20204.mail.yahoo.com>  3 --- Fabio Cardoso <fabiopenvms@yahoo.com.br> wrote:u3 > I upgraded my Alphaserver 4100 memory from 4GB toh$ > 6GB. My software configuration is: > . > OpenVMS is 7.2 / RDB 7.0-31 and Gembase 6.0. >  > 0 > I have some questions because I think for some2 > specific applciations I didnt see an improvment.0 > My application are much more io / memory bound2 > I have 4 CPUs EV56 and they allways stay between > 200-300 % at the monitor sys.  > 0 > I hope this goes formatet below.  It is my mon. > lock/curr and I believe the values are high. >    CORRECTr   MONITOR IO/CURRENT  # Direct I/O Rate                1369c Buffered I/O Rate		322 Mailbox Write Rate		  7h   Split Transfer Ratee$ Log Name Translation Rate        814$ File Open Rate                    20  # Page Fault Rate                 723 # POage Read Rate                 177i# Page Read I/O Rate               85o Page Write Ratee Page Write I/O Rateo Inswap% Free List Size                 307459e% Modified List Size              86385      MONITOR LOCK/CURRENT   New ENQ Rate			1565u Converted ENQ Rate		 348   DEQ Rate 			1502 Blocking AST Rate		  30r  $ ENQ Forced To Wait Rate           17$ ENQ Not Queued Rate              154   Deadlock Search Rate Deadlock Find RAte   Total Locks			180197 Total Resources			 26838         =====Y ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazilh fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?2 Yahoo! Sports - Coverage of the 2002 Olympic Games http://sports.yahoo.coma   ------------------------------  % Date: Fri, 22 Feb 2002 11:30:37 -0500 2 From: "Sue Skonetski" <susan.skonetski@compaq.com>! Subject: OpenVMS IPF presentationc0 Message-ID: <clud8.78$fL6.1043@news.cpqcorp.net>   Dear Newsgroup,F  C If you are interested in what we are doing with porting the OpenVMSpF Operating System and Application to the Itanium Processor Family & theH Compiler Roadmaps here is the link.  This just went up on the web today.  D In particular if you have seen Gaitan Dantoni's presentation you may remember this prostrationl5 http://www.compaq.com/hps/ipf-enterprise/openvms.htmlo  
 Warm Regards,    Suei   ------------------------------  # Date: Fri, 22 Feb 2002 17:29:56 GMTr# From: "John Smith" <a@nonymous.com> % Subject: Re: OpenVMS IPF presentationi0 Message-ID: <oevd8.4691$4xE.2076@news1.bloor.is>   Sue,  D About 16 slides into the OpenVMS Customer Presentation, there is theE timeline showing the 'Alpha-Itanium Processor Family Operating Systemh	 Roadmap'.r  L I note that the green arrow labeled 'OpenVMS on Itanium Processor Family' isE noticably shorter than all the others, even Tru64, which I though was . already given last rites. What does this mean?        = "Sue Skonetski" <susan.skonetski@compaq.com> wrote in message * news:clud8.78$fL6.1043@news.cpqcorp.net... > Dear Newsgroup,o >cE > If you are interested in what we are doing with porting the OpenVMSrH > Operating System and Application to the Itanium Processor Family & theJ > Compiler Roadmaps here is the link.  This just went up on the web today. > F > In particular if you have seen Gaitan Dantoni's presentation you may > remember this prostrationm7 > http://www.compaq.com/hps/ipf-enterprise/openvms.htmlh >p > Warm Regards,C >R > Sue  >s >- >  >A >2   ------------------------------    Date: 22 Feb 2002 12:09:01 -0600+ From: young_r@encompasserve.org (Rob Young):% Subject: Re: OpenVMS IPF presentationa3 Message-ID: <RxjUAyboV7Kf@eisner.encompasserve.org>r  V In article <oevd8.4691$4xE.2076@news1.bloor.is>, "John Smith" <a@nonymous.com> writes: > Sue, > F > About 16 slides into the OpenVMS Customer Presentation, there is theG > timeline showing the 'Alpha-Itanium Processor Family Operating System  > Roadmap'.i > N > I note that the green arrow labeled 'OpenVMS on Itanium Processor Family' isG > noticably shorter than all the others, even Tru64, which I though wasv0 > already given last rites. What does this mean? >    	Good catch!  A 	As you have noticed, those arrows don't meant a continuation.... B 	those are special "stopping arrows" and as you duly noted OpenVMSC 	is discountined in March 2009, Linux is discountined in June 2009, G 	but indeed... Tru64 and Windows make it all the way until October 2009a. 	when BeOS is expected to take over the world.   				Robi   ------------------------------  % Date: Fri, 22 Feb 2002 12:56:59 -0500 5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>V% Subject: Re: OpenVMS IPF presentation / Message-ID: <tIvd8.87$fL6.706@news.cpqcorp.net>u  J Now you want to read meaning into the ability of someone creating graphicsD by hand in a slide?  Why not ask why the arrow for Linux is so thin?       John Smith wrote in message ...e >Sue,< >nE >About 16 slides into the OpenVMS Customer Presentation, there is themF >timeline showing the 'Alpha-Itanium Processor Family Operating System
 >Roadmap'. >uJ >I note that the green arrow labeled 'OpenVMS on Itanium Processor Family' isF >noticably shorter than all the others, even Tru64, which I though was/ >already given last rites. What does this mean?  >r >a >  >y> >"Sue Skonetski" <susan.skonetski@compaq.com> wrote in message+ >news:clud8.78$fL6.1043@news.cpqcorp.net...  >> Dear Newsgroup, >>F >> If you are interested in what we are doing with porting the OpenVMSI >> Operating System and Application to the Itanium Processor Family & the K >> Compiler Roadmaps here is the link.  This just went up on the web today.  >>G >> In particular if you have seen Gaitan Dantoni's presentation you maye >> remember this prostration8 >> http://www.compaq.com/hps/ipf-enterprise/openvms.html >> >> Warm Regards, >> >> Sue >> >> >> >> >> >o >g   ------------------------------  % Date: Fri, 22 Feb 2002 02:27:59 -0500 6 From: "Christopher Duncan" <Chris@ShowProgramming.com>$ Subject: OT - Programmer war stories? Message-ID: <Anmd8.3817$2v1.204919@e3500-atl1.usenetserver.com>d  
 Hi, folks.  0 I'm trying to reach as many seasoned programmers1 as I can, hope you don't mind the off topic post.a1 As part of the research I'm doing for a follow up * to my recent book, "The Career Programmer:2 Guerilla Tactics for an Imperfect World" (Apress),1 I'm compiling a book of war stories from those of 0 you who have been out there dealing with all the- insanity that Corporate America has to offer.e  ( In The Career Programmer, I talked about/ overcoming the real world obstacles we run into,* trying to get decent software out the door* (arbitrary deadlines, clueless management,1 Marketing, fuzzy requirements, never any time forn1 testing, corporate politics - all that stuff they-/ never taught us in school). I've had tremendous70 feedback to the book, and one of the things that. struck me was that as I told stories about the. ridiculous situations I've had to cope with in. my own career, everyone asked if I was looking* over their cubicle walls because they were/ dealing with the same things in their shop. So,n, if figured it would be fun to do a follow up. book relating the war stories that some of you/ have been through. None of my books are tied tos, a particular programming language, operating- system or platform, so it doesn't matter whatm+ kind of coding you do. We all deal with the? same sort of craziness.r  , I'll paraphrase the stories, and of course I0 won't use any real names (personal or company) -0 we get enough grief from management as it is, no1 point inviting any more. I'll be happy to include:0 your name in the Acknowledgements, but of course+ will omit it if you prefer your story to be6- anonymous. This book will be in the same, er, / conversational style (that's what my publishersl. told me was the politically correct phrase for/ being a wise guy) as The Career Programmer was.h. I try not to take myself too seriously in this1 business - keeps my antacid bills down. But then,h1 if you've read the last book you probably figuredf+ that out about the time you encountered the $ night watchman's attack Chihuahua...  , Anyway, if you've got a good war story about- programming in Corporate America, I'd love tot0 hear it. I may change the format of how I gather1 & organize the stories on my web site, so just goi2 to www.ShowProgramming.com/TheCareerProgrammer.asp0 and look for the link on War Stories. That gives/ me the flexibility to reorganize the database &r4 data input page whenever I want without leaving dead3 links. And of course, you know I'll have to rewritei1 the UI at least a couple of times. There's alwayse one more tweak.e  3 Oh, by the way, since I get emails from programmers 0 all over the world, I get teased a little here &0 there about my use of "Corporate America", since4 programming is obviously an international community.7 And from what I've heard, the business world is equally.5 insane in every country. However, America is the only:6 one I have experience with. And besides, it was easier3 to write than Corporate <Insert Your Country Here>..  Hey, I got lazy, what can I say?  7 Anyway, if you've got a good story about how crazy this.5 business can get and wouldn't mind me sharing it with 2 others, please drop by and add it to the database.7 Make sure you leave me an email address that will stills5 be good when the book comes out (these things tend toi3 take months to get out the door - sound familiar?).>  1 And thanks very much for taking the time, I trulyo appreciate it.  
 Happy Coding!r   Christopher Duncan   ------------------------------  % Date: Fri, 22 Feb 2002 14:13:45 +0100n/ From: Nigel Arnot <sysmgr@maxwell.ph.kcl.ac.uk> 9 Subject: re: OT gassless cars Was Re: "Crashless Windows"c7 Message-ID: <00A09F33.9A898C81.17@maxwell.ph.kcl.ac.uk>>   >  lP > well, you can apparently use used cooking oil which the food industry tends toQ > just dump otherwise.  Cleanliness, I can't remember the details on that, sorry.e > Z > However, apparently it is necessary in the UK to pay fuel tax on any cooking oil you use > to run your car :-(.    J IANAL, but I suspect not if it's used cooking oil that would normally haveH negative value (ie you'd have to pay someone to take it away and legallyF dump it). Any percentage of zero is zero. Plus its an environmentally-I friendly thing to do (run a car on a renewable crop resource that adds nos# carbon dioxide to the environment).d  H And if they did tax it, they'd have to tax cooking oil at the same rate,G and the outcry that would cause is inconceivable. (If they didn't thereoI would be massive fraud, quite probably including the fraudulent claiming tG back of tax not actually ever paid, as used to happen with VAT on gold)n     	Yours, 
 		Nigel Arnott- 		NRA@MAXWELL.PH.KCL.AC.UK                   e  7 		"In the beginning there was nothing, which exploded."d   ------------------------------    Date: 22 Feb 2002 09:59:46 -0600- From: koehler@encompasserve.org (Bob Koehler)s9 Subject: Re: OT gassless cars Was Re: "Crashless Windows"t3 Message-ID: <C1bOasvxEUqf@eisner.encompasserve.org>h  g In article <3C743188.E8898616@blueyonder.co.uk>, Tim Llewellyn <tim.llewellyn@blueyonder.co.uk> writes:o  M > according to a recent motoring programon TV in the UK people are converting . > diesel cars to run on ordinary cooking oil.   B    Diesel tried to design an engine that can run on any fuel.  His    favorite was coal dust.  H    He did fairly well.  Peanut farmers used to run their diesel tractorsE    on peanut oil when it was cheap.  Must have been pretty cheap, toohG    since most states don't charge fuel taxes on farmers like they do onn    truckers.   ------------------------------    Date: 22 Feb 2002 10:07:53 -0600- From: koehler@encompasserve.org (Bob Koehler)e9 Subject: Re: OT gassless cars Was Re: "Crashless Windows"e3 Message-ID: <OIuINGgi4e0s@eisner.encompasserve.org>C  i In article <55f85d77.0202210434.2908924e@posting.google.com>, P.Young@unsw.EDU.AU (Patrick Young) writes:-   > At AUS$1.0 I would try to2H > source home heating oil used in some states during the 1ms time period/ > that is winter here or use such alternatives.4  H    Here in the US, the only real difference between home heating oil andC    diesel fuel is the tax paid on the diesel fuel and the dye addedt    to show the tax was paid.  G    Since agriculture was exempt from that tax my grandfather always ranl&    his deisel tractors on heating oil.  G    Major fine to get caught running a motor vehicle without having paids    that tax.   ------------------------------  % Date: Fri, 22 Feb 2002 16:13:47 +0000r% From: Alan Greig <a.greig@virgin.net>r9 Subject: Re: OT gassless cars Was Re: "Crashless Windows"t8 Message-ID: <15rc7u09mqhe93433i8s184d9clvlujk7v@4ax.com>  / On Fri, 22 Feb 2002 14:13:45 +0100, Nigel Arnotr$ <sysmgr@maxwell.ph.kcl.ac.uk> wrote:   > K >IANAL, but I suspect not if it's used cooking oil that would normally havesI >negative value (ie you'd have to pay someone to take it away and legallyaG >dump it). Any percentage of zero is zero. Plus its an environmentally-y  F Fuel tax is not a percentage. It is a fixed amount per litre. Think ofD the budget: "The chancellor today announced a 1.5p rise per litre on fuel"n  J >friendly thing to do (run a car on a renewable crop resource that adds no$ >carbon dioxide to the environment). >CI >And if they did tax it, they'd have to tax cooking oil at the same rate,   E Nope just as diesel for farm tractors is not charged at the same rate  as diesel for cars.   H >and the outcry that would cause is inconceivable. (If they didn't thereJ >would be massive fraud, quite probably including the fraudulent claiming   F Well they get round that with diesel by requiring different colours byF law for different uses. If you are caught with low tax diesel in a car> then the car is confiscated. This relies on informants mainly.  E However the UL government has not (to best of my knowledge) initiatedEE legal action against anyone using plant based diesel. The German govtt" seems to be testing the waters.... -- Alan   ------------------------------  % Date: Fri, 22 Feb 2002 10:00:49 -0800|, From: "James Gessling" <jgessling@yahoo.com>9 Subject: Re: OT gassless cars Was Re: "Crashless Windows" 4 Message-ID: <a5610j$4uft7$1@ID-46415.news.dfncis.de>  K We are way off topic but enjoying it.  For details of making biodiesel fuel  see:  / http://journeytoforever.org/biodiesel_mike.html    Jim    ------------------------------  % Date: Fri, 22 Feb 2002 12:23:02 -0500t  From: norm.raphael@jamesbury.com Subject: Re: PGP for OpenVMS?o4 Message-ID: <C2256B68.005F2099.00@jklh21.valmet.com>  9 Can anyone confirm or deny or elaborate on the following:v  N "The latest mention of it on comp.os.vms from Glenn and Mary Everhart suggestsM it's on the Fall 2001 Sig Tapes in the [NET] directory. I haven't looked so I A don't know whether this is still GNUPG 1.0.4 or a later version."    -Norm    ------------------------------  + Date: Fri, 22 Feb 2002 17:38:23 +0000 (UTC)  From: david20@alpha1.mdx.ac.uk Subject: Re: PGP for OpenVMS? + Message-ID: <a55vmf$2nn$1@aquila.mdx.ac.uk>   W In article <C2256B68.005F2099.00@jklh21.valmet.com>, norm.raphael@jamesbury.com writes:  >  > : >Can anyone confirm or deny or elaborate on the following: > O >"The latest mention of it on comp.os.vms from Glenn and Mary Everhart suggests N >it's on the Fall 2001 Sig Tapes in the [NET] directory. I haven't looked so IB >don't know whether this is still GNUPG 1.0.4 or a later version." >  >-Norm >   3 I've now looked on ftp://mvb.saic.com/vmslt01b/net/   and it appears to be GNUPG 1.0.6  ' However the file is in .TAR-BZ2 format. M I seem to recall seeing a decompression program which could handle .BZ2 files L some time ago on VMS but haven't got it on my system. I'd expect to find theK decompression tools on the same system as the ftp archive but can't seem to_J spot them. Anyone know where I can pick up a tool to decompress this file.    
 David Webb VMS and Unix team leader CCSS Middlesex University     ------------------------------  + Date: Fri, 22 Feb 2002 18:05:54 +0000 (UTC)h From: david20@alpha1.mdx.ac.uk Subject: Re: PGP for OpenVMS?!+ Message-ID: <a561a2$3dp$1@aquila.mdx.ac.uk>1  L In article <a55vmf$2nn$1@aquila.mdx.ac.uk>, david20@alpha1.mdx.ac.uk writes:X >In article <C2256B68.005F2099.00@jklh21.valmet.com>, norm.raphael@jamesbury.com writes: >> >>; >>Can anyone confirm or deny or elaborate on the following:  >>P >>"The latest mention of it on comp.os.vms from Glenn and Mary Everhart suggestsO >>it's on the Fall 2001 Sig Tapes in the [NET] directory. I haven't looked so ICC >>don't know whether this is still GNUPG 1.0.4 or a later version."s >> >>-Normo >> > 4 >I've now looked on ftp://mvb.saic.com/vmslt01b/net/! >and it appears to be GNUPG 1.0.6  >R( >However the file is in .TAR-BZ2 format.N >I seem to recall seeing a decompression program which could handle .BZ2 filesM >some time ago on VMS but haven't got it on my system. I'd expect to find the L >decompression tools on the same system as the ftp archive but can't seem toK >spot them. Anyone know where I can pick up a tool to decompress this file.W >e >O  N Sorry should have looked a bit harder before posting. bzip2 is on the freeware 5.0 cd.8  
 David Webb VMS and Unix team leader CCSS Middlesex University  
 David Webb VMS and Unix team leader CCSS Middlesex University     ------------------------------  % Date: Fri, 22 Feb 2002 04:32:53 -0500k  From: John Santos <JOHN@egh.com>F Subject: Re: Poor Performance with OpenVMS Advanced Server and TCPWare6 Message-ID: <1020222042850.42335A-100000@Ives.egh.com>  " On 22 Feb 2002, Ian Burgess wrote:  * >  Alan Greig <a.greig@virgin.net> writes: > >=: > >On Wed, 20 Feb 2002 11:26:34 -0500, "McCarthy Kevin P." > ><McCarthyKP@BWSC.ORG> wrote:o > >eK > >>We are on the verge of moving all of our Pathworks shares to NT/Win2000-L > >>servers because of this.  We are running Pathworks 6.1 and TCPware on anJ > >>ES40 833 MHz with HSG 80s and Gig Ethernet, but for serving data filesI > >>to AutoCAD 2002 a 866 MHz Win NT server is 1 or 2 orders of magnitudes@ > >>faster.  A single user with winzip made our PW server crawl. > >dH > >With  UCX SET PROT TCP/NODELAY_ACK I see uploads to Pathworks serversF > >go from 300KB/sec to around 3MBsec. Unfortunately, as you say, thisH > >workaround does not seem to be available for TCPWARE. But for extremeI > >slow read performance you want to stick the Pathworks cache way up (ifsC > >you have not already done so) The default is ludicrously low. In2H > >ADMIN/CONFIG (was this still run by a .COM file in 6.1?) I have  DataE > >Cache Size (Kbytes): 65536. I get about 5MB/sec on read with this.  > >  > >--  > >Alan  > K > Pathworks performance has always been a concern and the major application-J > is going to another platform because of it.  (Data cache size is 65536.) > H > An improvement even for the final few months would be worthwhile, so I( > tried setting no delay on acknowledge. > , > But I seem to be missing something here... > " > From SYSTEM with all privileges, >  >  $ tcpip show versioni > D >    DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A - ECO 1; >    on a AlphaServer 1200 5/533 4MB running OpenVMS V7.2-1n' >  $ tcpip show protocol tcp /parameterp >  TCP$ >    Delay ACK:              enabled >    ...! >  $ tcpip set proto tcp /NODelay   $ $ tcpip set config proto tcp/nodelay ------------^^^^^^   Then stop and restart tcp/ip.a  ? This is just a guess, based on HELP and other TCP/IP behaviour./  C Maybe it can't change the "/NODELAY" parameter while it is running, > so "tcpip set proto ..." does nothing.  "tcpip set config ..."= changes the permanent database, but has no effect on the liveg= system, so you need to stop and restart TCP/IP for the configb change to take effect.   HTH.    ( >  $ tcpip show protocol tcp /parameter  >  TCP$ >    Delay ACK:              enabled >    ... > > > Restart UCX (TCP/IP Services for OpenVMS) or even reboot and0 > "show protocol" still shows Delay ACK enabled. > ( > The Help says that command is correct. >    > Ian Burgess  > University of Queensland > I.Burgess[at]its.uq.edu.au > www.its.uq.edu.au  >  >    -- h John Santos6 Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Fri, 22 Feb 2002 10:52:29 +0100o& From: A Razzak <razzakaSP@Mcompaq.com>F Subject: Re: Poor Performance with OpenVMS Advanced Server and TCPWare+ Message-ID: <3C7614DD.AC1C9482@Mcompaq.com>    Ian Burgess wrote: > * >  Alan Greig <a.greig@virgin.net> writes: > >O: > >On Wed, 20 Feb 2002 11:26:34 -0500, "McCarthy Kevin P." > ><McCarthyKP@BWSC.ORG> wrote:- > >-K > >>We are on the verge of moving all of our Pathworks shares to NT/Win2000eL > >>servers because of this.  We are running Pathworks 6.1 and TCPware on anJ > >>ES40 833 MHz with HSG 80s and Gig Ethernet, but for serving data filesI > >>to AutoCAD 2002 a 866 MHz Win NT server is 1 or 2 orders of magnitudew@ > >>faster.  A single user with winzip made our PW server crawl. > >.H > >With  UCX SET PROT TCP/NODELAY_ACK I see uploads to Pathworks serversF > >go from 300KB/sec to around 3MBsec. Unfortunately, as you say, thisH > >workaround does not seem to be available for TCPWARE. But for extremeI > >slow read performance you want to stick the Pathworks cache way up (if C > >you have not already done so) The default is ludicrously low. IntH > >ADMIN/CONFIG (was this still run by a .COM file in 6.1?) I have  DataE > >Cache Size (Kbytes): 65536. I get about 5MB/sec on read with this.  > >r > >--> > >Alano > K > Pathworks performance has always been a concern and the major applicationeJ > is going to another platform because of it.  (Data cache size is 65536.) > H > An improvement even for the final few months would be worthwhile, so I( > tried setting no delay on acknowledge. > , > But I seem to be missing something here... > " > From SYSTEM with all privileges, >  >  $ tcpip show versione > D >    DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A - ECO 1; >    on a AlphaServer 1200 5/533 4MB running OpenVMS V7.2-1s' >  $ tcpip show protocol tcp /parameter> >  TCP$ >    Delay ACK:              enabled >    ...! >  $ tcpip set proto tcp /NODelayt' >  $ tcpip show protocol tcp /parametert >  TCP$ >    Delay ACK:              enabled >    ... > > > Restart UCX (TCP/IP Services for OpenVMS) or even reboot and0 > "show protocol" still shows Delay ACK enabled. > ( > The Help says that command is correct. > 
 > Ian Burgess' > University of Queensland > I.Burgess[at]its.uq.edu.au > www.its.uq.edu.aue  = 	$ SET CONFIG PROTO TCP /NODELAY      ! in Permanent databasee 	  -- Abdul Razzak
 Munich, GY   ------------------------------  % Date: Fri, 22 Feb 2002 10:15:42 +0000 % From: Alan Greig <a.greig@virgin.net>yF Subject: Re: Poor Performance with OpenVMS Advanced Server and TCPWare8 Message-ID: <6h6c7ukvmc4g0bt8d68d9ohqo6ipr573tp@4ax.com>  C On 22 Feb 2002 00:34:30 GMT, ccburgess@uqstu.jdstory.uq.edu.au (Ianh Burgess) wrote:f   >t >gC >   DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A - ECO 1t: >   on a AlphaServer 1200 5/533 4MB running OpenVMS V7.2-1& > $ tcpip show protocol tcp /parameter > TCPb# >   Delay ACK:              enabledm >   ...i  > $ tcpip set proto tcp /NODelay' > $ tcpip show protocol tcp /parameter t > TCP # >   Delay ACK:              enablede  5 The display is incorrect. I see the same on my systemo      g TCPIP> sho ver  A   DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A - ECO 2y5   on a Compaq AlphaServer ES40 running OpenVMS V7.2-1   C but the change *has* taken effect. Try an upload of a large file tosC the Pathworks server. I think ECO 2 has some fixes which autodetects@ certain Windows stacks and get the delayed acks timing right. WeE tested a pre-release versiojn of this patch and it worked for most NTu V4 systems but not W2K.. ...  >k= >Restart UCX (TCP/IP Services for OpenVMS) or even reboot andy/ >"show protocol" still shows Delay ACK enabled.  >a' >The Help says that command is correct.r > 
 >Ian Burgess r >University of Queenslandr >I.Burgess[at]its.uq.edu.aue >www.its.uq.edu.au   -- Alan   ------------------------------  % Date: Fri, 22 Feb 2002 13:47:31 +0100e= From: Oswald Knoppers <Oswald.Knoppers@contrastmediagroep.nl>.F Subject: Re: Poor Performance with OpenVMS Advanced Server and TCPWare5 Message-ID: <3C763DE3.8D1CE4A5@contrastmediagroep.nl>m   Alan Greig wrote:  > E > On 22 Feb 2002 00:34:30 GMT, ccburgess@uqstu.jdstory.uq.edu.au (Ianm > Burgess) wrote:r >  > >  > >fE > >   DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A - ECO 1e< > >   on a AlphaServer 1200 5/533 4MB running OpenVMS V7.2-1( > > $ tcpip show protocol tcp /parameter > > TCPt% > >   Delay ACK:              enabled 	 > >   ...e" > > $ tcpip set proto tcp /NODelay( > > $ tcpip show protocol tcp /parameter > > TCPl% > >   Delay ACK:              enabled  > 7 > The display is incorrect. I see the same on my systemg >  >  > TCPIP> sho ver > C >   DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A - ECO 2q7 >   on a Compaq AlphaServer ES40 running OpenVMS V7.2-1o > E > but the change *has* taken effect. Try an upload of a large file to E > the Pathworks server. I think ECO 2 has some fixes which autodetect?B > certain Windows stacks and get the delayed acks timing right. WeG > tested a pre-release versiojn of this patch and it worked for most NTa > V4 systems but not W2K.    You might check with sysconfig:r  # TCPIP> sysconfig -q inet nodelayackb  E Should show 0 if delay ack is enabled and 1 if delay ack is disabled.o   Regards,   Oswald   ------------------------------  % Date: Fri, 22 Feb 2002 15:02:32 +0000o% From: Alan Greig <a.greig@virgin.net>-F Subject: Re: Poor Performance with OpenVMS Advanced Server and TCPWare8 Message-ID: <t3nc7uoag8219qetfae1ngck3km66e0ng5@4ax.com>  3 On Fri, 22 Feb 2002 13:47:31 +0100, Oswald KnoppersD. <Oswald.Knoppers@contrastmediagroep.nl> wrote:   >l  >You might check with sysconfig: > $ >TCPIP> sysconfig -q inet nodelayack > F >Should show 0 if delay ack is enabled and 1 if delay ack is disabled.$ TCPIP> sysconfig -q inet tcpnodelack inet:i tcpnodelack = 1e   TCPIP> sho prot tcp/parc TCPg!   Delay ACK:              enabled.!   Window scale:           enabledt!   Drop count:                   0a!   Probe timer:                  0h  ? If I change with the SET PROT TCP command the tcpnodelack valuen; changes but the display shown by SHO PROT TCP/PAR does not.t   BUG!!o -- Alan   ------------------------------  % Date: Fri, 22 Feb 2002 08:22:00 +0100m( From: Rainer Lehrig <lehrig@t-online.de>= Subject: Re: POSIX for OpenVMS - is there something similar ?o+ Message-ID: <3C75F198.4BE31A96@t-online.de>h   Thank you very much !a   Jouk Jansen wrote: >  > Rainer Lehrig wrote: > > : > > The product "POSIX for OpenVMS" was retired in VMS 7.2L > > If I got it right you had the opportunity to run a unix compatible shell > > "psx>". K > > Within this shell you could use files in unix syntax. - Exactly this is- > > what I'm searching for.b > >  > > Problem:@ > > I have a portable GUI library Qt (http://www.trolltech.com).H > > This library runs on all Unix-Systems, Windows and MAC and embedded.4 > > I have ported an older version of Qt to OpenVMS.@ > > There you had "only" to handle the filenames (VMS <-> UNIX). > >tG > > I'm searching for a subsystem like "POSIX for OpenVMS", which wouldeA > > allow me to run programs which think of files in unix syntax.DB > > (DEVICE:[dir.subdir]file.ext  <-> /device/dir/subdir/file.ext) > >t= > I think you will have to look for the following 2 packages: ! > 1) The OpenVMS porting library: A > http://www.openvms.compaq.com/openvms/products/ips/porting.html ( > which is a great help at compile time. > 2) GNV' > http://gnv.sourceforge.net/readme.htm- > For a Unix-like shell. >  >                 Jouk   ------------------------------    Date: 22 Feb 2002 10:18:40 -0600- From: koehler@encompasserve.org (Bob Koehler)e= Subject: Re: POSIX for OpenVMS - is there something similar ?e3 Message-ID: <HpwXl7C6fxVW@eisner.encompasserve.org>.  V In article <3C74A657.D6A56B5C@t-online.de>, Rainer Lehrig <lehrig@t-online.de> writes:  E > I'm searching for a subsystem like "POSIX for OpenVMS", which wouldt? > allow me to run programs which think of files in unix syntax. @ > (DEVICE:[dir.subdir]file.ext  <-> /device/dir/subdir/file.ext)  G    All the I/O routines in the C library already do this.  Not perfect,i?    but probably what you want and they're getting better.  i.e.r7    with no code changes the following WILL work on VMS:          FILE *ip;c  4       ip = fopen("/device/dir/subdir/file.ext","r");  '    You also get this with C++ and Java.    ------------------------------  % Date: Fri, 22 Feb 2002 18:51:14 +0900n& From: "David Lee" <phongle@kornet.net> Subject: Postscript printing+ Message-ID: <a553ls$mo5$1@news1.kornet.net>a  I I have no problems printing text file to my printer.  However, whenever IhL print postscript file to this printer, nothing happen.  Just like it doesn't" understand what I am sending over.  J It is a postscript printer type (HP 4000N).  Am I missing some software onL the ES-40 machine running VMS 7.2 or am I missing something from the printer itself.n   Does anyone  have any ideas?  	 Thank youf   ------------------------------  % Date: Fri, 22 Feb 2002 05:22:13 -0500l- From: JF Mezei <jfmezei.spamnot@videotron.ca>e  Subject: Re: Postscript printing, Message-ID: <3C761BD4.E18C1503@videotron.ca>   David Lee wrote: > K > I have no problems printing text file to my printer.  However, whenever ImN > print postscript file to this printer, nothing happen.  Just like it doesn't$ > understand what I am sending over.  N You mean the printer prints the postscript file as if it were a text file (eg:7 lists the postscript program instead of executing it ?)b  K Is your printer setup to autodetect print type (text, HPCL, Postscript), orh set to print postscript only ?  M Try to set the printer to print postscript only (no auto detect) and see whatm# happens when you print a text file.   L also try print/para=(data_type=POSTSCRIPT) to force the printing software to. treat the print file as a postscript progrtam.   ------------------------------  % Date: Fri, 22 Feb 2002 10:36:21 -0500|0 From: Paul Anderson <paul.r.anderson@compaq.com>  Subject: Re: Postscript printing; Message-ID: <220220021036211714%paul.r.anderson@compaq.com>n  5 In article <3C761BD4.E18C1503@videotron.ca>, JF MezeiF% <jfmezei.spamnot@videotron.ca> wrote:e  N > also try print/para=(data_type=POSTSCRIPT) to force the printing software to0 > treat the print file as a postscript progrtam.  < Assuming, of course, that you are running DCPS, which is the= recommended way to print to PostScript printers from OpenVMS.   F If you are running DCPS, put the printer in auto-sensing or PostScriptC mode.  If not, and you want to print text files, put the printer in  auto-sensing mode.   Paul   -- t  Paul Anderson   OpenVMS Engineering    Compaq Computer Corporations   ------------------------------  % Date: Fri, 22 Feb 2002 02:40:56 -0500-- From: JF Mezei <jfmezei.spamnot@videotron.ca> 2 Subject: Re: Profitability and the survival of VMS, Message-ID: <3C75F607.BF161F32@videotron.ca>  # Rob Young wrote list of C2 systems:B2 > Data General Corporation AOS/VS II, Release 3.10A > Digital Equipment Corporation OpenVMS VAX and Alpha Version 6.1eQ > Microsoft Corporation Windows NT Workstation and Windows NT Server, Version 4.0r  L VMS is not in very good company in that rating. Of course, if NT can get it,K it means that the rating is meaningless. Also, the fact that the rating wasr$ not maintained past 6.1 tells a lot.   ------------------------------    Date: 22 Feb 2002 11:13:16 -0600- From: koehler@encompasserve.org (Bob Koehler)m2 Subject: Re: Profitability and the survival of VMS3 Message-ID: <d9hkM4qMOiMc@eisner.encompasserve.org>S  \ In article <3C75F607.BF161F32@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes:% > Rob Young wrote list of C2 systems: 3 >> Data General Corporation AOS/VS II, Release 3.10dB >> Digital Equipment Corporation OpenVMS VAX and Alpha Version 6.1R >> Microsoft Corporation Windows NT Workstation and Windows NT Server, Version 4.0 > N > VMS is not in very good company in that rating. Of course, if NT can get it,M > it means that the rating is meaningless. Also, the fact that the rating wasu& > not maintained past 6.1 tells a lot.  G    C2 doesn't mean much.  But the NT system only qualifies if it is notlG    connected to a network.  Such details are the reality of what little     meaning there is.   ------------------------------    Date: 22 Feb 2002 07:53:59 -0800" From: VMSfan@hotmail.com (VMS Fan)J Subject: Re: Proof that EV8 was fantastic ... and Capellas "gave" it away!= Message-ID: <d0a53e6e.0202220753.746206da@posting.google.com>r  m bob@instantwhip.com (Bob Ceculski) wrote in message news:<d7791aa1.0202210820.3c4ac5d0@posting.google.com>...r@ >                  McKinley      EV7        Madison        EV8  B > Introduction       2H02        2H02       2H03 est.    2H04 est.E > Technology      0.18 um Al  0.18 um Cu   0.13 um Cu  0.13 um SOI/Cui@ > Clock Speed      1.0 GHz     1.2 GHz    1.8 GHz est.   1.8 GHzA > On-chip Cache   3.0 MB L3   1.75 MB L2   6.0 MB L3    3.0 MB L2iB > SPECint_base2k   800 est.    1050 est.    1250 est.    2000 est.F > SPECint_fp2k    1400 est.    1500 est.    2200 est.  3200/4500* est.$ > *Auto-parallelized, run with SMT.   E Yes, from these figures the EV8 appears to be over twice the speed of-C the McKinley.  But if you look closer, you'll see that EV8 here waskE estimated to come out 2 years after McKinley, and in a .13 um processO instead of .18.e  D Comparing EV8 with Madison would be a more fair comparison, but evenF there Madison is estimated here to come out a year earlier than EV8 --? and who knows whether Intel might have done a process shrink oft  Madison by the time EV8 arrived?  A Paul DeMone seems to be convinced that EV8 would still have had awD performance lead, and he's probably right.  But Compaq was unwilling# to pay the price to make it happen.   @ The sad fact, too, is that even during the many years that AlphaC enjoyed a 2x performance (and price/performance) advantage over the F competition, Digital (and later Compaq) still somehow didn't have whatF it took (either in ability, or in willpower) to make it succeed in the marketplace.  F I enjoyed listening to the AMD X86-64 presentation someone here kindlyF pointed out (http://murl.microsoft.com/LectureDetails.asp?690).  AMD'sC belief is that implementation is more important to performance than B instruction set.  (Which, ironically, actually supports EPIC, in a way.)   E (If the instruction set hasn't really turned out to matter that much, E that probably implies that, in retrospect, considering how well Intel E has done with X86, Digital could probably have made a VAX chip run asoA fast or faster than the SPARC, et. al., if it had been willing tocC devote the resources necessary to doing so.  Of course, by going too3 Alpha, we users also got the benefit of 64 bits...)r  A My biggest worry right now is not performance of IA64, but marketsD acceptance.  Overcoming the software-incompatibility hurdle will notF be easy.  (Here Intel ironically faces the same challenges as Alpha onF NT did.)  But Intel is apparently prepared and willing to switch to anB X86-64 clone rapidly if IA64 tanks.  But then that would force big, changes in a lot of HP's and Compaq's plans. ---s- "Rumors of my death are greatly exacerbated."o!                            -- VMS    ------------------------------  # Date: Fri, 22 Feb 2002 18:13:53 GMTn* From: "Bill Todd" <billtodd@metrocast.net>J Subject: Re: Proof that EV8 was fantastic ... and Capellas "gave" it away!C Message-ID: <BTvd8.156363$d34.11702958@bin8.nnrp.aus1.giganews.com>a  / "VMS Fan" <VMSfan@hotmail.com> wrote in messageu7 news:d0a53e6e.0202220753.746206da@posting.google.com...t5 > bob@instantwhip.com (Bob Ceculski) wrote in messagep9 news:<d7791aa1.0202210820.3c4ac5d0@posting.google.com>... @ > >                  McKinley      EV7        Madison        EV8D > > Introduction       2H02        2H02       2H03 est.    2H04 est.G > > Technology      0.18 um Al  0.18 um Cu   0.13 um Cu  0.13 um SOI/CuoB > > Clock Speed      1.0 GHz     1.2 GHz    1.8 GHz est.   1.8 GHzC > > On-chip Cache   3.0 MB L3   1.75 MB L2   6.0 MB L3    3.0 MB L20D > > SPECint_base2k   800 est.    1050 est.    1250 est.    2000 est.H > > SPECint_fp2k    1400 est.    1500 est.    2200 est.  3200/4500* est.% > > *Auto-parallelized, run with SMT.r >dG > Yes, from these figures the EV8 appears to be over twice the speed oftE > the McKinley.  But if you look closer, you'll see that EV8 here was G > estimated to come out 2 years after McKinley, and in a .13 um processo > instead of .18.a >oF > Comparing EV8 with Madison would be a more fair comparison, but evenH > there Madison is estimated here to come out a year earlier than EV8 --A > and who knows whether Intel might have done a process shrink off" > Madison by the time EV8 arrived?  E If Intel could have done one, so could EV8:  IBM doesn't lag Intel inwJ process technology, and EV8 was designed with 100 nm process technology asJ its sweet spot.  But according to Intel's schedules the Itanic line is farH from the first to receive shrinks when the technology allows them (if itE were, McKinley would have been released at 130 nm).  So a 60% SPECint E performance advantage at EV8 release over the Itanic competition is as reasonable estimate.  L But of course SPECint tells only a small part of the story.  Madison remainsB crippled by McKinley's PC-style bus approach to MP rather than theI EV7/POWER4/Hammer on-chip switching/routing facilities, and by McKinley'saH external memory controller rather than EV7's/Hammer's on-chip controllerA with far lower access latency.  And even if you ignore memory andtG inter-processor bandwidth effects and concentrate only on SPECint-style H measures, due to its SMT facilities each EV8 would have delivered over 3K times the throughput of each Madison in anything other than single-threaded " use (i.e., in typical server use).   - bill   ------------------------------  % Date: Fri, 22 Feb 2002 08:55:11 +0100V' From: Brass Christof <welcome@spam.net>a% Subject: Re: Screen Saver Problem (?) ( Message-ID: <3C75F95F.40B5F329@spam.net>   Jim Strehlow wrote:t > ; > I have survived various fixable problems with DECwindows. : > Solutions ranged from changing a setting on the graphics! > adapter board inside the box to 6 > loading OpenVMS patches for the proper graphics card > for the version of OpenVMS.r > 8 > You should mention what type of graphics card, version. > of OpenVMS, and maybe a PRODUCT SHOW HISTORY, > to see what patches are already installed.   Of course, as always.b  ! DS20E, single cpu 500 MHz, 512MB.-= Two PowerStorm 300, devices GBB0 (head 1) and GBA0 (head 2). r   $ product show history *  9 PRODUCT                             KIT TYPE    OPERATION ; ----------------------------------- ----------- -----------r7 DEC AXPVMS VMS721_UPDATE V3.0       Patch       Installu7 DEC AXPVMS VMS721_PCSI V1.0         Patch       Install 7 DEC AXPVMS DECNET_OSI V7.2-1        Full LP     Install07 DEC AXPVMS DWMOTIF V1.2-5           Full LP     Installp7 DEC AXPVMS OPENVMS V7.2-1           Platform    InstallC7 DEC AXPVMS TCPIP V5.0-10            Full LP     Installc7 DEC AXPVMS VMS V7.2-1               Oper System Installp   $ type decw$server_0_error.log  6 This is the DECwindows X11 display server for OpenVMS 3 AXP V7.1-001212 compiled on Dec 12 2000 at 16:34:31I ...    -- S6 moc dot slupofni at ssarb - please revert the sequence   ------------------------------   Date: 22 Feb 2002 18:15:26 GMT- From: "Jim Strehlow" <jimSnoSpam@data911.com>e% Subject: Re: Screen Saver Problem (?) 0 Message-ID: <a561ru$4qi@dispatch.concentric.net>  * Update v3.0 includes VMS721_GRAPHICS-V0400- so you should be okay on that side of things.c  A Some of the other patches that I have applied below are mandatorys and others are optional.$ Maybe something else needs patching?  L ----------------------------------- ----------- ----------- ---------------- ----9 PRODUCT                             KIT TYPE    OPERATIONrL ----------------------------------- ----------- ----------- ---------------- ----7 DEC AXPVMS VMS721_DCL V3.0          Patch       Installa7 DEC AXPVMS VMS721_SYSLOA V2.0       Patch       Installj7 DEC AXPVMS VMS721_LAN V3.0          Patch       Installd7 DEC AXPVMS VMS721_IPC V2.0          Patch       Installe7 DEC AXPVMS VMS721_BACKUP V2.0       Patch       Install37 DEC AXPVMS VMS721_DRIVER V4.0       Patch       Install)7 DEC AXPVMS VMS721_CLIUTL V3.0       Patch       InstalloF DEC AXPVMS VMS721_SYS V11.0         Patch       Install       (warning regarding reboot)M7 DEC AXPVMS VMS721_RMS V3.0          Patch       InstallC7 DEC AXPVMS VMS721_MOUNT96 V3.0      Patch       Install 7 DEC AXPVMS VMS721_MAIL V1.0         Patch       Installo7 DEC AXPVMS VMS721_LAT V1.0          Patch       Install 7 DEC AXPVMS VMS721_FIBRE_SCSI V5.0   Patch       Installo7 DEC AXPVMS VMS721_DW_MOT_MUP V1.0   Patch       InstallC7 DEC AXPVMS VMS721_AUDSRV V2.0       Patch       Install 7 DEC AXPVMS TCPIP_ECO V5.1-153       Patch       Install 7 DEC AXPVMS VMS721_LIBRTL V3.0       Patch       Install 7 DEC AXPVMS VMS721_UPDATE V3.0       Patch       Install    I hope that might help??  ! Jim Strehlow, Data911, Alameda CAe  4 "Brass Christof" <welcome@spam.net> wrote in message" news:3C75F95F.40B5F329@spam.net... > Jim Strehlow wrote:e > >t= > > I have survived various fixable problems with DECwindows.f< > > Solutions ranged from changing a setting on the graphics# > > adapter board inside the box to 8 > > loading OpenVMS patches for the proper graphics card > > for the version of OpenVMS.0 > >r: > > You should mention what type of graphics card, version0 > > of OpenVMS, and maybe a PRODUCT SHOW HISTORY. > > to see what patches are already installed. >. > Of course, as always.0 >3# > DS20E, single cpu 500 MHz, 512MB.l> > Two PowerStorm 300, devices GBB0 (head 1) and GBA0 (head 2). >t > $ product show history * >n; > PRODUCT                             KIT TYPE    OPERATIONa= > ----------------------------------- ----------- -----------t9 > DEC AXPVMS VMS721_UPDATE V3.0       Patch       Install/9 > DEC AXPVMS VMS721_PCSI V1.0         Patch       Installa9 > DEC AXPVMS DECNET_OSI V7.2-1        Full LP     Install 9 > DEC AXPVMS DWMOTIF V1.2-5           Full LP     InstallS9 > DEC AXPVMS OPENVMS V7.2-1           Platform    Install 9 > DEC AXPVMS TCPIP V5.0-10            Full LP     Install 9 > DEC AXPVMS VMS V7.2-1               Oper System Install  >2  > $ type decw$server_0_error.log > 7 > This is the DECwindows X11 display server for OpenVMS 5 > AXP V7.1-001212 compiled on Dec 12 2000 at 16:34:31A > ...l > --8 > moc dot slupofni at ssarb - please revert the sequence   ------------------------------    Date: 22 Feb 2002 06:52:46 -0800( From: bob@instantwhip.com (Bob Ceculski)" Subject: Re: Shannon on the merger< Message-ID: <d7791aa1.0202220652.16a43bd@posting.google.com>  u "Bill Todd" <billtodd@metrocast.net> wrote in message news:<sChd8.144851$Aw2.10337398@bin7.nnrp.aus1.giganews.com>...UG > Terry has now weighed in at openvms.org and tru64.org in favor of theiJ > merger, listing the same tired arguments the principals have trotted outH > repeatedly and unconvincingly for 5 months now.  Why he is so securely/ > ensconced in Compaq's pocket remains unknown.e > M > From Compaq's viewpoint, proceding with the merger probably *is* preferableoK > to continuing on its current course with the same morons at the helm, butuJ > from HP's viewpoint the merger makes no sense whatsoever.  What would beL > best for both companies would be to scrap the merger, scrap the managementK > geniuses who thought it up, and get on with their potentially recoverable L > and profitable businesses, but that obviously won't happen without outsideM > influence - and since their boards seem intent on following their CEOs down0C > the path to oblivion, the most likely such outside influence is ajG > stockholder vote-down of the merger, after which the rest will becomeH > possible.i >  > - bill  7 he ought to read the true story of EV8 and be outraged!a   ------------------------------  # Date: Fri, 22 Feb 2002 15:07:42 GMTe4 From: "Terry C. Shannon" <terryshannon@mediaone.net>" Subject: Re: Shannon on the merger9 Message-ID: <29td8.8524$ro5.3746692@typhoon.ne.ipsvc.net>c  5 "Bob Ceculski" <bob@instantwhip.com> wrote in messageC6 news:d7791aa1.0202220652.16a43bd@posting.google.com...7 > "Bill Todd" <billtodd@metrocast.net> wrote in messaget? news:<sChd8.144851$Aw2.10337398@bin7.nnrp.aus1.giganews.com>...*I > > Terry has now weighed in at openvms.org and tru64.org in favor of the L > > merger, listing the same tired arguments the principals have trotted outJ > > repeatedly and unconvincingly for 5 months now.  Why he is so securely1 > > ensconced in Compaq's pocket remains unknown.s > >nD > > From Compaq's viewpoint, proceding with the merger probably *is*
 preferableI > > to continuing on its current course with the same morons at the helm,d buttL > > from HP's viewpoint the merger makes no sense whatsoever.  What would beC > > best for both companies would be to scrap the merger, scrap the.
 managementA > > geniuses who thought it up, and get on with their potentially  recoverablenF > > and profitable businesses, but that obviously won't happen without outsidelJ > > influence - and since their boards seem intent on following their CEOs downE > > the path to oblivion, the most likely such outside influence is a I > > stockholder vote-down of the merger, after which the rest will becomem
 > > possible.c > >f
 > > - bill >n9 > he ought to read the true story of EV8 and be outraged!a  J As previous stated, I *have* read the story. In fact, I have a citation onH same appearing in the next action-packed issue of Shannon Knows Mergers,: Marketing Malfeasance, and Vince Fosterized Architectures.  " Paul Demone is a damn good writer.  K I would be a lot more outraged had financial circumstances not impacted theuJ EV8 decision. I *am* outraged by the financial circumstances. If you wouldE like to see a bit of common sense on the CPQ Board, get the CommunistaL Government (one candidate per position) rules changed at CPQ, and support myJ effort to run for the Board. Remember: a vote for me is a vote for Charlie- Matco. And *not* for Ken (remember him?) Lay!g   ------------------------------  % Date: Fri, 22 Feb 2002 09:24:24 -0000e. From: "Ken Hagan" <K.Hagan@thermoteknix.co.uk> Subject: Re: Singular theyB Message-ID: <1014369791.18636.0.nnrp-12.3e31ffea@news.demon.co.uk>  3 "Alan E. Feldman" <SPAMSINK2001@YAHOO.COM> asked...t >i > Come on, would you say >c& >     Everyone took his coat with him. >t > or would you say >p* >     Everyone took their coats with them.  % "Everyone took their coat with them"?r  < Oh, and just to rattle a few cages, are we all familiar with the use of "themself"?   ------------------------------  % Date: Fri, 22 Feb 2002 11:28:41 +0000 4 From: Bob Cousins <bob@$NOSPAM$lintilla.demon.co.uk> Subject: Re: Singular they8 Message-ID: <doac7ug5r3h5af6puq1lb81j0slvnc0rmp@4ax.com>   Ken Hagan wrote:  4 >"Alan E. Feldman" <SPAMSINK2001@YAHOO.COM> asked... >> >> Come on, would you say6 >>' >>     Everyone took his coat with him.  >> >> or would you sayr >>+ >>     Everyone took their coats with them.  >i& >"Everyone took their coat with them"? >s= >Oh, and just to rattle a few cages, are we all familiar withn >the use of "themself"?. >   1 Jeez, whomever cares! English is a dumb language.c -- , Bob Cousins.   ------------------------------  % Date: Fri, 22 Feb 2002 06:06:51 -0800n# From: "Tom Linden" <tom@kednos.com>  Subject: RE: Singular they9 Message-ID: <CIEJLCMNHNNDLLOOGNJIIEEAEDAA.tom@kednos.com>h   > -----Original Message-----5 > From: Ken Hagan [mailto:K.Hagan@thermoteknix.co.uk]S) > Sent: Friday, February 22, 2002 1:24 AMs > To: Info-VAX@Mvb.Saic.Com5 > Subject: Re: Singular they >  > 5 > "Alan E. Feldman" <SPAMSINK2001@YAHOO.COM> asked...t > >t > > Come on, would you say > > ( > >     Everyone took his coat with him. > >l > > or would you say > >V, > >     Everyone took their coats with them. > ' > "Everyone took their coat with them"?T > > > Oh, and just to rattle a few cages, are we all familiar with > the use of "themself"? >  > $ You mean themselves, don't you ? :-)   ------------------------------  # Date: Fri, 22 Feb 2002 13:32:25 GMTe' From: alexc@world.std.com (Alex Colvin)uU Subject: Re: Singular they was Re: Sv: Younger recruits versus experienced veterans (a& Message-ID: <GrxsA1.1n9@world.std.com>  
 >... snip ...- >> - >> Come on, would you saya >> 2' >>     Everyone took his coat with him.t >> c >> or would you say: >> P+ >>     Everyone took their coats with them."   Damn! no one left me a coat!   -- 2
 	mac the nafd   ------------------------------    Date: 22 Feb 2002 09:10:32 -0600- From: koehler@encompasserve.org (Bob Koehler)iU Subject: Re: Singular they was Re: Sv: Younger recruits versus experienced veterans ( 3 Message-ID: <bxPKeuUzaDV9@eisner.encompasserve.org>   s In article <1014285398.11194.0.nnrp-12.3e31ffea@news.demon.co.uk>, "Ken Hagan" <K.Hagan@thermoteknix.co.uk> writes:G >  > Everyone is singular.  > Everyone are not plural. >   =    Back to reality.  The subject of the first sentence is not         everyone -- all the people      it'sn%       everyone -- the word 'everyone'0      3'    Of course a single word is singular.          Hats is plural.h       Hat is not plural.*       Fish is only singular in New Jersey.   ------------------------------  ! Date: Fri, 22 Feb 02 09:41:48 GMTo From: jmfbahciv@aol.comcY Subject: Re: Singular they was Re: Sv: Younger recruits versus experienced veterans ( was + Message-ID: <a55bmg$dm2$2@bob.news.rcn.net>   9 In article <3c75d402$0$79564$3c090ad1@news.plethora.net>, ,    seebs@plethora.net (Peter Seebach) wrote:* >In article <3C72ACD0.1CDB6F93@cisco.com>,( >J Ahlstrom  <jahlstro@cisco.com> wrote:C >>>I was therefore DELIGHTED a few years back to read a fascinating J >>>article (the citation, alas, escapes me)  that listed examples  of the  use ofF >>>"they" and "their" as third person singular in Thackeray, Dickens, 	 Carlisle,nF >>>Shakespeare, and Shaw--and several other luminaries I've forgotten.E >>>Since then, I've kept an eye out, and most recently discovered therB >>>singular "they"  in C. S. Lewis' *The Lion, the Witch, and the 
 Wardrobe*.D >>>That example was along the lines of "everyone had what THEY likedI >>>best to eat."  If it's good enough for CSL--no mean stylist--it's goodt >>>enough for me.t > 0 >"everyone" is a fairly widespread special case.
 ROTFLMAO.    /BAH  ' Subtract a hundred and four for e-mail.u   ------------------------------  % Date: Fri, 22 Feb 2002 09:58:32 -0000l5 From: "Robert A.M. van Lopik" <lopik@mail.telepac.pt>a Subject: SOAP on OpenVMS?u$ Message-ID: <3c76168d$1@news.wau.nl>  L While looking into the whole subject of Web Services, I noticed the ubiquityL of SOAP implementations. I would also want to have a (more or less complete)D list of SOAP implementations for VMS. So, if anybody knows of a SOAP5 implementation for VMS, could you please let me know.c   Thanks in advancec  
 rob van lopikl   ------------------------------  % Date: Fri, 22 Feb 2002 10:36:42 +0000l) From: Greg Thomas <thomasgd@omc.bt.co.uk>r Subject: Re: SOAP on OpenVMS? 8 Message-ID: <ek7c7uo5l2mdgiqpfam9mosd4arafg89fu@4ax.com>  ; On Fri, 22 Feb 2002 09:58:32 -0000, "Robert A.M. van Lopik"sB <lopik@mail.telepac.pt> wrote in article <3c76168d$1@news.wau.nl>:   >So, if anybody knows of a SOAP-6 >implementation for VMS, could you please let me know.  + http://search.cpan.org/search?dist=SOAP andjC http://xml.apache.org/soap/index.html spring to mind. I've used the!( latter, but not on the OpenVMS platform.   Greg -- .5 This post represents the views of the author and doesi5 not necessarily accurately represent the views of BT.o   ------------------------------   Date: 22 Feb 2002 04:20 CSTg' From: carl@gerg.tamu.edu (Carl Perkins)s@ Subject: Re: Strange status value - when can I expect to get it?- Message-ID: <22FEB200204204200@gerg.tamu.edu>0  1 JF Mezei <jfmezei.spamnot@videotron.ca> writes...o }Bob Koehler wrote:  }> > mc sysgen sh timeprompwaitS- }> > show a great unit : the microfortnight !z }> .$ }>    Goes back to VMS 1.x on VAXen. }  } + }I seem to be living on a different planet.  }  }for me, the above does: } $ }SYSGEN-E-NOPARAM, no such parameter } I }(and for the real parameter, it just professionally displays the value).t }  }  }VAX-VMS 7.2  9 It is, of course "timepromptwait" when spelled correctly.   8 If you look at it, in the "Units" column you'll see that6 it specified "uFortnights", which are microfortnights.8 (As the help for the parameter on recent versions of VMS9 notes, the microfortnights are approximated in the actual : implementation by using seconds instead - a microfortnight7 is 1.2096 seconds so they are short by a bit over 17%).    --- Carl   ------------------------------    Date: 22 Feb 2002 11:16:41 -0600- From: koehler@encompasserve.org (Bob Koehler)c@ Subject: Re: Strange status value - when can I expect to get it?3 Message-ID: <NGP0jxCXdsFY@eisner.encompasserve.org>w  \ In article <3C756007.54987BD9@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > Bob Koehler wrote: >> > mc sysgen sh timeprompwaitc- >> > show a great unit : the microfortnight !  >>  $ >>    Goes back to VMS 1.x on VAXen. >  > , > I seem to be living on a different planet. >   
    VMS 7.2-1:c       $ mc sysgen sh timeprompTwaitO Parameter Name           Current    Default     Min.      Max.     Unit Dynamic O --------------           -------    -------    -------   -------   ---- -------yL TIMEPROMPTWAIT              65535         -1         0         -1 uFortnight   ------------------------------    Date: 22 Feb 2002 11:19:34 -0600- From: koehler@encompasserve.org (Bob Koehler)o@ Subject: Re: Strange status value - when can I expect to get it?3 Message-ID: <6QUSb6VMPI7O@eisner.encompasserve.org>h  \ In article <3C75BECF.67945434@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > I > Are you saying that Compaq intends to productize those easter eggs, and5N > require a special licence ($$$) to access those easter eggs in VMS ?  (would3 > those be considered system integrated products ?)>  >    Compaq can't read calendars.  They laid their egg 2HCY2001.   ------------------------------    Date: 22 Feb 2002 11:25:20 -0600- From: koehler@encompasserve.org (Bob Koehler)t@ Subject: Re: Strange status value - when can I expect to get it?3 Message-ID: <ZwNO0Y6x0emt@eisner.encompasserve.org>a  p In article <3c75ba7e.524144494f47414741@radiogaga.harz.de>, martin@radiogaga.harz.de (Martin Vorlaender) writes: > > > Gross approximation, I'd say: 1 uFortnight = 1.2096 seconds.- > But properly documented. What a great joke!1  H    1 Fortnight is nearly an even multiple of 60Hz line clock frequency.0    I think some PDP-11/70 fellow was having fun.   ------------------------------  % Date: Fri, 22 Feb 2002 11:06:48 +0300V4 From: "Ruslan R. Laishev" <Laishev@SMTP.DeltaTel.RU>/ Subject: Re: SYSPRV & $GETUAI - Dumb question ?C0 Message-ID: <3C75FC18.EC610851@SMTP.DeltaTel.RU>   Hello Richard,( 	thanks for you great examples and help.8 	Now I have a yet another example uwss written on DEC C.  5 	http://starlet.deltatel.ru/~laishev/delta/beaweb/*.*V   Richard Maher wrote: > K > For Pete's sake just do what all of the VMS and Rdb engineers do and dropo  > the /PROTECT linker qualifier! > N > Make sure to protect the psects that need protecting against user mode write4 > access with the protect= linker options qualifier. > N > As I've previously stated. the documentation is ambiguous on this matter butK > the gist (and backed up by Stephen Hoffman) is that it is a huge securityrI > hole. (See my example accessing SYSUAF via RMS for a "secure" example).0 > M > Unfortunately, to date, no one has bothered to explain why. They could tellI% > us but then they'd have to kill us.  >    -- o Cheers, F +OpenVMS [Sys|Net] HardWorker .......................................+E  Russia,Delta Telecom Inc,                    Cel:  +7 (901) 971-3222 E  191119,St.Petersburg,Transportny per. 3                     116-3222-F +http://starlet.deltatel.ru ................. SysMan rides HailStorm +   ------------------------------    Date: 22 Feb 2002 11:03:34 -0600- From: koehler@encompasserve.org (Bob Koehler) - Subject: Re: Using Edit inside a procedure...V3 Message-ID: <aZN4ZK+9hUly@eisner.encompasserve.org>V  W In article <94cd8.2823$Yf1.17988@typhoon.bart.nl>, "Hans Vlems" <hvlems@iae.nl> writes: F > Well the error message explains it all, EDIT/EDT (nor EDIT/TPU) only> > works from a terminal. Try the HELP contents on EDIT/SUMSLP.! > Perhaps that will work for you.X  A    Nonsense.  I've been using EDT from command files for a couple_    decades.   ) > <valdemir-@uol.com.br> wrote in messageA2 > news:200202161306.LAA01603@walters.uol.com.br... >>8 >> How can I edit a file inside a procedure ? Like this: >> >> vax1> type x.com  >>       $ EDIT/READ LOGIN.COM >>       $ EXIT  >>5 >> When I execute this procedure, I get this message:a >>2 >> Change mode can be entered only from a terminal  B    Depends on what you're really trying to do.  You have a startupB    script for EDT that includes the 'change mode' command (usuallyG    just 'c').  If you want to bring up the editor from the command filesA    and let the user edit a file, you need to redirect sys$input. 3C    sys$input is the command file while you're using the '@' commandiB    unless you change it.  sys$command is the terminal all the time    you're logged in:  .       $ define/suer_mode sys$input sys$command       $ edit/read login.como       $ exit  D    If you actually want to script edits, simply put them in a scriptD    as line mode EDT commmands, and tell eDT to use that command file@    instead of the one with the 'c' in it.  In this example, I'llE    leave sys$input pointing at the DCL command file and use it as theB    EDT script:  ,       $edit/read login.com/command=sys$input       $deckr       type whole
       quit
       $eod       $exit    ------------------------------    Date: 22 Feb 2002 07:08:43 -0800( From: bob@instantwhip.com (Bob Ceculski)6 Subject: Re: What is the best NFS package for OpenVMS?= Message-ID: <d7791aa1.0202220708.5a06a479@posting.google.com>C  e Mark Daniel <mark.daniel@wasd.vsm.com.au> wrote in message news:<3C6EA55F.9EBF516@wasd.vsm.com.au>...  > Bob Ceculski wrote:-F > > I have been on vms for over 16 years now "without" an os crash ...H > > We have been on purveyor for two years now w/o a crash or lockup ...B > > We have been on TCPware for over 2 years now w/o a problem ...I > > These are called testimonials ... do you have any, including your own2I > > that suggests that wasd is reliable ... thats all I am asking for ...rJ > > My point is, you have a job, and wasd is not your primary job, it is aI > > side hobby, and it is great that you are trying to help others w/free K > > shareware, but software that you pay for is generally better (i.e. vms)BL > > than shareware (i.e. linux) because you are paying for it (the exceptionH > > to this is microsoft) ... wasd may be an exception, and testimonials3 > > would answer any questions ... do you have any?t > J > Now it has been argued that V7.3 should have been designated V8.0 and weF > may actually have yet another example of why we are inclined to step? > over dot-zero releases of anything but within a discussion ofp0 > "production quality" this is most significant. > C > Now perhaps we might all live quieter lives if we stuck to an n-2 @ > version of whatever, where we live with less features and moreG > stability, but for many that is not an option.  Also, despite all thecJ > Alpha, Beta and Gamma (dot-zero) testing, until something becomes widelyE > deployed the nastier, latent gremlins often will not surface.  It'soH > dirty work but then someone, somewhere, sometime will have to do it by > necessity. >   H which may explain why I have never had a crash is when version 7.3 comesE out we will move to 7.2 if necessary, otherwise we do not live on theiG cutting edge ... we tried that w/apache and got cut!  Couldn't you have I waited a little longer for 7.3 as it is not a significant change from 7.2  as far as features?e   ------------------------------    Date: 22 Feb 2002 18:24:28 -0000= From: Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]>o6 Subject: Re: What is the best NFS package for OpenVMS?6 Message-ID: <20020222182428.28699.qmail@gacracker.org>  9 On 22 Feb 2002, bob@instantwhip.com (Bob Ceculski) wrote:o; >Mark Daniel <mark.daniel@wasd.vsm.com.au> wrote in message1+ >news:<3C6EA55F.9EBF516@wasd.vsm.com.au>...  >> Bob Ceculski wrote:G >> > I have been on vms for over 16 years now "without" an os crash ...eI >> > We have been on purveyor for two years now w/o a crash or lockup ...sC >> > We have been on TCPware for over 2 years now w/o a problem ...sJ >> > These are called testimonials ... do you have any, including your ownJ >> > that suggests that wasd is reliable ... thats all I am asking for ...K >> > My point is, you have a job, and wasd is not your primary job, it is asJ >> > side hobby, and it is great that you are trying to help others w/freeL >> > shareware, but software that you pay for is generally better (i.e. vms)M >> > than shareware (i.e. linux) because you are paying for it (the exception I >> > to this is microsoft) ... wasd may be an exception, and testimonialso5 >> > would answer any questions ... do you have any?   >> tK >> Now it has been argued that V7.3 should have been designated V8.0 and weeG >> may actually have yet another example of why we are inclined to step @ >> over dot-zero releases of anything but within a discussion of1 >> "production quality" this is most significant.h >> rD >> Now perhaps we might all live quieter lives if we stuck to an n-2A >> version of whatever, where we live with less features and morenH >> stability, but for many that is not an option.  Also, despite all theK >> Alpha, Beta and Gamma (dot-zero) testing, until something becomes widely F >> deployed the nastier, latent gremlins often will not surface.  It'sI >> dirty work but then someone, somewhere, sometime will have to do it bye >> necessity.  >>   > I >which may explain why I have never had a crash is when version 7.3 comes.F >out we will move to 7.2 if necessary, otherwise we do not live on theH >cutting edge ... we tried that w/apache and got cut!  Couldn't you haveJ >waited a little longer for 7.3 as it is not a significant change from 7.2 >as far as features?  K The biggest problem is what if everybody adopted that approach? There would J be no serious real-world testing of the new releases and bugs would remainK uncovered thus propagating to the next-released version. Sure, bigger sites H can run their test systems on later versions of the operating system butI this will, regardless of OS origins, never find all the bugs that running I at full tilt on production will. Sooner or later somebody has to take the0 risk.8     Doc. -- m6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                              http://vmsbox.cjb.net>   ------------------------------    Date: 22 Feb 2002 05:33:48 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) & Subject: Re: World Class Sportfishing!3 Message-ID: <YbV3tU6yUN+e@eisner.encompasserve.org>r  p In article <b4ld8.7669$ro5.3537103@typhoon.ne.ipsvc.net>, "Terry C. Shannon" <terryshannon@mediaone.net> writes:L > Oh for cripe's sake, Larry, if this is the biggest thing you have to bitch  E I am not here to complain, I am here to learn.  But random characters-! through the text gets in the way.   H > about, then just killfile me. Or bitch at my ISP. Or Mike Capellas. Or > Elvis.  D Sorry, but my killfile is reserved for those who never have anythingC worthwhile to say.  It takes more than a purchasing error to earn at place there :-)y   Larry Kilgallen   N > FYI my mailer didn't say anything about minme until after the fact, not that8 > an excuse is worth a tinkers damn in this environment. >  > < > "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message/ > news:6bdenkKkgV0u@eisner.encompasserve.org...oF >> In article <Ctfd8.7542$ro5.3284809@typhoon.ne.ipsvc.net>, "Terry C.. > Shannon" <terryshannon@mediaone.net> writes:1 >> > This is a multi-part message in MIME format.  >> >0 >> > ------=_NextPart_000_0063_01C1BB07.2E3BCA00 >> > Content-Type: text/plain; >> > charset="iso-8859-1"r0 >> > Content-Transfer-Encoding: quoted-printable >>) >> Please do not post MIME in newsgroups.eM >> Contact your software vendor if the method for avoiding MIME is not clear.i   ------------------------------   End of INFO-VAX 2002.104 ************************