1 INFO-VAX	Sun, 06 Jul 2003	Volume 2003 : Issue 369       Contents: Re: "VMS for Developers"- "VMS for Developers" (was: Are HP listening?) 1 Re: "VMS for Developers" (was: Are HP listening?) 1 Re: "VMS for Developers" (was: Are HP listening?) % Alpha nr. 2 on list of supercomputers  Re: Are HP listening? ) more on DSSI disks and allocation classes - Re: more on DSSI disks and allocation classes - Re: more on DSSI disks and allocation classes ( Re: Q on TS10 emulator: only TEXT BASED? Re: Running VMS off CD1 Re: Standard Digital Markup Language (SDML) DTD ? ! RE: Upcoming release of VMS 7.3-2 ! Re: Upcoming release of VMS 7.3-2   F ----------------------------------------------------------------------  % Date: Sat, 05 Jul 2003 18:42:42 -0500 ( From: Michael Rice <marice@whiteice.com>! Subject: Re: "VMS for Developers" / Message-ID: <vgeojjdd81vg73@corp.supernews.com>   # On 7/5/2003 6:14 PM, rob kas wrote: C >>Furthermore, bundling compilers severly cramps the style of third # >>parties trying to sell compilers.  >>F >>Why is it that you want to drive third party vendors away from VMS ? > M >   Other then the Awful GNU C Compiler what other C Compilers ara available?  >   F What is the most recent version of the GNU compiler available for VMS H (VAX or Alpha)?  I saw a reference to GCC 2.95.3, but couldn't find any 	 binaries.    ------------------------------   Date: 5 Jul 2003 14:13:35 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) 6 Subject: "VMS for Developers" (was: Are HP listening?)3 Message-ID: <Zci3mxi3Mvsm@eisner.encompasserve.org>   P In article <ov1egv4rifn1ahguk5otplu7a6s44j2o38@4ax.com>, Dave Brennan <> writes:  B > I agree with the analysis. I would however bundle as many of the/ > layered products as possible into the system.  > F > This would produce packages such as "VMS for developers" which would, > have all the development products bundled.  B For commercial product developers, that would be the ASAP/CSA/DSPP program.  H > The development costs of the layered products have been recoverd yearsH > ago. Why keep the DEC model for selling layered products, who buys a C* > compiler at the price HP sell it on VMS.  D Bundling a C compiler doesn't do much for those of us who develop inC other language.  Not pricing things separately hides information as 4 to how popular various language processors might be.  A Furthermore, bundling compilers severly cramps the style of third ! parties trying to sell compilers.   D Why is it that you want to drive third party vendors away from VMS ?   ------------------------------  $ Date: Sat, 5 Jul 2003 19:14:07 -0400# From: "rob kas" <rob@paychoice.com> : Subject: Re: "VMS for Developers" (was: Are HP listening?)8 Message-ID: <ajINa.346$jB5.46252169@news.netcarrier.net>  C > Furthermore, bundling compilers severly cramps the style of third # > parties trying to sell compilers.  > F > Why is it that you want to drive third party vendors away from VMS ?    K   Other then the Awful GNU C Compiler what other C Compilers ara available?   8                                                   Thanks7                                                     Rob    ------------------------------   Date: 5 Jul 2003 21:26:16 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) : Subject: Re: "VMS for Developers" (was: Are HP listening?)3 Message-ID: <ODAWpodzPkoc@eisner.encompasserve.org>   ^ In article <ajINa.346$jB5.46252169@news.netcarrier.net>, "rob kas" <rob@paychoice.com> writes: > D >> Furthermore, bundling compilers severly cramps the style of third$ >> parties trying to sell compilers. >>G >> Why is it that you want to drive third party vendors away from VMS ?  >  > M >   Other then the Awful GNU C Compiler what other C Compilers ara available?   E I have no particular interest in C, but Dibol and PL/I compilers have  been mentioned recently.   ------------------------------  % Date: Sat, 05 Jul 2003 22:57:48 +0200  From: Dirk Munk <munk@home.nl>. Subject: Alpha nr. 2 on list of supercomputers2 Message-ID: <be7ejg$uke$1@news3.tilbu1.nb.home.nl>  K An Alpha SC system with 8192 CPU's on 1.25 GHz is the nr. 2 on the list of  Q supercomputers. It consists out of ES45 building blocks according to the article  P in CT. The performance of the system is 13880 Teraflops, the performance of nr. I 1 (a NEC system) is 35860 Teraflops. The Alpha belongs to the Los Alamos   National Laboratory.  C There is also a Superdome 875 on the list, it is ranked as nr. 172.    ------------------------------  % Date: Sat, 05 Jul 2003 18:42:05 +0100  From: Dave Brennan <>  Subject: Re: Are HP listening?8 Message-ID: <ov1egv4rifn1ahguk5otplu7a6s44j2o38@4ax.com>  5 Doc.Cypher <Use-Author-Address-Header@[127.1]> wrote:  > . >Here's my correspondent's list of suggestions > > >1.  Get a major application such as SAP running on VMS again. > 8 >2.  Get Oracle to treat VMS with a little more respect. > ' >3.  Market outside the installed base.  > ) >4.  Invest in the Educational Programme.  > L >5.  Reintroduce lower-cost hardware and offer a low-cost turnkey web server >    platform. >   @ I agree with the analysis. I would however bundle as many of the- layered products as possible into the system.   D This would produce packages such as "VMS for developers" which would* have all the development products bundled.  0 "VMS for Enterprise" a cheap VMS cluster option.  F What I am saying is that given that most users are migrating from VMS,' why not make it easy to migrate to VMS.   F The development costs of the layered products have been recoverd yearsF ago. Why keep the DEC model for selling layered products, who buys a C( compiler at the price HP sell it on VMS.  E The real problem is that VMS is still run by people who live in a DEC B world. The time has come to move out of the 1970's and think about# packaging VMS for the next century.    ------------------------------  * Date: Sat, 5 Jul 2003 20:25:14 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)2 Subject: more on DSSI disks and allocation classes$ Message-ID: <be7c79$vpi$1@online.de>  G A while back, I posted some questions here regarding how to change the  E allocation class on a DSSI bus connected to a VAX I inherited.  (I'm  E still not sure whether it's completely necessary, but I figured it's  ? best to change it to make sure there is no conflict with other  C allocation classes in the cluster I intend to introduce it to.)  I  G recently installed VMS 7.3 on the DSSI disk.  I don't plan to use this  C as the system disk in the long term, but will use an external SCSI  E shadow set.  First, I just wanted to make sure the VAX is OK, adjust  4 system parameters etc and try it out in the cluster.  D Going over the responses including the VMS FAQ, I finally tried this out:    9    $ set host /dup /server=mscp$dup/task=params node_name   # which got me to the PARAMS> prompt.   H I then did SET ALLCLASS and SET NODENAME and entered WRITE.  I was told I this needed an initialisation and was asked if I wanted to do it now.  I  	 said yes.   * This has resulted in an unbootable system!  B When I got back to the DCL prompt, a message appeared about an ID F conflict or something similar.  I was logged in to the console and it  froze.  E Even power-cycling the system results in "bootstrap failure" when it   tries to boot.    I The problem is, I can't change anything from the console (with regard to  G the DSSI bus or to the VAX itself (for example, whether it should boot  H on power-up and from which disk)) since I can't even get to the console  prompt.   H Any help would be much appreciated, as soon as possible since I'll only H be in the neighbourhood of this VAX for another 28 hours or so, and not B again for another couple of weeks, but it would be nice to get it $ running again within those 28 hours.  ? Replies by email are encouraged for speed; I'll post a summary.    ------------------------------  * Date: Sat, 5 Jul 2003 21:03:18 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)6 Subject: Re: more on DSSI disks and allocation classes$ Message-ID: <be7eem$1me$1@online.de>  D In article <be7c79$vpi$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    ; >    $ set host /dup /server=mscp$dup/task=params node_name   , > This has resulted in an unbootable system!  I I find it really hard to believe that I can execute a command at the DCL  I prompt which has the result that after power-cycling the system, I can't   even get to the console prompt!    ------------------------------  * Date: Sat, 5 Jul 2003 22:05:23 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)6 Subject: Re: more on DSSI disks and allocation classes$ Message-ID: <be7i33$4ei$1@online.de>  D In article <be7eem$1me$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    = > >    $ set host /dup /server=mscp$dup/task=params node_name  > . > > This has resulted in an unbootable system! > K > I find it really hard to believe that I can execute a command at the DCL  K > prompt which has the result that after power-cycling the system, I can't  ! > even get to the console prompt!   B It seems I spoke too soon.  When I tried power-cycling, I left the> system off for about 10 seconds or so.  This has always seemedE sufficient.  As I said, I couldn't even get to the console prompt.  I E then left the power off for a much longer time, an hour or so.  After F that, the system came up normally and I DID get to the console prompt.H After connecting to the DSSI controller from the console, I saw that the? changes I made had taken effect.  After that, the system booted 
 normally.   G Thus, it seems that the conflicting parameters causing the problem got  0 cached somewhere, even though the power was off!  E Of course, it's often possible to do something and lock oneself out,  I such as connecting to something via TCPIP then changing the IP address.   I However, a power cycle is usually sufficient to get one into a situation  H where one can at least connect again somehow.  (In this case, it was as , well, but the power had to stay off longer.)   ------------------------------  # Date: Sat, 05 Jul 2003 20:53:49 GMT ) From: Andrew Balaam <abalaam@yahoo.co.uk> 1 Subject: Re: Q on TS10 emulator: only TEXT BASED? 4 Message-ID: <20030705.20534900.248152911@imagnu.geo>  I I have run X applications from a TS10 emulated VAX on my Linux box - the=  =20 ' performance is not great, but it works.   6 >>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<  I On 04/07/03, 12:57:05, winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston -=20 I SSRL Admin Cmptg Mgr") wrote regarding Re: Q on TS10 emulator: only TEXT=  =20  BASED?:     B > In article <adeef81b.0307040348.7a52f998@posting.google.com>,=20. shevchenko@ikp.tu-darmstadt.de (Sheva) writes:I > >> The TS10 vax emulator uses a virtual Ethernet solution (TUN/TAP) th=  atI > >> allows the host box to hold network conversations with the emulated=   VAX > > ( > >In the description for TS10 they say:$ > >Environment: Console (Text Based) > > B > >Does that mean that I cannot use X-WINDOWS applications on thatF > >emulated VAX, but only those programs which give the text output to > >the console?  > > D > >Or had someone managed to use TS10-emulated VAX with DECWINDOWS ?  I > I haven't used TS10 at all, but I'm aware that you can make a VAX or=20  Alpha G > that doesn't have a graphics card at all run DECwindows software with I > displays somewhere else.  If the emulated VAX thinks it has an Etherne=  t I > card, it should be able to pop open DECterms anywhere else you want to=    > allow it to.  	 > -- Alan  > -- >=20I =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= I =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= I =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=  =3D=3D=3D=3D=3D=3D=3D 2 >  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUE >  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone: =20  650/926-3056I >  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA =   =20 94025  >=20I =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= I =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= I =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=  =3D=3D=3D=3D=3D=3D=3D    ------------------------------  % Date: Sat, 05 Jul 2003 22:59:23 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: Running VMS off CD ' Message-ID: <3F079E9B.933DE84F@fsi.net>    "Doc.Cypher" wrote:  > I > On Fri, 04 Jul 2003, "David J. Dachtera" <djesys.nospam@fsi.net> wrote:  > >"Doc.Cypher" wrote: > >> [snip] L > >> On a slightly related note, does anyone know of a package (available toP > >> Hobbyists) which gives similar functionality to the *ix package "tripwire"? > > A > >Rather depends on what "tripwire" is and what it does. Care to 
 > >elucidate?  > / > I came up with half of the solution myself...  > I > $ SET SECURITY/ACL=((ALARM=SECURITY,ACCESS=WR+DE+CO+SUCCESS+FAILURE), - N >                     (AUDIT=SECURITY,ACCESS=WR+DE+CO+SUCCESS+FAILURE)  <file>  ! I'd add that to my VMS wish list:   & $ DEFINE/SYSTEM/EXEC SECURITY_NOTIFY -* 	SYS$MANAGER:NOTIFY_BY_EMAIL_AND_PAGER.COM+ $! KERNEL MODE LOGICAL WOULD BE MOST SECURE  $ SET SECURITY/ACL=(-  	(ALARM=SECURITY,-# 		ACCESS=WR+DE+CO+SUCCESS+FAILURE,-  		ACTION=@SECURITY_NOTIFY), - < 	(AUDIT=SECURITY,ACCESS=WR+DE+CO+SUCCESS+FAILURE))  <object>  H ...or perhaps SET AUDIT/ACTION would be more appropriate. Dunno, haven'tH thought it through much yet. In any case, the process environment withinG which the action executes should provide environment symbols from which F information can be retrieved such as the identity of the user trippingD the ALARM or AUDIT, the identity and class of the target object, the; action attempted, the result (SUCCESS or FAILURE), etc. ...   H As for SET AUDIT/LISTENER, the listener process should probably run in a proc. similar to:   
 $ set noon	 $restart:  $ run <listener> $ wait 00:00:01  $ goto restart  D The one(1) second wait allows a chance to kill it if it "runs away".   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Sat, 05 Jul 2003 19:50:20 GMT ( From: Alder <PGDEHMKOKIMD@spammotel.com>: Subject: Re: Standard Digital Markup Language (SDML) DTD ?; Message-ID: <0_FNa.36776$ZS3.3836761@news2.telusplanet.net>    Hoff Hoffman wrote:   Z > In article <3f03583e$1@obsidian.gov.bc.ca>, "Alder" <MUNDDGNTDYTV@spammotel.com> writes:M > :The DTD, or Document Type Definition, is merely a formal definition of the - > :tags and attributes of a markup language.   > K >   I am not aware of any publicly available DTD for DECdocument.  Manuals, G >   certainly, but DIGITAL never offered a DTD for the package.  (There I >   were some converters around that skirted this area and may well have  & >   had internal DTD-like structures.) > I >   Being at least partitially TeX based, you might also be able to prise K >   at least some certain related information out of the product TeX files.  > I >   You might want to discuss this with Touch Technologies, the owners of I >   DECdocument -- I do not know if the Touch folks follow the newsgroups D >   -- as Touch would know what is presently available in this area: > 0 >     http://www.ttinet.com/tti/decdocument.html >   
 Thanks, Hoff.   ? OK, so this little project of mine is not going to be easy. :-)   F What I am after is a more-or-less comprehensive list of SDML tags and I attributes, preferably in an ASCII text file.  Using this file, I should  A be able to build a simple plugin for the well-known HTML editor,  L HTML-Kit, allowing authors to write SDML with colorized tags and attributes.  D I don't have access to a DECdocument installation to review the TeX = files as you suggest, so I'm asking c.o.v. readers to either:   B a) send me SDML source documents which I can parse into a tag listC b) send me tag lists from other sources, like the DECdocument docs.   G Better yet, if there is already a Windows-based editor out there which  F can colorize SDML tags and attributes, or otherwise make it easier to 5 build SDML source files, I'd love to hear about them.    Cheers,    Alder    ------------------------------  $ Date: Sat, 5 Jul 2003 16:16:40 -04000 From: "Andrew Robert" <shardservant@hotmail.com>* Subject: RE: Upcoming release of VMS 7.3-2/ Message-ID: <vgech4qrk0trba@corp.supernews.com>    Good afternoon everyone,  H I recently upgraded my production cluster to VMS 7.3-1 but then saw that$ VMS7.3-2 is due to come out shortly.  K Does anyone know of any compelling performance reasons to upgrade to the -2  release?  L I read the brief release features as listed on http://www.openvms.compaq.comH and the only things I can see that might be interesting are the fastpath options and working set lists.  K Does anyone know what kind of performance enhancement I can expect in going  to VMS 7.3-2 with these?  K When we moved to the improved XFC file caching under VMS 7.3-1, we improved  speedy buy approximately 20%.   K My current system consists of an Alpha GS-140 with 8G of memory and six EV6  processors.   < Any insight anyone can provide would be greatly appreciated.  
 Thank you,
 Andrew Robert  Principal Systems Analyst $ Operations Systems Support - OpenVMS  Massachusetts Financial Services E-mail: arobert@mfs.com    ------------------------------  $ Date: Sat, 5 Jul 2003 19:35:34 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> * Subject: Re: Upcoming release of VMS 7.3-2: Message-ID: <7hJNa.8129$eF3.1139972@news20.bellglobal.com>  ; "Andrew Robert" <shardservant@hotmail.com> wrote in message ) news:vgech4qrk0trba@corp.supernews.com...  > Good afternoon everyone, > J > I recently upgraded my production cluster to VMS 7.3-1 but then saw that& > VMS7.3-2 is due to come out shortly. > M > Does anyone know of any compelling performance reasons to upgrade to the -2 
 > release? > N > I read the brief release features as listed on http://www.openvms.compaq.comJ > and the only things I can see that might be interesting are the fastpath  > options and working set lists. > M > Does anyone know what kind of performance enhancement I can expect in going  > to VMS 7.3-2 with these? > M > When we moved to the improved XFC file caching under VMS 7.3-1, we improved  > speedy buy approximately 20%.  > M > My current system consists of an Alpha GS-140 with 8G of memory and six EV6 
 > processors.  > > > Any insight anyone can provide would be greatly appreciated. >  > Thank you, > Andrew Robert  > Principal Systems Analyst & > Operations Systems Support - OpenVMS" > Massachusetts Financial Services > E-mail: arobert@mfs.com  >   L I recently attended an Open VMS Technical Seminar in Ottawa and the HP folksK convinced me that an upgrade would be desirable. See my previous post under K the subject "OpenVMS Technical Seminar Highlights (some)" for a list of the . details. BTW, this release is due in OCT-2003.  M One thing I didn't put in my previous post was the return of $ANA/ERROR which M currently requires the installation of some optional layered software. People 5 who used this with VAX/VMS really missed it on Alpha.    <PERSONAL OPINION>L I noticed a big jump in performance just going from 7.3 to 7.3-1 and anotherJ performance increase is expected going to 7.3-2 (mostly related to variousN spin locks). I believe that other companies would do the "marketing thing" andM release 7.3-1 as 7.4 and 7.3-2 as 7.5 so don't ignore it just because this is  a dash release.  </PERSONAL OPINION>     
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------   End of INFO-VAX 2003.369 ************************