1 INFO-VAX	Sat, 09 Jun 2001	Volume 2001 : Issue 318       Contents:, Re: "Talk Baq" to Compaq - or perhaps to Sun, Re: "Talk Baq" to Compaq - or perhaps to Sun+ Re: %FOR-F-NOMSG, where's the real message?  ACMS & DECtrace ) Re: Astronomy programs for Alpha / VMS???  Re: Bloated software Re: Bloated software8 Re: Global warming? was: (OT) Current Microsoft v DOJ st( Re: How can i set a DLT tapedrive online( Re: How can i set a DLT tapedrive online( Re: How can i set a DLT tapedrive online Re: I demand your respect  Re: I demand your respect  NCP counter question" Re: NCP to NCL command equivilants" Re: NCP to NCL command equivilants Need help with LIB$TABLE_PARSE" Re: Need help with LIB$TABLE_PARSE" Re: Need help with LIB$TABLE_PARSE Re: NYSE# Re: Portability. The myth continues . Re: SKC Writeup on VMS Diamond and Tech Forums. Re: SKC Writeup on VMS Diamond and Tech Forums Re: The future of VMS  Re: The future of VMS  Re: The future of VMS  Re: The future of VMS  Re: Threads.- Re: Urgent: Oracle 8 parallel server on VMS 7 - Re: Urgent: Oracle 8 parallel server on VMS 7 ) Re: [OT] Current Microsoft v DOJ status ?   F ----------------------------------------------------------------------    Date: 09 Jun 2001 14:01:24 +0800, From: Paul Repacholi <prep@prep.synonet.com>5 Subject: Re: "Talk Baq" to Compaq - or perhaps to Sun - Message-ID: <87ae3itcnf.fsf@prep.synonet.com>   / JF Mezei <jfmezei.spamnot@videotron.ca> writes:    > "Terry C. Shannon" wrote:   E > > As for McNealy, he was damned conspicuous in his absence from the @ > > earnings warning of Thursday last. He musta been out selling% > > another thousand StarFires, eh...   E > You can find all the faults in Sun, but at least Sun isn't actively E > preventing its best products from being marketed and has no problem E > pitching is own products against Microsoft's.  Compaq's handling of C > VMS is not *that* significantly different from Palmer's. They may C > have stopped actively trying to kill VMS, but it is still told to D > stay put in the small market niches where it can't hurt Microsoft.  A Yeah, I still think the 100 with a billypen card is the best idea  they have had in along time.  H Q vs Compaq? What have the Q done? Where are the new, non DEC originatedH stuff? EV8 is about it, and it is a long way away still. I can not think9 of one forward inovative initive since Mrs Palmer left...    If you know of one, speak now.   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.H Spam-To: uce@ftc.gov,enforcement@sec.gov,sness@fcc.gov,hfurchtg@fcc.gov,#   mpowell@fcc.gov,gtristan@fcc.gov     ------------------------------  # Date: Sat, 09 Jun 2001 15:00:50 GMT 4 From: "Terry C. Shannon" <terryshannon@mediaone.net>5 Subject: Re: "Talk Baq" to Compaq - or perhaps to Sun : Message-ID: <CSqU6.3272$Tc.648974@typhoon.ne.mediaone.net>  9 "Paul Repacholi" <prep@prep.synonet.com> wrote in message ' news:87ae3itcnf.fsf@prep.synonet.com... 1 > JF Mezei <jfmezei.spamnot@videotron.ca> writes:  >  > > "Terry C. Shannon" wrote:  > G > > > As for McNealy, he was damned conspicuous in his absence from the B > > > earnings warning of Thursday last. He musta been out selling' > > > another thousand StarFires, eh...  > G > > You can find all the faults in Sun, but at least Sun isn't actively G > > preventing its best products from being marketed and has no problem G > > pitching is own products against Microsoft's.  Compaq's handling of E > > VMS is not *that* significantly different from Palmer's. They may E > > have stopped actively trying to kill VMS, but it is still told to F > > stay put in the small market niches where it can't hurt Microsoft. > C > Yeah, I still think the 100 with a billypen card is the best idea  > they have had in along time. > J > Q vs Compaq? What have the Q done? Where are the new, non DEC originatedJ > stuff? EV8 is about it, and it is a long way away still. I can not think; > of one forward inovative initive since Mrs Palmer left...  >   > If you know of one, speak now.  J ProFusion, ServerNet II/III, VersaStor,QuickBlade, ICE, iPAQ, Grommit, EV9J (now in advanced development, finally), Piranha, Marvel come to mind. PokeJ around www.compaq.com/research; there's a lot of intriguing stuff there as well.   K Compaq's marketing problem isn't that they're not marketing, it's that they L could be marketing a lot smarter. They could learn a lot from Sun and IBM in this regard.  F And if you're of the opinion that Compaq's marketing and direction are" wrong, by all means tell 'em so at  ! http://www.compaqworkinggroup.org   H Capellas and Elias and Marcello, et al don't read newsgroups and I doubtJ they pay much attention to fan mail over the Internet. They will, however,5 see the results of the survey being conducted online.    ------------------------------   Date: 9 Jun 2001 10:12 CDT' From: carl@gerg.tamu.edu (Carl Perkins) 4 Subject: Re: %FOR-F-NOMSG, where's the real message?, Message-ID: <9JUN200110125299@gerg.tamu.edu>  0 Steve Lionel <Steve.Lionel@compaq.com> writes...B }On 8 Jun 2001 10:11 CDT, carl@gerg.tamu.edu (Carl Perkins) wrote: }>Danger, Will Robinson! }>K }>Something that it doesn't tell you until *after* you have installed it is I }>that if you are running VMS V7.1 or 7.2 (including -x) you will need to M }>reboot if you install this and have never installed Fortran v 7.3 or later:  }>K }>     If you did not previously install Compaq Fortran version 7.3, or are O }>     installing under OpenVMS 7.1 or 7.2 (including any "-x" updates), please N }>     reboot this system and any other Alpha systems in this OpenVMS cluster.L }>     Failure to do so may cause image activation and other errors due to aJ }>     problem in the OpenVMS operating system involving replacement of anK }>     image with a shareable address section (DPML$SHR.EXE).  This problem H }>     is corrected in OpenVMS 7.3.  (If you previously installed CompaqO }>     Fortran 7.3 or later, the new version of DPML$SHR.EXE is already present B }>     and was not replaced, therefore a reboot is not necessary.) }>E }>In many cases this would be a bad thing to have to do unexpectedly.  } B }Sorry about that - you're right.  Our installation guide does say6 }this, but folks reading this wouldn't have seen that. } . }Steve Lionel (mailto:Steve.Lionel@compaq.com)  A I also suggest that you mention it in the release notes that come ? with the thing, say in section 1.1 "Overview and Installation". B (Not that I actually looked before I leaped, but if I had it would not have helped.)    --- Carl   ------------------------------   Date: 9 Jun 2001 04:54:02 -0700 + From: robert_jm_barron@hotmail.com (Robert)  Subject: ACMS & DECtrace= Message-ID: <15945534.0106090354.6944b68e@posting.google.com>    Hi,   @ Does anyone have any experience with using DECTrace to tune ACMS
 applications?   F I use ACMS with servers which we wrote & I'm trying to get DECtrace to( give me information on specific servers.   Thanks in advance,   Robert.    P.S.C I seem to be having newgroup problems, so please send a copy to my   email.   ------------------------------   Date: 9 Jun 2001 15:39:12 +0100 O From: pmoreau@dev.ath.cena.fr (Patrick MOREAU, CENA Athis, Tel: 01.69.57.64.40) 2 Subject: Re: Astronomy programs for Alpha / VMS???  Message-ID: <AYdj8G05K8O7@sable>  4 In article <8wPT6.1168$fi2.29968@news.cpqcorp.net>, 4 hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) writes: > ..& > : http://decwarch.free.fr/astro.html > ..   9 >   The links from there are invalid and/or inaccessable. . >   The WKU stuff moved to process.com, AFAIK.  K Yes, I know, I need to update some pages on the archive, there is no space  1 actually at process.com to hold the DECW archive.   . >   The CNAM server directories are protected.  9 I've just tested and its OK. May be a transient problem ?   M I plan to put more kits on free.fr (actually only kits made since 01-JAN-2000 N are there), but I need to open more accounts (1 account = 100 Mbytes) 'cause I! have near 500 Mb of archive data/    Patrick  --O =============================================================================== O pmoreau@cena.dgac.fr  (CENA)     ______      ___   _           (Patrick MOREAU) 4 moreau_p@decus.fr (DECUS)       / /   /     / /|  /|J CENA/Athis-Mons FRANCE         / /___/     / / | / |   __   __   __   __  N BP 205                        / /         / /  |/  |  |  | |__| |__  |__| |  |N 94542 ORLY AEROGARE CEDEX    / /   ::    / /       |  |__| | \  |__  |  | |__|N http://www.ath.cena.fr/~pmoreau/            http://www.multimania.com/pmoreau/O ===============================================================================    ------------------------------  % Date: Sat, 09 Jun 2001 10:22:06 +0200 ) From: Christof Brass <brass@infopuls.com>  Subject: Re: Bloated software , Message-ID: <3B21DCAE.EFBE3F07@infopuls.com>   JF Mezei wrote:  > = > I have always worked on systems that were not well endowed.    [SNIP]  I > Now, often, one hears about bloated microsoft or windows software. I am E > curious whether the bloating on personal computers is the result of D > programmers not having the concept of limited resources and codingP > inefficiently, or whether the OS lacks tools that would allow them to use lessM > memory resources, or whether the applications truly absolutely require that N > much memory even when coded efficiently due to marketing requirements for so > many functions.   @ Although it is said that a typical Windoze program spends 80% of> its time in WAPI calls there is no real reason to blame the OS@ or lack of tools for not handling resources properly within your= program. Especially the Windoze personally concept allows the = skilful programmer to be as efficient as possible by directly ) using Cutler's kernel. Rembember Interix?    [SNIP]  K > If the folks who designed ALLIN1 had made a X-windows port to support GUI K > stuff, would they have developped something that would have been far less P > memory/CPU demanding than MS products, or would the end result have been quite > similar ?   ? Depends how it would have been done. The problem with Micro$hit 9 SW is that it mostly developed using the completely wrong = techniques like assembly language and C and a little bit C++, 7 that they hire the cheapest programmers and don't offer = acceptable working conditions (most programmers work the same @ amount of overtime they work at normal hours) and that they have> a very high personelle turnaround, i.e. the average programmer> leaves the company after two years or less IIRC, and all their? programs suffer from featureitis, i.e. a lot of features 90% of = the users would never think about using it even if they know.   @ And don't forget that every Micro$hit programs contains valuable< add-ons like flight simulator (Excel) and other very helpful> games (helpful in the meaning that you don't have to cope with> the program's erroneously so called main task). Companies love' that kind of SW and so do employees ;-)    ------------------------------   Date: 9 Jun 2001 10:18:35 -0700 " From: cstranslations@msn.com (Joe) Subject: Re: Bloated software = Message-ID: <d56d1c2d.0106090918.76984bdd@posting.google.com>   a JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message news:<3B21217F.99241825@videotron.ca>... = > I have always worked on systems that were not well endowed.  >    <snip>  I > Now, often, one hears about bloated microsoft or windows software. I am E > curious whether the bloating on personal computers is the result of D > programmers not having the concept of limited resources and codingP > inefficiently, or whether the OS lacks tools that would allow them to use lessM > memory resources, or whether the applications truly absolutely require that N > much memory even when coded efficiently due to marketing requirements for so > many functions.   E Some of it comes from not knowing how to write code. I could show you D some of the source code for stuff we have running at work that wouldC make your skin crawl. We have whole libraries of routines that look @ like they were concieved of and coded by someone that got a D inE programming 101. I get stuck in meetings and discussions on a regular C basis where the focus is "what can we do to improve performance?" I > keep saying, "change the way you write code." Four weeks ago ID happened to be in the right place at the right time when someone wasB complaining about the length of time it took to run a certain dataF upload. I made two suggestions. With those two suggestions rather thanF the whole prosess taking 2.5 days it now takes 5 minutes. I'll save of@ another time the description of all the huffing and puffing thatC occured because I stepped on the toes of the person that originally E designed the particular process. We have two month end jobs that will F run for 5 or 6 days, use and hour or two of CPU time, and generate 10+C billion I/Os. Everyone wonders why the 4100 is consistently sucking ( wind. I get tired of arguing about it...   As far as Micro$oft bloat...  C Poor design. That's not all however. What is Windows suppose to be? 8 Stop and think about it be for you answer. What was DOS?  D DOS was Disk Operating System. Windows is suppose to be an operating@ system. An operating system controls the hardware. Nothing more.D Nothing less. A web browser integrated into the OS has nothing to do? with controlling the hardware. The instant messaging which will > supposedly be integrated into WindozeXP has nothing to do with? controlling the hardware (unless perhaps you live in some small E eastern European country or in China and are writing a virus). I just E love reading some of the posts on ZDNet's "talk back" - from the yoyo D that manages to include the phrase "Microsoft fights the good fight"C everytime he says something to all the people saying that all these < integrated features improve competition in the market place.  B Windows is Bill's onging attempt to turn anything computer related into his own little surfdom.   Joe    ------------------------------  % Date: Sat, 09 Jun 2001 09:17:16 +0100   From: Paul Sture <paul@sture.ch>A Subject: Re: Global warming? was: (OT) Current Microsoft v DOJ st + Message-ID: <VA.000003ba.467b07f9@sture.ch>   > In article <3B1F89E4.8AB70366@bbc.co.uk>, Tim Llewellyn wrote:/ > From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>  > Newsgroups: comp.os.vms G > Subject: Re: Global warming? was: Re: (OT) Current Microsoft v DOJ st ' > Date: Thu, 07 Jun 2001 15:04:20 +0100  >  >  >  > Bill Todd wrote: > J > > A study by the National Academy of Sciences (commissioned by the WhiteK > > House) was released (I think) just today.  It stated unequivocally that I > > global warming was real and that human activity was a major causative K > > factor.  Not that this is likely to convince those who still have their 2 > > heads stuck firmly in the sand (or elsewhere). > >  > # > Some people like being ostriches. I > Of course, you could blame it all on the water floridation programme...   & You mean chad infested water? Eek! :-)  - > Then again, us Brits have the lead pipes...    Could explain a lot... ___ 	 Paul Sure  Switzerland    ------------------------------   Date: 09 Jun 2001 06:45:35 GMT- From: "Bill Pedersen" <pedersen@ccsscorp.com> 1 Subject: Re: How can i set a DLT tapedrive online 0 Message-ID: <3b21c63f$1@kerberos.linuxpuppy.net>  G Well, the SET MAGTAPE probably would not work...  The following is from , OpenVMS 5.5-2 (its been there for a while...  G The SET MAGTAPE command is valid for magnetic tape devices mounted with  foreign volumes.   Format   SET MAGTAPE device-name[:]   Bill.  --
 Bill Pedersen  CCSS Corporation CCSS Interactive Learning 
 www.VMS.St 831-336-2708 ================  ? "Robert Deininger" <rdeininger@mindspring.com> wrote in message F news:rdeininger-0806012318530001@user-2ive75t.dialup.mindspring.com...K > In article <3B210615.4845DB57@hsc.vcu.edu>, Jim Agnew <Agnew@hsc.vcu.edu>  wrote: >  > > Actually, it's > > J > > set (devicename, in this case MAGTAPE is definitely a logical pointing to the > > devicename) /nounload  >J >2 > I don't see a  >u > $ SET devicename > 2 > in HELP.  And the SET command doesn't like this. >  > ???  >p > > or, one cano > >  > > MOUNT/NOUNLOAD TAPE: > @ > Ok, this is in HELP for VMS 7.1-2.  I never noticed it before. >V	 > Thanks.  >a > G > My baby steps into GKDRIVER to build a program to issue a SCSI "load"rI > command to a DLT aren't going well.  It seems I can't use GKDRIVER on a L > unit where MKDRIVER is already configured.  So if I did get the program toI > work, all I could do would be the load command.  Then I could reboot tooB > configure MKDRIVER, and do something useful with the drive.  :-) > / > Not too useful.  I'm looking for another way.E >RF > I did dig out the DLT 4000 manual.  There's definitely a "load tape"K > command, as well as a command to read out the actual compression ratio ofu- > the tape.  So this puzzle is worth solving.e >  > -- > Robert Deininger > rdeininger@mindspring.come   ------------------------------   Date: 9 Jun 2001 07:15:08 -0700v1 From: nothome@spammers.are.scum (Malcolm Dunnett)o1 Subject: Re: How can i set a DLT tapedrive online , Message-ID: <OpEcQvyH18Mg@malvm5.mala.bc.ca>  M In article <rdeininger-0806012318530001@user-2ive75t.dialup.mindspring.com>, c8     rdeininger@mindspring.com (Robert Deininger) writes: > F > I did dig out the DLT 4000 manual.  There's definitely a "load tape"K > command, as well as a command to read out the actual compression ratio ofo- > the tape.  So this puzzle is worth solving.i > I     FWIW, the "mt online" command on Tru64 Unix will load a tape in a DLT  drive.  =     Is this something the SCSI* programs in SYS$ETC could do?d   ------------------------------  % Date: Sat, 09 Jun 2001 12:12:06 -0400b2 From: rdeininger@mindspring.com (Robert Deininger)1 Subject: Re: How can i set a DLT tapedrive onlineCL Message-ID: <rdeininger-0906011212060001@user-2ive6e5.dialup.mindspring.com>  F In article <OpEcQvyH18Mg@malvm5.mala.bc.ca>, nothome@spammers.are.scum (Malcolm Dunnett) wrote:    ? >     Is this something the SCSI* programs in SYS$ETC could do?   J Dunno.  I haven't looked there in a while, but it sounds like a good idea.   --   Robert Deininger rdeininger@mindspring.comt   ------------------------------  % Date: Sat, 09 Jun 2001 09:17:17 +0100a  From: Paul Sture <paul@sture.ch>" Subject: Re: I demand your respect+ Message-ID: <VA.000003bd.467b0f67@sture.ch>o  S In article <8af17fe1.0106071415.499052b7@posting.google.com>, Aaron Sakovich wrote:-> > From: alphaman-nixspam@hsv.sungardtrust.com (Aaron Sakovich) > Newsgroups: comp.os.vmss$ > Subject: Re: I demand your respect! > Date: 7 Jun 2001 15:15:25 -0700: >  > fabio_compaq@ep-bc.petrobras.com.br wrote in message news:<OF1B29F8DC.6ED724D3-ON03256A64.003AB5EB@ep-bc.petrobras.com.br>...-I > > How many sysadmins in the world installed DECNET/OSI sucessfully ????s > >  > > 1,2, or 3 ?u > >  > > Regardsq > >  > > FC > G > Hate to disappoint you folks, but we've got several hundred customers>A > running it, and multiple systems in-house.  True, we only use asE > Local:. namespace, and try not to do anything terribly mind-numbingt > with it, but it does work. > G > Now, if only we could learn the commands.  Installation from a scriptnH > is one thing, running Decnet_Register is fine.  But I've had to resortB > to creating multiple symbols (e.g., Loop :== "ncl loop loop applF > name") to perform any kind of productive troubleshooting.  *sigh*, I> > miss the old "ncp loop node" syntax.  Almost English-like... > G Learning the commands is tricky and frustrating at first, but it _does_o come with practice.h  L Which is the rub really, since when I first came across it I would typicallyI only need to delve into NCP a couple of times a year. NCP would come backk! quickly in a way that NCL didn't.s  I I personally feel that the much of the abuse directed towards Phase V was J down to the lack of adequate ^H^H^H^H^H^H^H^H _any useful_ examples in theL documentation. By contrast, the Phase IV docs were full of working examples, in true VMS fashion. ___t
 Paul Sture Switzerlandi   ------------------------------  % Date: Sat, 09 Jun 2001 09:17:16 +0100   From: Paul Sture <paul@sture.ch>" Subject: Re: I demand your respect+ Message-ID: <VA.000003bc.467b0aca@sture.ch>i   In article ,I <35666012DF4CD411BE940090279FA240010BEF9C@ppnt41.physics.ox.ac.uk>, John f Macallister wrote:: > From: John Macallister <J.Macallister1@physics.ox.ac.uk> > Newsgroups: comp.os.vmst$ > Subject: RE: I demand your respect' > Date: Thu, 07 Jun 2001 11:53:20 +0100B > I > > How many sysadmins in the world installed DECNET/OSI sucessfully ????d >  > >1,2, or 3 ? > M > The answer is not 1,2,3 or even 42. It's 0. If a successful installation isEF > one which works, at least for basic things, DECnet/OSI cannot yet beB > installed successfully. For example, the DECnet/IP feature hangsF > consistently on network connections which have packet loss making it( > impossible to transfer large datasets. >  Now who's talking ROT? :-)  L I've got it running on over 30 systems at work. And on my Hobbyist Alpha at  home. Easy peasy.g  N OK I'll admit I did used to curse it, but it has definitely improved over the  years.   ___r
 Paul Sture Switzerland    ------------------------------  $ Date: Sat, 9 Jun 2001 13:02:37 -0500% From: Keith Brown <kbrown780@isd.net>  Subject: NCP counter question / Message-ID: <ti4ovm9pm8ua48@corp.supernews.com>e  K I support a VAX 4500 running OpenVMS 6.2 with Decnet phase 4, UCX v4.2 and   LAT.   $ mcr ncp show known line countC   gives the following output...  Line = ISA-0  n'       >65534  Seconds since last zeroed "      1251483  Data blocks received'       949453  Multicast blocks received             0  Receive failure     120744315  Bytes received&     84810051  Multicast bytes received            0  Data overrun       102533  Data blocks sent#         9207  Multicast blocks sent .           84  Blocks sent, multiple collisions+          132  Blocks sent, single collision -          649  Blocks sent, initially deferredr     80364475  Bytes sent"      1670383  Multicast bytes sent            0  Send failure,            0  Collision detect check failure,            6  Unrecognized frame destination'          217  System buffer unavailabler%            0  User buffer unavailablea  - Note 217 System buffer unavailable messages. t  F Question:  What do I need to tweek to fix this, is it cause for worry?   Thanks in advancee -- a Keith Brown  kbrown780@isd.neto   ------------------------------  % Date: Sat, 09 Jun 2001 09:17:18 +0100o  From: Paul Sture <paul@sture.ch>+ Subject: Re: NCP to NCL command equivilantsa+ Message-ID: <VA.000003be.467b124d@sture.ch>-  M In article <rdeininger-0806012358450001@user-2ive75t.dialup.mindspring.com>, c Robert Deininger wrote:M4 > From: rdeininger@mindspring.com (Robert Deininger) > Newsgroups: comp.os.vmsi- > Subject: Re: NCP to NCL command equivilantsg' > Date: Fri, 08 Jun 2001 23:58:45 -0400  > 7 > In article <3B211A11.E6A0E422@videotron.ca>, JF MezeiC' > <jfmezei.spamnot@videotron.ca> wrote:  > * > > "Brian Schenkenberger, VAXman-" wrote:> > > > I nominate this post for inclusion into the OpenVMS FAQ! > > P > > I think it should go on the on-line HELP in VMS under the NCL help topic for. > > those unfortunatle to have to run DECNET 5 > I > There is some NCP --> NCL conversion information in the DECnet 5 docs. M > For example: > R > http://www.openvms.compaq.com:8000/72final/6499/6499pro_001.html#convert_ncp_ncl > L > There's also the x-windows interface to NCL, which can be told to spit out2 > ncl command corresponding to whatever you click. > I Has anyone ever found that utility of any use except for read only mode? e > Q > > It isn't so much the commands themselves. One can use help to figure out. But N > > it is the structure of the configuration of your network and the lack of a > > configuration file.  > ; > There are a bunch of configuration files, written in NCL.e > I > > When I had installed decnet5, it had dumped a whole bunch of files in Q > > sys$manager (or is it sys$startup ?) and finding out what files did what, andrL > > which file issues the command that generates too many opcom messages was > the challenge. >  > Yes, that was annoying.a > R > > So it in my case, it wasn't so much what the syntax was, but rather where wereQ > > all the commands hidden in those files, and what could I do to override those|R > > commands and be sure that my overriding them will stay and not be crushed when; > > I run some decnet utility that regenerates those files./ > S This is almost certainly totally unsupported, but in SYS$SYSTEM you will find some wR NET$CHECKSUM*.DAT files. It is possible to edit the various startup .NCL files andO then preserve them by running CHECKSUM against the new version, and putting the0. result in the relevant NET$CHECKSUM*.DAT file.  N I just went through this last week when I was setting up a development clusterJ with multiple versions of VMS. I discovered that the MOP info had been setF up to load the system image from DISK$LABEL: instead of SYS$SYSDEVICE:  K A real pain to work through the NET$CONFIGURE menu for even one client, letn; alone multiple ones, so direct editing was the way to go...c  I > I discovered fairly early that DECnet 5 is a lot easier to use once youcJ > read the manuals.  Because it is slightly weird, it doesn't pay to startJ > in the middle of the manual like many in this group are probably used to > doing. > M Presuming you do have the full manual set. But that last sentence is spot on.  ___a
 Paul Sture Switzerlands   ------------------------------  % Date: Sat, 09 Jun 2001 12:22:42 -0400S2 From: rdeininger@mindspring.com (Robert Deininger)+ Subject: Re: NCP to NCL command equivilantszL Message-ID: <rdeininger-0906011222430001@user-2ive6e5.dialup.mindspring.com>  @ In article <VA.000003be.467b124d@sture.ch>, paul@sture.ch wrote:  O > In article <rdeininger-0806012358450001@user-2ive75t.dialup.mindspring.com>,   > Robert Deininger wrote:d  N > > There's also the x-windows interface to NCL, which can be told to spit out4 > > ncl command corresponding to whatever you click. > > K > Has anyone ever found that utility of any use except for read only mode? i   Well, I haven't.   ...u  P > I just went through this last week when I was setting up a development clusterL > with multiple versions of VMS. I discovered that the MOP info had been setH > up to load the system image from DISK$LABEL: instead of SYS$SYSDEVICE: > M > A real pain to work through the NET$CONFIGURE menu for even one client, let = > alone multiple ones, so direct editing was the way to go...e  E It looks like the MOP part of DECnet-plus is gradually falling out ofeI favor.  The LANCP version of MOP seems to get more engineering attention,wD and is available even without DECnet.  I've stopped using the MOP in DECnet-plus.  H CLUSTER_CONFIG_LAN.COM (instead of CLUSTER_CONFIG.COM) uses LANCP behindF the scenes instead of NCL.  For small MOP jobs, like strange satelliteJ boots to upgrade firmware, I just use LANCP directly; a couple of commands$ are enough to support a simple boot.  I In a large-scale environment, DECnet-plus and NCL might be easier for MOPB stuff...  K > > I discovered fairly early that DECnet 5 is a lot easier to use once youzL > > read the manuals.  Because it is slightly weird, it doesn't pay to startL > > in the middle of the manual like many in this group are probably used to
 > > doing. > > O > Presuming you do have the full manual set. But that last sentence is spot on.i  H Well, it's on the VMS CDs, and most of the manuals are on the web.  FullE hardcopy sets are a bit scarce these days, like all the VMS manuals. nA DECnet-4 hardcopies are easy to get; just find an old gray wall. s  DECnet-plus is post-gray wall...   -- i Robert Deininger rdeininger@mindspring.comh   ------------------------------  $ Date: Sat, 9 Jun 2001 11:22:47 +0400D From: "Ruslan R. Laishev" <junk-     laishev@mail.dls.net     -junk>' Subject: Need help with LIB$TABLE_PARSEe@ Message-ID: <e7kU6.15300$Ff4.752281@e420r-atl3.usenetserver.com>    Hello All, K         I'm need your advise or help, I'd like to use this stuff to parse a  string likeaK "TokenA: bla-bla-1 TokenB: bla-bla-2" to extract bla-bla-1 and bla-bla-2 ase "values"J of the tokens (TokenA and TokenB). Can someone write or explain to me what is the parser's table must be ?   ------------------------------  % Date: Sat, 09 Jun 2001 05:55:10 -0400h  From: Kuff@Tessco.Com (Hal Kuff)+ Subject: Re: Need help with LIB$TABLE_PARSEaO Message-ID: <2DCAFDDCC6C8E44F.9EE77F4A302CFA06.097E0A901D961FB3@lp.airnews.net>h  H In article <e7kU6.15300$Ff4.752281@e420r-atl3.usenetserver.com>, "Ruslan= R. Laishev" <junk-     laishev@mail.dls.net     -junk> wrote:   
 >  Hello All, M >         I'm need your advise or help, I'd like to use this stuff to parse a 
 > string likeeM > "TokenA: bla-bla-1 TokenB: bla-bla-2" to extract bla-bla-1 and bla-bla-2 ase
 > "values"L > of the tokens (TokenA and TokenB). Can someone write or explain to me what > is the > parser's table must be ?        +    Here is something from a samples library      PROGRAM NOTES:  I LIB$TABLE_PARSE is a new RTL function which is available in OpenVMS.  The F call to LIB$TABLE_PARSE() can be replaced with a call to LIB$TPARSE().E However, the compile commands for tparse.c and parse.mar will need to D include the /tie qualifier and the link command will need to includeD the /nonative qualifier.  If LIB$TABLE_PARSE() is to be called, thenD the compile and link commands in this document can be used as shown.  H Below is a command file that will create the C file, the Macro file, and= a linker options file, and then compile and link the program:g   $ set verify $ create tparse.c  /*B ** Program TPARSE uses LIB$TPARSE.  The main routine is written in. ** DEC C, with the state table coded in Macro.4 ** The state table has specified one action routine.F ** This action routine will be called if the state transition is true. **B ** This example of LIB$TPARSE will parse the SHOW DEVICES command: **$ **     SHOW DEVICES [device-name[:]] **. **     Command Qualifiers             Defaults **     /ALLOCATEDl, **     /BRIEF                         /BRIEF
 **     /FILESc, **     /FULL                          /BRIEF **     /MOUNTED  **     /[NO]SYSTEM **B ** If the program is successful, the action routine will print outF ** the name of the device you entered and the message 'syntax ok' willH ** be printed.  If there is an error, the message 'syntax error' will be ** printed.  */   /*9 ** The external symbols are defined in the macro routine.a */   #include <tpadef.h>i #include <libdef.h>  #include <lib$routines.h>t #include <stdio.h> #include <string.h>y #include <stdlib.h>r #include <ctype.h>  
 int tpars[9];  extern int ufd_state, ufd_key;  
 int action();s  
 int main() {v       char upstring[80];       char *bstring;"       char string[80], *stringptr;       int i, stringcnt; !       int status, options, count;,       int len, pos;l   /* ** Set parameter block count */         tpars[0] = TPA$K_COUNT0;   /*< ** Set options.  Process the blanks and allow abbreviations. */  -       tpars[1] = TPA$M_BLANKS | TPA$M_ABBREV;-   /* ** Read input string from user.9 */  B       printf(" Enter a SHOW DEVICES [device-name[:]] command.\n");       printf(" TPARSE>");D       gets(string);    /*! ** Convert lowercase to UPPERCASEp */  &       for (i=0; i<strlen(string); i++)*            upstring[i]=toupper(string[i]);$       upstring[strlen(string)]='\0'; /*9 ** Set the string length and address in the TPARSE vectorm */  #       stringcnt = strlen(upstring);        stringptr = upstring;    /* ** Call LIB$TABLE_PARSES */         tpars[2] = stringcnt;n        tpars[3] = (int)stringptr;<       status = lib$table_parse(tpars, &ufd_state, &ufd_key);   /* ** Print out the return values */  %       if (status == LIB$_SYNTAXERR) {n,           pos = strlen(upstring) - tpars[2];6           printf(" Syntax error at position %d", pos);:           printf(" The character is %c", upstring[pos-1]);       }d  2       if ( status == 1) printf(" SYNTAX IS OK\n");         return EXIT_SUCCESS; }    /* ** Here is the Action RoutineH */   int action() {n       char string[80];   /*D ** This is an Action Routine.  The state transition that occured, isF ** that a character was found were the device specification should be.A ** This routine will print the device specification, then return.  */  3       strncpy( string, (char *)tpars[5], tpars[4]);e       string[tpars[4]] = '\0'; /*8 ** Byte array string now contains the string.  Print it. */  +       printf(" The device is %s\n",string);          return 1;  }o   $ create parse.marE ; ************************* PARSE.MAR *******************************           .TITLE PARSEg ;+5 ; This macro routine will set up the state tables for. ; the show devices command.t ;-   ;+C ; Define parameter block.  You may fill the count and options fieldo- ; here or in the high level language routine.g ;-6        .PSECT TPARS,PIC,OVR,GBL,SHR,NOEXE,REL,WRT,LONG8 PARM:   .LONG   0                       ; LONGWORD COUNT0        .LONG   0                       ; OPTIONSA        .BLKL   7                       ; DEFINE REMAINING STORAGE  ;e ; Define control block offsets ;.        $TPADEF GLOBALd        $LIBDEF GLOBAL0 ;sA ; Begin definition of the state table.  UFD_STATE and UFD_KEY are ( ; the names of the state and key tables. ;u)        $INIT_STATE     UFD_STATE, UFD_KEYa ;oB ; The $STATE Macro designates that we are entering a new state and< ; a name for that state.  The name is used for branching to.H ; The $TRAN Macro specifies a keyword and, optionally, a state to branch ; to.L        $STATE  COMMAND        $TRAN   'SHOW', FUNCs ; C ; If the Token does not match any of the transition for that state, 7 ; an error indicating an syntax error will be returned.  ;b        $STATE  FUNCc        $TRAN   TPA$_BLANKt
        $STATEC        $TRAN   'DEVICES', DEV1 ;21 ; Command is SHOW DEVICES.  Check for / or blank.t ;e        $STATE  DEV!        $TRAN   TPA$_BLANK, DEVNAM.        $TRAN   '/', QUAL"        $TRAN   TPA$_EOS, TPA$_EXIT ;CC ; Blank found, check for device name.  If device name present, calli ; the action routine.s ;k        $STATE  DEVNAM +        $TRAN   TPA$_SYMBOL, ALLDONE, ACTION  ;-  ; Qualifiers found, check syntax ;         $STATE  QUAL7        $TRAN   'ALLOCATED', DEVl        $TRAN   'BRIEF', DEVm        $TRAN   'FILES', DEV         $TRAN   'FULL', DEV        $TRAN   'MOUNTED', DEV0        $TRAN   'SYSTEM', DEV        $TRAN   'NOSYSTEM', DEV        $TRAN   'WINDOWS', DEVn        $TRAN   TPA$_BLANK, QUAL "        $TRAN   TPA$_EOS, TPA$_FAIL ;: ; All done, return, status ok  ;u        $STATE  ALLDONE"        $TRAN   TPA$_ANY, TPA$_EXIT"        $TRAN   TPA$_EOS, TPA$_EXIT          $END_STATE6        .END- $ create tpars.opt. PSECT=TPARS,PIC,OVR,GBL,SHR,NOEXE,REL,WRT,LONG $ cc tparseI
 $ macro parse? $ link tparse,parse,tpars/opt  $ set noverify   ------------------------------  $ Date: Sat, 9 Jun 2001 14:52:03 +0400D From: "Ruslan R. Laishev" <junk-     laishev@mail.dls.net     -junk>+ Subject: Re: Need help with LIB$TABLE_PARSEn@ Message-ID: <qbnU6.15777$Ff4.770907@e420r-atl3.usenetserver.com>   Thanks Hal,,H     In my case I have not a separator like "/" and I'm need to extract aB character string between tokens but not from token till separator.  - "Hal Kuff" <Kuff@Tessco.Com> wrote in messagecI news:2DCAFDDCC6C8E44F.9EE77F4A302CFA06.097E0A901D961FB3@lp.airnews.net...gJ > In article <e7kU6.15300$Ff4.752281@e420r-atl3.usenetserver.com>, "Ruslan? > R. Laishev" <junk-     laishev@mail.dls.net     -junk> wrote:u >t > >  Hello All,tG > >         I'm need your advise or help, I'd like to use this stuff tot parse at > > string likeeL > > "TokenA: bla-bla-1 TokenB: bla-bla-2" to extract bla-bla-1 and bla-bla-2 as > > "values"I > > of the tokens (TokenA and TokenB). Can someone write or explain to mes what
 > > is the > > parser's table must be ? >^ >  >y >e- >    Here is something from a samples libraryt >h >h > PROGRAM NOTES: > K > LIB$TABLE_PARSE is a new RTL function which is available in OpenVMS.  The H > call to LIB$TABLE_PARSE() can be replaced with a call to LIB$TPARSE().G > However, the compile commands for tparse.c and parse.mar will need tosF > include the /tie qualifier and the link command will need to includeF > the /nonative qualifier.  If LIB$TABLE_PARSE() is to be called, thenF > the compile and link commands in this document can be used as shown. >sJ > Below is a command file that will create the C file, the Macro file, and? > a linker options file, and then compile and link the program:  >t > $ set verify > $ create tparse.c  > /*D > ** Program TPARSE uses LIB$TPARSE.  The main routine is written in0 > ** DEC C, with the state table coded in Macro.6 > ** The state table has specified one action routine.H > ** This action routine will be called if the state transition is true. > **D > ** This example of LIB$TPARSE will parse the SHOW DEVICES command: > **& > **     SHOW DEVICES [device-name[:]] > **0 > **     Command Qualifiers             Defaults > **     /ALLOCATEDt. > **     /BRIEF                         /BRIEF > **     /FILESs. > **     /FULL                          /BRIEF > **     /MOUNTEDl > **     /[NO]SYSTEM > **D > ** If the program is successful, the action routine will print outH > ** the name of the device you entered and the message 'syntax ok' willJ > ** be printed.  If there is an error, the message 'syntax error' will be
 > ** printed.t > */ >p > /*; > ** The external symbols are defined in the macro routine.  > */ >  > #include <tpadef.h>i > #include <libdef.h>6 > #include <lib$routines.h>e > #include <stdio.h> > #include <string.h>  > #include <stdlib.h>s > #include <ctype.h> >  > int tpars[9];   > extern int ufd_state, ufd_key; >1 > int action();y >  > int main() > {  >       char upstring[80]; >       char *bstring;$ >       char string[80], *stringptr; >       int i, stringcnt;l# >       int status, options, count;s >       int len, pos;  >  > /* > ** Set parameter block count > */ >   >       tpars[0] = TPA$K_COUNT0; >  > /*> > ** Set options.  Process the blanks and allow abbreviations. > */ >t/ >       tpars[1] = TPA$M_BLANKS | TPA$M_ABBREV;d >i > /*! > ** Read input string from user.i > */ > D >       printf(" Enter a SHOW DEVICES [device-name[:]] command.\n"); >       printf(" TPARSE>");  >       gets(string);k >  > /*# > ** Convert lowercase to UPPERCASEi > */ >e( >       for (i=0; i<strlen(string); i++), >            upstring[i]=toupper(string[i]);& >       upstring[strlen(string)]='\0'; > /*; > ** Set the string length and address in the TPARSE vectora > */ >-% >       stringcnt = strlen(upstring);  >       stringptr = upstring;5 >0 > /* > ** Call LIB$TABLE_PARSEg > */ >  >       tpars[2] = stringcnt;m" >       tpars[3] = (int)stringptr;> >       status = lib$table_parse(tpars, &ufd_state, &ufd_key); >o > /*  > ** Print out the return values > */ >8' >       if (status == LIB$_SYNTAXERR) {A. >           pos = strlen(upstring) - tpars[2];8 >           printf(" Syntax error at position %d", pos);< >           printf(" The character is %c", upstring[pos-1]);	 >       }  >!4 >       if ( status == 1) printf(" SYNTAX IS OK\n"); >M >       return EXIT_SUCCESS; > }  >s > /* > ** Here is the Action Routine  > */ >r > int action() > {  >       char string[80]; >  > /*F > ** This is an Action Routine.  The state transition that occured, isH > ** that a character was found were the device specification should be.C > ** This routine will print the device specification, then return.r > */ > 5 >       strncpy( string, (char *)tpars[5], tpars[4]);u  >       string[tpars[4]] = '\0'; > /*: > ** Byte array string now contains the string.  Print it. > */ > - >       printf(" The device is %s\n",string);r >f >       return 1;o > }r >t > $ create parse.marG > ; ************************* PARSE.MAR *******************************s >r >        .TITLE PARSE  > ;+7 > ; This macro routine will set up the state tables forl > ; the show devices command.i > ;- >r > ;+E > ; Define parameter block.  You may fill the count and options fieldo/ > ; here or in the high level language routine.  > ;-8 >        .PSECT TPARS,PIC,OVR,GBL,SHR,NOEXE,REL,WRT,LONG: > PARM:   .LONG   0                       ; LONGWORD COUNT2 >        .LONG   0                       ; OPTIONSC >        .BLKL   7                       ; DEFINE REMAINING STORAGEe > ;   > ; Define control block offsets > ;  >        $TPADEF GLOBALo >        $LIBDEF GLOBALg > ;tC > ; Begin definition of the state table.  UFD_STATE and UFD_KEY areu* > ; the names of the state and key tables. > ;M+ >        $INIT_STATE     UFD_STATE, UFD_KEYu > ;rD > ; The $STATE Macro designates that we are entering a new state and> > ; a name for that state.  The name is used for branching to.J > ; The $TRAN Macro specifies a keyword and, optionally, a state to branch > ; to.e >        $STATE  COMMAND >        $TRAN   'SHOW', FUNCh > ;eE > ; If the Token does not match any of the transition for that state,s9 > ; an error indicating an syntax error will be returned.E > ;F >        $STATE  FUNC  >        $TRAN   TPA$_BLANKo >        $STATEi >        $TRAN   'DEVICES', DEV  > ;v3 > ; Command is SHOW DEVICES.  Check for / or blank.e > ;  >        $STATE  DEV# >        $TRAN   TPA$_BLANK, DEVNAM  >        $TRAN   '/', QUAL$ >        $TRAN   TPA$_EOS, TPA$_EXIT > ;rE > ; Blank found, check for device name.  If device name present, callm > ; the action routine.s > ;c >        $STATE  DEVNAMl- >        $TRAN   TPA$_SYMBOL, ALLDONE, ACTION- > ;-" > ; Qualifiers found, check syntax > ;S >        $STATE  QUALs! >        $TRAN   'ALLOCATED', DEVc >        $TRAN   'BRIEF', DEVi >        $TRAN   'FILES', DEV- >        $TRAN   'FULL', DEV >        $TRAN   'MOUNTED', DEVc >        $TRAN   'SYSTEM', DEV  >        $TRAN   'NOSYSTEM', DEV >        $TRAN   'WINDOWS', DEV0! >        $TRAN   TPA$_BLANK, QUALm$ >        $TRAN   TPA$_EOS, TPA$_FAIL > ;r > ; All done, return, status okt > ;, >        $STATE  ALLDONE$ >        $TRAN   TPA$_ANY, TPA$_EXIT$ >        $TRAN   TPA$_EOS, TPA$_EXIT >e >        $END_STATEy
 >        .ENDe > $ create tpars.opt0 > PSECT=TPARS,PIC,OVR,GBL,SHR,NOEXE,REL,WRT,LONG
 > $ cc tparse  > $ macro parsee > $ link tparse,parse,tpars/opt  > $ set noverify   ------------------------------  # Date: Sat, 09 Jun 2001 11:51:12 GMTe From: LBohan@dbc.spam_less..coms Subject: Re: NYSEC8 Message-ID: <d224itkv7tffvorc47hc21be9ud5q8qu0b@4ax.com>  C On 8 Jun 2001 10:24:34 -0700, svieth@wi.rr.com (Scott Vieth) wrote:w  E >One of the four exchanges in Chicago runs on VMS.  I know one of theuC >other ones in Chi-town(CBOT?) made a big deal out of the fact thato  >they're running the show on NT. >-Scotte  5 the Chicago exchange, the one supposedly running NT, n9 (a regional stock exchange iirc), used to be a VMS shop, I3 one that coded most of their applications in-house   using  VAX-Basic.   / I'd heard, second/thrird-hand, that over time, I' that their in-house trading s/w became m3 the s/w from hell,  that noone still  there really r6 understood it anymore,  the original coders long gone.  9 despite the NT boxes in front, I believe they still have C/ a few Vaxen in the bgnd, not yet decommisioned.t (missing source code?)  ; given that this sort of software lifetime development cyclef0 is all too common/ingrained, I'd wager in about 5 10 years they'll be climbing back onto the same boat,u albeit w/ Visual Basic.    ------------------------------  % Date: Sat, 09 Jun 2001 11:08:54 +0200a) From: Christof Brass <brass@infopuls.com> , Subject: Re: Portability. The myth continues, Message-ID: <3B21E7A6.66837D23@infopuls.com>   Paul Repacholi wrote:. > - > Christof Brass <brass@infopuls.com> writes:t > ( > > These are a good but hard questions:@ > > - should we aim for portability from VMS to other platforms?< > > - how do we get people to use the language(s) of choice? > H > Well, what 'other platform'? Oh, unix, well, to which language on thatD > other platform? C, yes fine, do you mean with pcc call convention,. > 'standard' C calling, or the gcc variety?...  ; Good point. Normally you define portability on the languagea@ level in the hope the language is that well defined that it will> layer/hide the HW or the implementation specifics. But when it; comes to OS APIs there are only a few ways to go. The one I-? normally choose is to use my abstractions which are implemented.@ in the most efficient manner on each platform. This is more work* than to define a least common denominator.  A > Note, that this sort of thing bites before you even have a full1C > program, and that it ignores the hackmem problem. So at the best,v1 > portability is a myth, and is normally a fraud.   > Unfortunately I have to agree. If you've read some of my posts< and suggestions to this topic you know that I'm very sceptic$ about this general portability idea.  @ > If we get further up the chain, it gets worse. How do you do aA > 'portable' ast? How do you portably turn them off in a critical?@ > section? Note you don't even have to use an AST to be 'got' by
 > the second.$ > F > If I could have one thing happen to enhance portability, it would beD > for T64 to have *A* call standard, and all languages adhere to it,> > and for it to be forced into COE as an absolute requirement.  > Or standardise on the language and the language's libraries as@ opposed on the lower level like .NET. This will exclude a lot of
 languages.   >  > --> > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda.oB >                                              West Australia 60760 > Raw, Cooked or Well-done, it's all half baked.J > Spam-To: uce@ftc.gov,enforcement@sec.gov,sness@fcc.gov,hfurchtg@fcc.gov,$ >   mpowell@fcc.gov,gtristan@fcc.gov   ------------------------------  % Date: Sat, 09 Jun 2001 09:17:16 +0100e  From: Paul Sture <paul@sture.ch>7 Subject: Re: SKC Writeup on VMS Diamond and Tech Forumsr+ Message-ID: <VA.000003bb.467b0957@sture.ch>   G In article <iXMT6.18128$QP6.7702608@typhoon.ne.mediaone.net>, Terry C. s Shannon wrote:6 > From: "Terry C. Shannon" <terryshannon@mediaone.net>B > Newsgroups: comp.os.vms,vmsnet.alpha,comp.org.decus,comp.sys.dec5 > Subject: SKC Writeup on VMS Diamond and Tech Forumsc% > Date: Thu, 07 Jun 2001 15:18:38 GMTo > L > A review of the recent VMS Forums in Europe may be found on the cyberpagesC > of The Enquirer. And yes, the review is formatted in HTML. Enjoy!* > ) > http://www.theinquirer.net/07060111.htmr > M And for those interested in the CockpitMgr product mentioned in the Brussels  I presentation, but who cannot understand Netherlands (Dutch) or French to u5 navigate there, here's the url for the presentations:x  L http://www.compaq.be/main/frameset.asp?reference=08-01.01-13&lang=fr&undefin en  / (Unfortunately all .PPS Powerpoint slide shows), ___i
 Paul Sture Switzerlandn   ------------------------------  # Date: Sat, 09 Jun 2001 14:38:38 GMT*4 From: "Terry C. Shannon" <terryshannon@mediaone.net>7 Subject: Re: SKC Writeup on VMS Diamond and Tech Forums : Message-ID: <OxqU6.3262$Tc.634511@typhoon.ne.mediaone.net>   > >BE > And for those interested in the CockpitMgr product mentioned in thet BrusselsJ > presentation, but who cannot understand Netherlands (Dutch) or French to7 > navigate there, here's the url for the presentations:/ >  > L http://www.compaq.be/main/frameset.asp?reference=08-01.01-13&lang=fr&undefin > e  >p1 > (Unfortunately all .PPS Powerpoint slide shows)  >g  @ Blame CPQ for that one, not me ;-}. Anyhow, I was impressed withL CockpitMgr... I worked for a VMS system management software vendor (CIS, nowK Heroix) back in the late 80's, and CockpitMgr does some very nice stuff. IttH also integrates with Tecsys Development Inc.'s Consoleworks, which shipsJ with every GS-Series system. Both products are well worth taking a look at5 if you're considering enterprise management software.a   terry sp   ------------------------------  % Date: Sat, 09 Jun 2001 11:15:33 +0200]) From: Christof Brass <brass@infopuls.com>n Subject: Re: The future of VMS+ Message-ID: <3B21E935.78B03E3@infopuls.com>e   Chris Casey wrote: > E > Christof Brass wrote in message <3B2146CF.490706F3@infopuls.com>...i > >[SNIP; > >This makes completely sense. The tide integration of then@ > >functions you mentioned explains the short development times.? > >But I suspect that other tools are more appropriate today as(0 > >some of the restrictions don't apply anymore. > 0 > The original terseness was used to save space.E > Many people write much more readable code nowadays and I have stills@ > to see anything as powerful with such short development times.  ; And one of my favourite topics: oo. I don't expect MUMPS to > offer that kind of features. And this is a severe drawback for creating huge systems.  ; > >I studied the Cach Web site and was rather impressed. IuA > >couldn't find a clear indication and even sent an email asking 9 > >that what PL has to be used to use the powerful stored,= > >procedures and object DB interface but the example I foundO > >looked like C++.. >  > The options are .... > G > CacheObjectScript - derived from the original Mumps and usable on anyG > supported platform.l > C.? > Any COM enabled language - VB, Delphi, C+ etc on Billy boxes.  > Corba. > Java.o
 > ODBC, JDBC.lJ > and several others - I can't be bothered to go to the site to refresh my > memory of the full list. > > > >There is nothing to hate with that concept as long as it is > >consistent.B > >If you look at the current object DBs they try to offer similar > >features.C > >The idea is to make the boundary between persistent and volatile < > >data almost invisible, but of course completely under the > >control of the programmer.t > J > This is exactly what we have had in Mumps for 20 years before the object > paradigm gained popularity.N  ; But this is only one ascpect for efficiency. If you have to'= organise the work for a bunch of programmers and want to haveb@ easy to reuse pieces of SW you need additional features, e.g. oo for creating an oo framework.   C > >Do you know about the DSM implementation compared to others likee > >the one on AIX? > J > DSM (Digital Standard Mumps) was the main implementation on VMS for some! > time and is still very popular. N > The other competitors were ISM (Intersystems Standard Mumps) which came fromM > the Unix platforms and got ported to VMS and GT.M (a fully compiled version  > now owned by Sanchez). > G > DSM was the one used by Digital when they benchmarked against Rdb andY	 > Oracle. J > They were trying to promote Rdb at the time so when DSM came in 10 times > faster; > at 1/20th of the cost they forgot to promote the figures. B > They sold DSM to Intersystems as part of the big sell off phase.  @ Okay then the question should be re-phrased: do you know how the3 DSM/ISM implementation on VMS compares to the MUMPSO3 implementation on AIX? I heard that all these MUMPS,= implementations offer very good portability. Is this correct?    ------------------------------  % Date: Sat, 09 Jun 2001 14:25:11 +0200l) From: Christof Brass <brass@infopuls.com>: Subject: Re: The future of VMS, Message-ID: <3B2215A7.5C957521@infopuls.com>   Hunter Goatley wrote:h > P > On Fri, 08 Jun 2001 23:53:48 +0200, Christof Brass <brass@infopuls.com> wrote: >  > >Brian Tillman wrote:n > >>E > >> >What I would like to see is a tool which automatically converts A > >> >the command line interface of foreign programs to a CLD/VMSp > >> >style interface. > >>F > >> It's a simple thing to produce.  Use the following as a template: > >> > >> DEFINE VERB foreign6 > >>     IMAGE foreigndir:foreign, CLIFLAGS( FOREIGN ) > >'A > >I don't understand this. Will this provide all the options andt@ > >parameter conditions automatically that I could run the image > >like any other VMS image? > ? > No, it just lets you define a real command to invoke an imageeA > you normally have to define a foreign symbol for.  There's more @ > to implementing DCL-style /QUALIFIERS than that.  I've done it? > for several utilities, most notably ZIP and UNZIP.  It can befD > ridiculously easy to do, and it can be extremely tricky, depending- > on what the underlying application expects.. > A > The trick I used was to parse the DCL command line and create atC > new argv[] array that consists of all the equivalent UNIX commandcD > options.  This new argv[] array gets returned to the main routine.8 > The code underneath doesn't have to change, other than" > adding a call to the DCL parser.  ? This is a smart shortcut. What I like to see is a conversion of ? the main part into a typical VMS main part. The difference fromt> my perspective is that with the VMS style you don't need a lot4 of checks that are necessary in the non-VMS version.  1 > >BTW, on my VMS 6.2 there is no DEFINE VERB :-(T > @ > DEFINE VERB is recognized by the command SET COMMAND xxxx.CLD.A > What Brian gave above is a simple .CLD file that will eliminatei  > the need for a foreign symbol.  @ Thanks, I wrote CLD files myself several years ago but obviously! didn't remember the CLD language.i   > Hunter > ------; > Hunter Goatley, Process Software, http://www.process.com/ ; > goathunter@goatley.com     http://www.goatley.com/hunter/    ------------------------------  % Date: Sat, 09 Jun 2001 14:31:40 +0200 ) From: Christof Brass <brass@infopuls.com>i Subject: Re: The future of VMS, Message-ID: <3B22172C.9BFE86B8@infopuls.com>  " Shane.F.Smith@Healthnet.com wrote: > J > It's possible in any language, just not always worth it. I remember onceI > doing recursion in a language that didn't handle local arguments (and IiK > wish I could remember which one, now. It was easily twelve years ago.). IiI > had to save and restore the values for each layer of recursion manuallye- > across calls in my own stacks. Never again.u >  > Shane_   :-)1  > This reminds me to the essay by Bjarne Stroustroup (spelling?)3 about oo and C++ which came with the first AT&T C++,< pre-compilers. He stated that you could programm oo in every@ language but some languages only make it possible whereas others support it.u  9 I read a FORTRAN manual at that time which contained this*& statement about missing recursiveness.   ------------------------------   Date: 9 Jun 2001 10:23 CDT' From: carl@gerg.tamu.edu (Carl Perkins)  Subject: Re: The future of VMS, Message-ID: <9JUN200110230319@gerg.tamu.edu>  ? "Brian Tillman" <tillman_brian@notnoone.notnohow.com> writes...rA }>What I would like to see is a tool which automatically converts = }>the command line interface of foreign programs to a CLD/VMS] }>style interface. } B }It's a simple thing to produce.  Use the following as a template: }  }DEFINE VERB foreign2 }    IMAGE foreigndir:foreign, CLIFLAGS( FOREIGN ) }-- B }Brian Tillman                   Internet: tillman_brian at si.com  C That didn't convert squat. It didn't change the command line flags,)B say "-foo" for example, to qaulifiers (like "/foo" for example) orA anything. It just kept everything the same except that once it is @ in the command tables you don't have to define the symbol to run it foreign.2  D Automatically converting interfaces would be quite difficult - you'dD have to parse the soruce code of the program to see what it actually@ does with the command line data, including deciding if any givenE element should be a qualifier or a parameter and figuring out if they C are grouped (example: the differece between "program arg1 arg2" and*E the very similar looking "program arg1 value1" , the former might end*C up as "program/arg1/arg2" and the latter as "program/arg1=value1").   A I have some serious doubts as to whether or not it is possibly to C do this interface style conversion automatically for anything other  than a very simple case.   --- Carl   ------------------------------   Date: 9 Jun 2001 07:42:23 -0700 1 From: martin.ejdestig@planit.se (Martin Ejdestig)  Subject: Re: Threads. = Message-ID: <19468064.0106090642.7f2477ca@posting.google.com>A  d wspencer@ap.nospam.org (Warren Spencer) wrote in message news:<ti2djo5slif6b9@news.supernews.com>...7 > martin.ejdestig@planit.se (Martin Ejdestig) wrote in  4 > <19468064.0106052337.5d3b3ce0@posting.google.com>: >  > M > Some extra info would help.  Are you using the DECthreads (cma) library or  * > pthreads?  Version?  Version of OpenVMS?  C From what I can understand the recommended way to use DECthreads isTD through the pthreads interface (all other ways are obsolete) when itF comes to the version Im running (see below). Correct me if Im wrong :)   > L > My comments below refer to the pthreads - the only interface I'm familiar < > with (on Linux, not OpenVMS, but it *should* be the same). > D > >I'm a bit confused when it comes to threads and asts. I'm running& > >version V7.1-2 on an Alpha machine. > > G > >1. When does asts interrupt the thread that requested it (per-threadf> > >AST) and when are they delivered to the process as a whole? > >rI > >2. If I use sys$setast() to disable asts is this on a per-thread basis F > >or for the process as a whole? If to the process, mustn't I preventI > >other threads from enable asts with sys$setast() (otherwise I can't betE > >sure that asts are disabled in the thread that disabled asts...you  > >know what I mean...). > H > AST's pre-date threads.  If you need your threads to wait for certain J > conditions, refer the documentation for pthread condition variables.  A F > (pthread) Mutex is the appropriate mechaism for protecting critical C > sections of code and serializing access to data structures.  Use  G > pthread_setcancelstate() and pthread_setcanceltype() to control when T? > threads can be destroyed, and what actions they take on exit.   @ Well...my library is supposed to be threadsafe. The library alsoB contains some thread functions. When these thread functions aren'tD used (single threaded proggie) I don't want ANY pthread functions to0 be called. And for that I use the tis interface.  C Btw...how does one accomplish the same thing on platforms where the0C tis interface isn't available (Linux and other unixes besides Tru64_D for example)? Must one use the pthread() functions. If yes...doesn'tE this cause unnecessary overhead in single threaded proggies that uses, the lib?   > H > >3. I've made a library that uses some global variables to store errorI > >information. This must be stored for each thread. To accomplish this IiE > >must use tis_key_create()....right (Im not 100% sure on how to use-G > >it...care to write an example :)? Is there some way I can initializesH > >this atomatically when linked with my library (ie without calling anyI > >MyLib_Initialize() routine) (and deinitlaize atomatically when my progw
 > >exits).  > Is there something like __init_ (the loader calls all __init_*> routines when the library is loaded) and __fini_ (all __fini_*F routines are called when the library is unloaded) routines on Tru64? IB tried to look at LIB$INITIALIZE but the documentation tells you toD avoid it and frankly I didn't get that much of it. The docs are wery* sparing when it comes to LIB$INITIALIZE :(  4 If there isn't I'll have to make do with tis_once().   > > G > >4. Ive made a "wait for one of these events to occur" function whichrG > >uses poll() in the unix version and asts, sys$hiber() and sys$wake()TD > >(in the asts) in the vms version. I've understood that sys$wake()H > >wakes all threads that have called sys$hiber(). To solve this problemH > >Ive gotta check somehow if the sys$wake() was for me (the thread) andH > >if it wasn't do sys$hiber() again. Is there a better way (ie a way toI > >just sys$wake() the thread that that should be awaken)?...note that itu- > >should work in a single threaded proc. to.r > N > A pthread condition variable seems appropriate here, although an Event Flag N > might work equally well, and would typically be the mechanism of choice for 4 > this type of problem in non-threaded environments.  F Aren't eventflags shared between threads? If I use a pthread_cond_waitD mustn't the pthread library be loaded even if it's a single threadedE prog? Do you mean that I should use pthread_cond_signal() in my asts?o  = tis_cond_wait doesn't behave appropiate when single threaded.c   > C > >5. When a thread has called the "wait for one of these events tooE > >occur" function above, it must be able to be notified from anotheraE > >thread that it should exit. Since I use poll() on unix I acomplisheG > >this by using pipe() and write()...dunno if this is any good though.aF > >What is the best approach on vms (see above for how Ive implemented9 > >the "wait for one of these events to occur" function)?r > I > You'd likely be able to implement this with fewer "moving parts" using tN > Event Flags or pthread condition variables, obviating the need for the pipe  > entirely.c >  > Hope this helps.  E A little maybe...I'm happy that someone responded at all :)...thoughtl0 I was going to be all by myself with this one ;)   >  > ws   /Martin Ejdestig   ------------------------------  $ Date: Sat, 9 Jun 2001 18:33:50 +0200, From: "Harald Thienel" <harald@h-thienel.de>6 Subject: Re: Urgent: Oracle 8 parallel server on VMS 7/ Message-ID: <9ftjc6$9ov$05$1@news.t-online.com>    Hi,v  A thanks for your response! The architecture of the database or thedC applications did not change vom Oracle Version 7, where it run very 
 satisfying1 What can we change to make the system run better?n   Regards  Harald  ; "chan01" <chandra_2000_us@yahoo.com> schrieb im Newsbeitrag 7 news:480cb9d4.0106081246.5320fecb@posting.google.com...l > Hi,vK >  I think it is not the problem with OPS Oracle, it is depending up on thecD > database architecture. slow because of the LMD. may be you have to: > change your archi. what kind of system you were running. >  > Thanks9 > "Harald Thienel" <harald@h-thienel.de> wrote in messageo+ news:<9fr6la$6jh$03$1@news.t-online.com>...gF > > Since one week we are running Oracle 8.1.7 on VMS-Alpha-Cluster (3	 machines)n2 > > in parallel server mode. VMS Version is 7.2.1.K > > there is one instance of the database on every machine. The database isR used3 > > by 200-300 users distributed on the 3 machines.  > >0H > > First problem is that Oracle 8 seems to be several times slower than Oracle > > 7 (maybe 4 times). > > I > > But the big problem is that after some hours (1 to 3), Oracle becomessG > > getting slower and slower and then stops processing completely. Alli users J > > are waiting forever. This seems to be a problem in the lock managementI > > (LMD-process). Sometimes this state can be cured by shutting down one K > > instance of the database (then the other two resume working), sometimest allwG > > three instances must be shut down and restarted. In a manufacturingkC > > environment like ours, these shutdowns are not very wellcome...n > >dB > > We are in contact with Oracle-Support and tweeking around with parameters,e' > > but this didn't really help so far.m > >r; > > The situation is escalating - any help is very welcome!r
 > > Thanks
 > > Harald   ------------------------------  $ Date: Sat, 9 Jun 2001 18:37:34 +0200, From: "Harald Thienel" <harald@h-thienel.de>6 Subject: Re: Urgent: Oracle 8 parallel server on VMS 7/ Message-ID: <9ftjc7$9ov$05$2@news.t-online.com>e   Hi,m  E yes we have many processes in MUTEX state. This is the state when the ) processes stop working and start waiting.p8 We will try the expire_time setting as soon as possible.   thanks Harald  > "Mitchell, David R." <mitchell@WPI.EDU> schrieb im NewsbeitragD news:FE1835D68492D311BF7900508B5BEB0D04499653@petra.admin.wpi.edu...I > We're not running parallel server, but we just upgraded to 8.1.7 on our. AlphaeD two weeks ago.  Immediately afterwards we started getting a bunch ofL processes in "MUTEX" state.  Over time we'd eventually start having problemsI with new connections and would have to shutdown and restart the database.sH Our DBA was working with Oracle support and eventually found a DCD (DeadI Connection Detection) parameter that was set to 10 minutes and eventuallylL caused resouce quota problems.  Setting the value of sqlnet_expire_time to 0L minutes (disabling it) in sqlnet.ora has fixed this problem for us.  Are youI seeing a large number of processes in a MUTEX state with a "show system"?f   DavidV   David Mitchell' Administrative Computer Systems Managera# Computing and Communications Centere Worcester Polytechnic Institutef e-mail:mitchell@wpi.edu  phone:(508)831-6044e     > -----Original Message-----3 > From: Harald Thienel [mailto:harald@h-thienel.de]e% > Sent: Friday, June 08, 2001 2:50 PM  > To: Info-VAX@Mvb.Saic.Coma4 > Subject: Urgent: Oracle 8 parallel server on VMS 7 >m >./ > Since one week we are running Oracle 8.1.7 onp  > VMS-Alpha-Cluster (3 machines)0 > in parallel server mode. VMS Version is 7.2.1.= > there is one instance of the database on every machine. Then > database is used1 > by 200-300 users distributed on the 3 machines.k > : > First problem is that Oracle 8 seems to be several times > slower than Oracle > 7 (maybe 4 times). >nG > But the big problem is that after some hours (1 to 3), Oracle becomess5 > getting slower and slower and then stops processing  > completely. All users H > are waiting forever. This seems to be a problem in the lock managementG > (LMD-process). Sometimes this state can be cured by shutting down one ? > instance of the database (then the other two resume working),l > sometimes all E > three instances must be shut down and restarted. In a manufacturingaA > environment like ours, these shutdowns are not very wellcome...o > ; > We are in contact with Oracle-Support and tweeking arounde > with parameters,% > but this didn't really help so far.  >n9 > The situation is escalating - any help is very welcome!  > Thanks > Harald >a >e >    ------------------------------   Date: 9 Jun 2001 10:27 CDT' From: carl@gerg.tamu.edu (Carl Perkins)u2 Subject: Re: [OT] Current Microsoft v DOJ status ?, Message-ID: <9JUN200110272307@gerg.tamu.edu>  5 "David J. Dachtera" <djesys.nospam@fsi.net> writes...  }Bob Koehler wrote:hT }> In article <9fqs30$hk2$7@husk.cso.niu.edu>, system@niuhep.physics.niu.edu writes: }> >E }> > I grew up 5 miles from Lake Michigan, anybody else here know then# }> > meaning of "lake effect snow"?  }> aH }> My mothers family survived for decades in the lake effect region when2 }> there was no 4WD anything outside the military. } H }Back in Jan. '67, my mom drove home from work (about seven(7) miles) in7 }circa. 24 inches of snow in a 1966 Ford Galaxie (RWD).e }  }--  }David J. Dachtera  F And did that car get better or worse mileage per gallon than a current day 4WD SUV?   --- Carl   ------------------------------   End of INFO-VAX 2001.318 ************************