1 INFO-VAX	Mon, 09 Oct 2000	Volume 2000 : Issue 564       Contents:C Re: ??== Delete an interface from the TCPIP$CONFIGURATION.DAT file.  ADV: Search Engine Registration ) Backup image - restore, free blocks delta - Re: Backup image - restore, free blocks delta ) Backup image - restore, free blocks delta  Re: Exec vs Super  Re: Exec vs Super 7 Re: Get Viagra at the lowest prices available anywhere! 7 RE: Get Viagra at the lowest prices available anywhere! ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS ' Re: Getting Compaq to advertise OpenVMS  IO$M_NOW and IO$M_NORSWAITF Re: Musical Chairmen in Houston... Rosen Steps Down, Capellas Steps Up1 Re: Seeking info/prices for OpenVMS and hardware. $ Re: Why has Compaq retired DECnet ??$ Re: Why has Compaq retired DECnet ??, Re: Why is it so hard to port things to VMS?, Re: Why is it so hard to port things to VMS?, Re: Why is it so hard to port things to VMS?  F ----------------------------------------------------------------------  % Date: Mon, 09 Oct 2000 04:56:07 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)L Subject: Re: ??== Delete an interface from the TCPIP$CONFIGURATION.DAT file.; Message-ID: <39e133c7.524144494f47414741@radiogaga.harz.de>   - Hans M. Aus (aus@vim.uni-wuerzburg.de) wrote: F : How can I delete an interface from the TCPIP$CONFIGURATION.DAT file?  ! TCPIP> SET CONFIG NOINTERFACE WE1    cu,    Martin --J One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer7 One OS to find them           | work: mv@pdv-systeme.de N One OS to bring them all      |       http://www.pdv-systeme.de/users/martinv/> And in the Darkness bind them.| home: martin@radiogaga.harz.de   ------------------------------  % Date: Mon, 09 Oct 2000 13:30:25 +0800  From: mike97@arabia.com ( Subject: ADV: Search Engine Registration/ Message-ID: <"200010090530.NAA17690"@netranger>    Removal instructions below  # I saw your listing on the internet.   % I work for a company that specializes # in getting clients web sites listed ! as close to the top of the major   search engines as possible.   # Our fee is only $29.95 per month to ! submit your site at least twice a ! month to over 350 search engines   and directories.  $ To get started and put your web site$ in the fast lane, call our toll free
 number below.      Mike Bender  888-532-8842  & To be removed call: 888-800-6339 X1377   ------------------------------  # Date: Mon, 09 Oct 2000 00:22:55 GMT 0 From: gilley@nospam.bravewc.com (Charles Gilley)2 Subject: Backup image - restore, free blocks delta7 Message-ID: <zd8E5.5369$zx.15079@skycache.prestige.net>   J I have recovered an image backup of an RF73.  Assuming that the system in M question was booted minimum (no locked files) and no files on the volume are  J set nobackup, under what circumstances would I expect to see fewer blocks J restored than on the original disk?  I know such an operation defrags the O drive, but would the restore have any other effect?  I have about a 140K block  J delta, which isn't much, but it is something.  Would the restore have any ! affect on a file's extent values?    Thanks   ------------------------------  % Date: Sun, 08 Oct 2000 21:14:15 -0500 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> 6 Subject: Re: Backup image - restore, free blocks delta- Message-ID: <39E129F7.A54B8364@earthlink.net>    Charles Gilley wrote:  > K > I have recovered an image backup of an RF73.  Assuming that the system in N > question was booted minimum (no locked files) and no files on the volume areK > set nobackup, under what circumstances would I expect to see fewer blocks K > restored than on the original disk?  I know such an operation defrags the P > drive, but would the restore have any other effect?  I have about a 140K blockK > delta, which isn't much, but it is something.  Would the restore have any # > affect on a file's extent values?   F Well, the first thing that comes to mind would be the INDEXF.SYS file.H The current usage (after the /IMAGE restore) represents ONLY the currentC usage, and not any usage of INDEXF.SYS that may have existed in the B past. That is, the INDEXF.SYS may have been extended in an earlier5 "life" to accomdate a higher utilization of the disk.   B The same might be true of directory files. Blocks are allocated to@ directories as they need to be extended. However, when files areD deleted, the allocation quantity of any directory will remain at theB highpoint and will, of course, not be released unless you manually" truncate them (SET FILE/TRUNCATE).  E This might not account for the entire difference, but can represent a  good chunk of it.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  % Date: Mon, 09 Oct 2000 00:32:34 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>2 Subject: Backup image - restore, free blocks delta7 Message-ID: <200010090032_MC2-B639-8188@compuserve.com>   C         If the two disks have different cluster sizes, all the file G allocations will be different.  A smaller cluster size generally wastes  less space.     & Message text written by Charles GilleyJ >I have recovered an image backup of an RF73.  Assuming that the system i= n =   H question was booted minimum (no locked files) and no files on the volume are =   J set nobackup, under what circumstances would I expect to see fewer blocks=  =  J restored than on the original disk?  I know such an operation defrags the=  =  H drive, but would the restore have any other effect?  I have about a 140K block =   J delta, which isn't much, but it is something.  Would the restore have any=  =  ! affect on a file's extent values?  <    ------------------------------  % Date: Sun, 08 Oct 2000 12:58:55 -0500 / From: Chris Scheers <chris@applied-synergy.com>  Subject: Re: Exec vs SuperO Message-ID: <4221A8E6E3BEC99D.1B538BBFB79618C5.5198E2D1C58FC71B@lp.airnews.net>    John Santos wrote: >  > J > NO!  The default for creation in DCL ($ assign, $ define) is SUPER mode,E > but the default for translation, as Arne says, is USER mode, so the C > user can override inner mode logical names by assigning them in a I > program or with assign/user, but a program can defeat this by insisting > > on inner mode names by explicitly specifying an access mode. > H > I checked HELP on Alpha VMS 7.2-1.  LIB$SET_LOGICAL explicitly createsE > SUPER mode logicals.  SYS$CRELNM seems to default to USER mode, but I > HELP isn't explicit.  (ACMODE is an optional argument, but HELP doesn't # > say what happens if you omit it.)   H SYS$CRELNM can create logicals at any mode, restricted by the privilegesG (specificly SYSNAM) and access mode of the caller.  Thus, a normal USER 0 mode program can only create USER mode logicals.  E LIB$SET_LOGICAL is a service provided by DCL.  Since DCL runs in SUPR 4 mode, LIB$SET_LOGICAL can create SUPR mode logicals.  H SYS$CRELNM is always available, but LIB$SET_LOGICAL is only available ifG you are running with DCL.  In particular, detached programs tend not to  be able to use LIB$SET_LOGICAL.   H BTW: USER mode logicals are removed at image rundown only if they are inG the process logical name table.  USER mode logicals in other tables are  not removed.  G ----------------------------------------------------------------------- $ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com     Fax: 817-237-3074    ------------------------------  % Date: Mon, 09 Oct 2000 09:11:58 +0010 % From: paddy.o'brien@zzz.tg.nsw.gov.au  Subject: Re: Exec vs Super5 Message-ID: <01JV4MCRY02Q00554U@tgmail.tg.nsw.gov.au>    Richard Gilbert wrote:  A >Message text written by INTERNET:paddy.o'brien@zzz.tg.nsw.gov.au D >>To go back to the original, and something that I have never really >understood:4 >I have (in the order listed by show logical), e.g.: > & >"SYS$DISK: [super,crelog] = "DKA100:" >"SYS$DISK: [exec] = "DKA100:" > I >what really is the distinction?  How can I use the logicals differently?  >  >Why are there two?  >< > K >        You evidently did not include the full output of SHOW LOGICAL here L >but if you did, it would show that the SUPER mode logical is in the process@ >table and the EXEC mode logical is in the system table.  So theH >distinctions are the access mode and the name of the logical name tableI >that each is in.  I had never noticed the "crelog" before you pointed it H >out but it suggests to me that the logical is being created by a reallyJ >primitive part of VMS; something that goes back to VMS V3.x or before.  =  	 [snipped]   K The logicals I quoted were taken from a $show logical/process/full command.   9 Interestingly, if I do $show logical/full sys$disk, I get   9 "SYS$DISK" [super,crelog] = "DKA100:" (LNM$PROCESS_TABLE) I "SYS$DISK" [exec] = "QUOLL$DKA0:" [concealed,terminal] (LNM$SYSTEM_TABLE)   P (Note that the translation of the exec logical in the system table is different Q from what I originally listed -- as quoted by Richard above.  DKA0 is the system  A disk and DKA100 is my login device, and where I was at the time.)   P If I do $show logical/process/full sys$disk, I get ONLY the super logical.  The P exec one in the process table does not appear seemingly when I name the logical.  K This also applies to logical sys$output, show logical/process/full shows a  P [super,confine] and an [exec], but once I name the logical I get only the super  one listed.   D This is on Alpha VMS 7.2 and also applies on my VAX VMS 6.2 machine.  Q Further experimenting shows that the super sys$disk takes the value of whichever  M disk I am on at the time, and the (hidden) exec logical in the process table  J stays as my login device.  The exec logical in the system table (which my : process cannot alter anyway) is the same as sys$sysdevice.   Regards, Paddy   Paddy O'Brien, Transmission Development, 
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australia    Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people, ; but that little whizz-bang apostrophe gives me little spam.    ------------------------------   Date: 8 Oct 2000 16:43:44 -0400 / From: jordan@lisa.gemair.com (Jordan Henderson) @ Subject: Re: Get Viagra at the lowest prices available anywhere!* Message-ID: <8rqma0$a58$1@lisa.gemair.com>  3 In article <200010081500.KAA62106@nadal.loyno.edu>, #  <cyberpharmacy1@excite.com> wrote: ^ >Get Viagra at the absolute lowest prices available.Make your sex life unlike it has ever been >before! >  >GO TO LINK BELOW!! / >http://home.earthlink.net/~gab4ever/bulla1.htm  >  >   F Wrong newsgroup.  You must be looking for a Eunuch's Operating System  discussion forum.    -Jordan Henderson  jordan@greenapple.com    ------------------------------  % Date: Sun, 08 Oct 2000 17:29:37 -0400 # From: John Vottero <John@MVPSI.com> @ Subject: RE: Get Viagra at the lowest prices available anywhere!D Message-ID: <C15945A9D9EFCF11BA8B08002BBF1CCC0CD835@berry.mvpsi.com>   > -----Original Message-----> > From: jordan@lisa.gemair.com [mailto:jordan@lisa.gemair.com]5 > In article <200010081500.KAA62106@nadal.loyno.edu>, % >  <cyberpharmacy1@excite.com> wrote: @ > >Get Viagra at the absolute lowest prices available.Make your " > sex life unlike it has ever been
 > >before! > >  > >GO TO LINK BELOW!! 1 > >http://home.earthlink.net/~gab4ever/bulla1.htm  > >  > >  > H > Wrong newsgroup.  You must be looking for a Eunuch's Operating System  > discussion forum.  >   K Oh, I don't know about that.  People that use VMS are interested in keeping  it up as long as possible! :)    ------------------------------  % Date: Sun, 08 Oct 2000 14:24:59 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> 0 Subject: Re: Getting Compaq to advertise OpenVMS, Message-ID: <39E0BBF7.1B009930@videotron.ca>   Veli K=F6rkk=F6 wrote: > =   & > But FMS is available on Alpha/VMS. =    J You can thank the ALL-IN-1 group for that. As a matter of fact, I wouldn'= t beE surprised if they took over maintenance of FMS. Remember that FMS was J "temporarily" killed during the Palmer years because they didn't realise = thatJ it was still being used by ALL-IN-1. All-in-1 had been given the green li= ght J to port to ALpha, but FMS had been directed to the great big software cem= etary.  J Digital has always had a very bad handling of their messaging products si= nce J the 90s. The whole mess of the ALL-IN-1-MAIL (which was a TOTALLY differe= ntJ product from ALL-IN-1), the whole mess of not offering an upgrade package=  from J message router to Maibus 400, and then not porting Message Router to Alph= a etc.  J But the biggest joke was the ALL-IN-1 group announcing the porting of its=  J office server to dec-unix and NT to make a multiplatform robust mail back= boneH product. Of course, Palmer didn't take notice of this, and a week later,J announces yet another blow job to Bill Gates where Digital will abandon i= tsJ messaging products and rollout exchange everywhere. It took about 2 weeks=  forJ Digital to realise the conflict and ordering the ALL-IN-1 group to cancel=  H their plans to port ALL-IN-1 to NT since it would have competed directly against Billy Gates.  J Had ALL-IN-1 been allowed to port its server to unix and NT, Digital woul= d J have gained a very credible messaging backbone that ran on multiple platf= orms.  But alas, that didn't happen.   J As a matter of fact, there are so many missed opportunities in the story = of VMS...   ------------------------------  % Date: Sun, 08 Oct 2000 13:21:38 -0500 / From: Chris Scheers <chris@applied-synergy.com> 0 Subject: Re: Getting Compaq to advertise OpenVMSO Message-ID: <91C246ECC3498628.A89C124058F93E44.9FAEBCC0546346C0@lp.airnews.net>    Jordan Henderson wrote:  > A > I would imagine that a lot of OpenVMS/VAX customers _prefer_ to > > avoid enhancements that they possibly don't need in order to@ > maintain stability in their present environment.  After all, aA > customer who cannot upgrade to Alpha is probably in a situation . > where stability is valued over enhancements.  G But the sites where stability is paramount do NOT upgrade VMS.  That is E why there are still so many VMS 5.5-2 and 6.2 sites out there.  Those  are nice stable releases.   > The sites that do upgrade VMS are looking for new features and
 capabilities.   G It appears that Compaq is taking the money for VAX support and using it E to fund Alpha development.  I think that Compaq should just be honest H and say that it isn't going to upgrade VAX/VMS anymore.  Of course, that1 isn't the way that business really works.  <sigh>   G ----------------------------------------------------------------------- $ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com     Fax: 817-237-3074    ------------------------------  % Date: Sun, 08 Oct 2000 15:20:57 -0400 , From: Howard S Shubs <hshubs@mindspring.com>0 Subject: Re: Getting Compaq to advertise OpenVMS> Message-ID: <hshubs-39342B.15205708102000@news.mindspring.com>   In article  @ <91C246ECC3498628.A89C124058F93E44.9FAEBCC0546346C0@lp.airnews.n5 et>, Chris Scheers <chris@applied-synergy.com> wrote:   $ >I think that Compaq should just be  >honest 8 >and say that it isn't going to upgrade VAX/VMS anymore.  ; Fine by me, as long as they continue Alpha VMS development.  --   Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------  % Date: Sun, 08 Oct 2000 16:44:34 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>0 Subject: Re: Getting Compaq to advertise OpenVMS7 Message-ID: <200010081644_MC2-B62E-CEEB@compuserve.com>a  J         Some sites avoid upgrading VMS because of the expense which can b= ek extreme!  E         I used to work for McGraw-Hill.  We upgraded a three node VAXtJ Cluster from VMS V5.5-2 to VMS V6.2 about three years ago.  In addition t= o H VMS, the cluster ran Rdb, CDD, ACMS, FMS, and about thirty other layeredJ products.  It was mission critical!  We did it because we needed some new=   features in Rdb.  J         We put over 3,000 man hours into the project.  The actual upgrade=  J (of two system disks) took just a few of those hours.  What took the time=  J and cost big bucks was building a test cluster, doing baseline testing of=  J the application, upgrading the test cluster, running all the tests again,=  J tuning Rdb to handle problem transactions better, rerunning the tests yet=  = again after all the tuning and fixing was complete, etc, etc.o  G         When we finally upgraded the production cluster, everything ran J perfectly.  If it had not, we could have put our division out of business= ! =   J It not only ran perfectly, it ran faster; mostly because Rdb 7.1 was fast= er- than Rdb 4.1 (after a lot of careful tuning).0  J         In my current job we are upgrading VMS 6.2-1H3, Oracle 7.1.3, and=  J Gembase 5.6 to VMS 7.2-1, Oracle 8.0.5, and Gembase 6.0.4.   We've been a= teJ it since May, and don't expect to be ready to go live tile next year some=  C time.  We needed to support new and faster hardware and also get totG supported and supportable releases of VMS, Oracle, etc.  I'd guestimate 7 that the bill for labor will run well into six figures.n  J         Once again, it's mission critical and everything must run perfect= ly' on the Monday morning after we go live!   J         Yes, there are situatations where a VMS upgrade can be done in tw= o.J hours whenever you feel like doing it.  Such situatations are not typical=   in the business world!  % Message text written by Chris Scheers:H >But the sites where stability is paramount do NOT upgrade VMS.  That isE why there are still so many VMS 5.5-2 and 6.2 sites out there.  Thosee are nice stable releases.a <t   ------------------------------  % Date: Sun, 08 Oct 2000 17:47:38 -0500 / From: Chris Scheers <chris@applied-synergy.com>u0 Subject: Re: Getting Compaq to advertise OpenVMSO Message-ID: <E1BD45B869E04045.C39C066BD86CBF5D.663B124EC03ECAB4@lp.airnews.net>g   Chris Scheers wrote: > I > It appears that Compaq is taking the money for VAX support and using it G > to fund Alpha development.  I think that Compaq should just be honest-J > and say that it isn't going to upgrade VAX/VMS anymore.  Of course, that3 > isn't the way that business really works.  <sigh>   F OTOH, it may be that there's so little VAX support money that the onlyB support that can be justified is "allowing" the VAX to remain in a" cluster with the latest Alpha VMS.  G -----------------------------------------------------------------------a$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com o   Fax: 817-237-3074e   ------------------------------  % Date: Sun, 08 Oct 2000 19:09:48 -0400u- From: JF Mezei <jfmezei.spamnot@videotron.ca>i0 Subject: Re: Getting Compaq to advertise OpenVMS, Message-ID: <39E0FEB4.CA425EF0@videotron.ca>   Chris Scheers wrote:I > But the sites where stability is paramount do NOT upgrade VMS.  That isdG > why there are still so many VMS 5.5-2 and 6.2 sites out there.  Thosel > are nice stable releases.2  K I somewhat disagree. My opinion on why there are so many site still left ataN 5.5-2 is that those are the sites that heeded Palmer's call for migration awayL from VMS and have left whatever they have not yet ported running on whatever> version of VMS they had and puit it in "low maintenance" mode.  F If they spent any money on their VMS platform, they would be upgradingN applications to keep up with the times and in doing so, woudl probably need to5 upgrade VMS to be able to upgrade their applications.t  M To me, a 5.5-2 site is one that has abandonned VMS, but also one where CompaqnL could have the most success in regaining as a customer and one that would beJ fairly easy to convince to give VMS a second look. If they still have someT apps on VMS it means that they have apps that don't migrate to NT or UNIX very well.   ------------------------------  # Date: Mon, 09 Oct 2000 02:17:04 GMTi4 From: "Terry C. Shannon" <terryshannon@mediaone.net>0 Subject: Re: Getting Compaq to advertise OpenVMS; Message-ID: <AU9E5.41732$tn.783128@typhoon.ne.mediaone.net>   0 "Brian Tillman" <tillman_brian> wrote in message news:39dde15d@news.si.com...E > >BTW - when you say you can't upgrade from VAX to Alpha, there's noh* > >TECHNICAL reason this needs to be true. > K > False.  The TECHNICAL reason is that the applications we use DON'T RUN ON K > THE ALPHA and the vendors WON'T PORT.  That's as about as technical as itn > gets  K The Charon-VAX folks now have a VAX/VMS emulator that runs on Alpha systemsgH so there is in fact a way to run unmodified VAX apps when you migrate toH Alpha. Dunno if the emulator is commercially available yet; I managed to scarf a copy at CETS2000.    ------------------------------   Date: 8 Oct 2000 22:40:27 -0400-/ From: jordan@lisa.gemair.com (Jordan Henderson)00 Subject: Re: Getting Compaq to advertise OpenVMS* Message-ID: <8rrb6r$ptu$1@lisa.gemair.com>  O In article <91C246ECC3498628.A89C124058F93E44.9FAEBCC0546346C0@lp.airnews.net>,f1 Chris Scheers  <chris@applied-synergy.com> wrote:e >Jordan Henderson wrote: >> fB >> I would imagine that a lot of OpenVMS/VAX customers _prefer_ to? >> avoid enhancements that they possibly don't need in order toeA >> maintain stability in their present environment.  After all, awB >> customer who cannot upgrade to Alpha is probably in a situation/ >> where stability is valued over enhancements.d >eH >But the sites where stability is paramount do NOT upgrade VMS.  That isF >why there are still so many VMS 5.5-2 and 6.2 sites out there.  Those >are nice stable releases. >b? >The sites that do upgrade VMS are looking for new features andi >capabilities. >o  E And the overwhelming majority of THOSE sites have moved or are moving- to Alpha.  -  I The primary reason for not migrating to Alpha given here was being stuck pH with third party code that won't run on Alpha.  If you are so dependent H on such vendor code, then you probably won't upgrade your OS either for B fear of breaking that code.  In this case, you don't need or want 
 enhancements.d  H >It appears that Compaq is taking the money for VAX support and using itF >to fund Alpha development.  I think that Compaq should just be honestI >and say that it isn't going to upgrade VAX/VMS anymore.  Of course, that 2 >isn't the way that business really works.  <sigh> >m  I They are updating OpenVMS/VAX, just not as aggressively as OpenVMS/Alpha.eH Another thing to consider is that it's increasingly expensive for CompaqI to support OpenVMS/VAX in that it has to be tested on a lot of antiquated'G configurations that are increasingly rare in the real-world.  So, a lotsG of that OpenVMS/VAX support money goes to maintaining a lot of test andsF support hardware configurations.  Compaq also has to maintain a lot of? expertise on old VAX hardware, both in development and support.D  H >-----------------------------------------------------------------------% >Chris Scheers, Applied Synergy, Inc.e >uD >Voice: 817-237-3360            Internet: chris@applied-synergy.com  >  Fax: 817-237-3074   -Jordan Henderson  jordan@greenapple.comL   ------------------------------  # Date: Mon, 09 Oct 2000 05:05:52 GMTt3 From: "MrSignor" <MrSignor@nospam_bellatlantic.net>e# Subject: IO$M_NOW and IO$M_NORSWAITe7 Message-ID: <QmcE5.9842$uO1.664510@typhoon1.ba-dsg.net>-   hi,-  0 In respect of sockets UCX/Multinet + TcpWare ...  3 Assuming you agree with the following statement ...p  9 IO$M_NORSWAIT seems to be a superset of IO$M_NOW, because-C IO$M_NORSWAIT is indicating "dont suspend" me if i run out of quota'@  (any quota, ast's, dio,bio etc), whereas IO$M_NOW would seem to6 indicate "dont suspend" me if there is NOT enough room$ in the socket to complete the write.  D Q1 When would you WANT to specify both IO$M_NORSWAIT + IO$M_NOW whenI       performing an asynchronous WRITE to any of the above socket devicesr  H Q2 Do all of the above IP stacks support both IO$M_NORSWAIT + IO$M_NOW ?   Thanks inadvance   -Fred1   ------------------------------  # Date: Sun, 08 Oct 2000 22:48:18 GMTf4 From: "Terry C. Shannon" <terryshannon@mediaone.net>O Subject: Re: Musical Chairmen in Houston... Rosen Steps Down, Capellas Steps Up = Message-ID: <SQ6E5.35606$pu4.3996441@typhoon.ne.mediaone.net>s  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:39D4F47F.F0EF92B5@videotron.ca...   >h1 > I have only seen 2 initiatives taken by Compaq:  >hE > -merging of the Compaq and Digital disk storage system into the news > Storageworks stuff8 > -Announcing that Tandem will eventually move to Alpha. >eJ > The Galaxy and Wildfire were, as far as I recall, projects that had been* > started prior to Compaq taking over DEC. >s  H Yep. But I would add the Zero Latency Enterprise (ZLE) initiative to the above.   ------------------------------  " Date: Sun, 8 Oct 2000 22:57:52 GMT( From: Terry Kennedy <terry@gate.tmk.com>: Subject: Re: Seeking info/prices for OpenVMS and hardware.' Message-ID: <G24vsG.EJ7@spcuna.spc.edu>   4 David Mathog <mathog@seqaxp.bio.caltech.edu> writes:K > You MUST buy SCSI - the EIDE disk is really slow.  Have your reseller get J > you a third party disk and have them buy the intraserver controller fromN > the manufacturer.  (If you buy the controller and disk from the Q you'll pay > 3X more for the same thing.)  I   The 3X-KZPCA-AA Ultra2/LVD controller from Compaq lists for $379. Given K historical DEC pricing, that isn't obscene. I don't know which board it is,e1 as they haven't managed to ship my order yet 8-(.s  4         Terry Kennedy             http://www.tmk.com5         terry@tmk.com             Jersey City, NJ USAw   ------------------------------  # Date: Sun, 08 Oct 2000 18:45:39 GMTe+ From: rjordan@mars.mcs.net (Richard Jordan)T- Subject: Re: Why has Compaq retired DECnet ?? 2 Message-ID: <nh3E5.265$VB1.34770@news.goodnet.com>  L >>  There is useful info on the DNPG site, but not much for older equipment,D >> or for items that Cabletron/Enterasys decided to keep (Roamabout)   > You mean RouteAbout?  I No; I believe that Routeabout info is available on the DNPG site.  I meansH Roamabout; the Digital version was pre-IEEE802.11, but _apparently_ withH late enough firmware, the Digital-made access points can use the currentG 802.11 cards (but the samples imply that more memory is required, and InD can't find info on that).  However, DNPG does not have the RoamaboutH info, Enterasys (one of the Cabletron by-products) does, and they _only_D list info for new firmware which does not support the DEC pre-802.11B cards.  That firmware was on the networks.digital.com site, as wasE the firmware for the actual PC cards (and DEC NICs and a lot of othert stuff).  All apparently gone.r   Rich Jordan  rjordan@mcs.nets   ------------------------------  % Date: Sun, 08 Oct 2000 12:13:04 -0700r9 From: Clemens Wermelskirchen <wermelsk@SLAC.Stanford.EDU>-- Subject: Re: Why has Compaq retired DECnet ??06 Message-ID: <39E0C740.6555DE29@ssrl.slac.stanford.edu>  G From what I heard from DNPG@CETS.LA, the wireless stuff did not make it-% to DNPG, it will stay with Cabletron.D  J I was surprised to see that Compaq has their own new wireless access point% and cards, all IEEE802.11 compatible.    Clemens Wermelskirchen   Richard Jordan wrote:>  N > >>  There is useful info on the DNPG site, but not much for older equipment,F > >> or for items that Cabletron/Enterasys decided to keep (Roamabout) >  > > You mean RouteAbout? > K > No; I believe that Routeabout info is available on the DNPG site.  I meaneJ > Roamabout; the Digital version was pre-IEEE802.11, but _apparently_ withJ > late enough firmware, the Digital-made access points can use the currentI > 802.11 cards (but the samples imply that more memory is required, and I-F > can't find info on that).  However, DNPG does not have the RoamaboutJ > info, Enterasys (one of the Cabletron by-products) does, and they _only_F > list info for new firmware which does not support the DEC pre-802.11D > cards.  That firmware was on the networks.digital.com site, as wasG > the firmware for the actual PC cards (and DEC NICs and a lot of other  > stuff).  All apparently gone.F > 
 > Rich JordanM > rjordan@mcs.nety   ------------------------------  % Date: Sun, 08 Oct 2000 16:53:45 -0400s" From: Dan Sugalski <dan@sidhe.org>5 Subject: Re: Why is it so hard to port things to VMS?a; Message-ID: <5.0.0.25.0.20001008165006.02474c60@24.8.96.48>t  4 At 10:56 AM 10/8/00 -0400, Richard B. Gilbert wrote:K >         DEC Fortran allows you to turn on array bounds checking which canL: >catch a lot of problems in the bud.  C has no equivalent.  D That's not true anymore. Dec C since around 6.0 can do array bounds J checking for you. There's a speed penalty, and it's got some limitations,  but it does work.w  J >         One of the underlying assumptions in C and Unix is that you knowK >what you are doing.  Ninety percent of the time this is a damned dangerouse >assumption to make!!!!c  I It's more the language is willing to sacrifice just about everything for eH that last little bit of speed. It also assumes (not entirely correctly) K that by the time software's rolled out into production that you've already rL found the bugs and don't want or need to pay the extra cost for correctness I checking. History's shown that this isn't a great assumption, but that's e the way these things go.   					Dan  I --------------------------------------"it's like this"------------------- 2 Dan Sugalski                          even samurai? dan@sidhe.org                         have teddy bears and event;                                       teddy bears get drunk-   ------------------------------   Date: 8 Oct 2000 22:10:20 -0400 / From: jordan@lisa.gemair.com (Jordan Henderson)g5 Subject: Re: Why is it so hard to port things to VMS?8* Message-ID: <8rr9ec$oeu$1@lisa.gemair.com>  ; In article <5.0.0.25.0.20001008165006.02474c60@24.8.96.48>,h$ Dan Sugalski  <dan@sidhe.org> wrote:5 >At 10:56 AM 10/8/00 -0400, Richard B. Gilbert wrote: L >>         DEC Fortran allows you to turn on array bounds checking which can; >>catch a lot of problems in the bud.  C has no equivalent.  >uE >That's not true anymore. Dec C since around 6.0 can do array bounds  K >checking for you. There's a speed penalty, and it's got some limitations, l >but it does work. >   G The array bounds checking that DECC does for you isn't terribly useful.dD It won't check the bounds on an array passed into a routine because ? arrays passed into an array _always_ decay into pointers.  Eg.:i  
   $ cc/ver+   Compaq C V6.2-008 on OpenVMS Alpha V7.2-1j)   $ cc/check=bounds/warn=(enable=all) x.c 
   $ link x   $ r xy   $ type x.c   3   void doodle( char buff[50] ) { buff[50] = '\0'; }h      int-   main( int argc, char **argv )-   {-           char buff[50];              doodle( buff );y              return 1;3   }   F You'll note that above the compiler _could_ flag an error in doodle(),E but doesn't.  This is because buff is already a pointer by this point E and the declaration of it as an array of 50 characters is essentiallya ignored.  E The above example is rather contrived, just to show how useless arrayoB bounds checking in typical C implementations is.  More typically, A doodle() would be declared as "void doodle( char *buff )" anyway,oA in which case array bounds checking is challenging, especially ina light of separate compilation.  G In a language like Ada or Pascal, you get real arrays that carry around A with them size information such that real run-time checks can be e@ inserted.  I believe, although I've not verified, that the DECC D arrays bound checking is completely compile time static analysis, no1 run-time checking at all.  Hmm... Let me check...v  )   $ cc/check=bounds/warn=(enable=all) x.c-
   $ link x   $ r xcR   %SYSTEM-F-SUBRNG, arithmetic trap, subscript out of range at PC=000000000002017C   , PS=0000001B 1   %TRACE-F-TRACEBACK, symbolic stack dump followsaL     image    module    routine             line      rel PC           abs PCQ    X  X  main                               884 000000000000017C 000000000002017CrQ    X  X  __main                               0 0000000000000120 0000000000020120=Q                                               0 FFFFFFFFBF3EB3F4 FFFFFFFFBF3EB3F4a   $ type x.c      int    main( int argc, char **argv )t   {h           int  i;e           char buff[50];              i = 50;c              buff[i] = '\0';o              return 1;g   }g  H Nope, I was wrong, it does check at run-time, but not once the array is L passed to a routine (see above), even if that routine declares the array as  having a certain size.  G As I said, not very useful.  Very few array bounds exceptions I see areeE from code so explicitly exceeding an array bound like the above.  Wayo more typical is something like:t  )   $ cc/check=bounds/warn=(enable=all) x.cc
   $ reca links
   $ link x   $ r x4   $ type x.c      #include <string.h>       int    main( int argc, char **argv )6   {3           char buff[50];   *           ( void )memset( buff, 'A', 50 );'           ( void )strcat( buff, "BB" );y              return 1;r   }t   Which produced no error.  I C could possibly pass around more information with pointers to allow realuH aggressive bounds checking on arrays and malloc'd areas, but this would G require that there be versions of all the run-time routines that could aG deal with (or ignore) the extra information.  You'd also probably need nB #pragmas or new keywords to allow you to generate code without theA extra information to interface to libraries without the checking.e  K >>         One of the underlying assumptions in C and Unix is that you know L >>what you are doing.  Ninety percent of the time this is a damned dangerous >>assumption to make!!!! >TJ >It's more the language is willing to sacrifice just about everything for   >that last little bit of speed.   I I think it's something of a combination of the two views expressed above. G Actually, C pointer aliasing inhibits the more aggressive optimizationsmG that compilers can perform to languages like Pascal and FORTRAN.  There6H have been several recommendations for extensions to C to ameliorate thisH problem, like an "alias" keyword.  I don't know what the status of those were in C99.  C C allows the programmer to write pretty efficient programs without 'G optimization, this being a necessity as the more extreme optimizations tC are at odds with very low level programming like device drivers andi Operating System kernels.   I >                               It also assumes (not entirely correctly)  L >that by the time software's rolled out into production that you've already M >found the bugs and don't want or need to pay the extra cost for correctness :J >checking. History's shown that this isn't a great assumption, but that's  >the way these things go.- >-  J Tony Hoare, the inventor of the Quick Sort algorithm (among other things),H related in his Turing Award lecture (I believe, I don't have it handy atG the moment) that the original Algol 60 compilers that were shipped for uI Burroughs machines had full array bounds checking on.  They had intended hJ to add an option to turn this off at a later time, but they couldn't find I a customer who wanted that option.  All the customers preferred that the s= programs abort if there was an errant array bounds exception.h  K This was in the early 60's.  We've known for a long time that array bounds y< checking was a good thing, C has just made that problematic.  	 >					Dani >nJ >--------------------------------------"it's like this"-------------------3 >Dan Sugalski                          even samurait@ >dan@sidhe.org                         have teddy bears and even< >                                      teddy bears get drunk >o >t   -Jordan HendersonS jordan@greenapple.comm   ------------------------------  + Date: Sun, 08 Oct 2000 22:24:15 -0400 (EDT)e/ From: Jordan Henderson <jordan@lisa.gemair.com>o5 Subject: Re: Why is it so hard to port things to VMS?h3 Message-ID: <200010090224.WAA25695@lisa.gemair.com>l   > * > Message text written by Jordan Henderson > M >         Other languages are quite as  portable as C.  A Fortran 77 compileruM > will compile standards compliant Fortran 77 code on any platform.  There is1L > a large standard library that comes with the compiler and/or the O/S.  The% > compiled code should run just fine.W > : >         An ANSI COBOL compiler will compile COBOL. . . . > L >         Ada is not a popular language outside of the Department of DefenseM > and certain shops writing "man rated" applications (somebody's life depends  > on its working correctly).L > Nevertheless, any Ada compiler must comply with the standard and pass someG > pretty tough tests before it can be called Ada.  Such a compiler willp( > compile standards compliant Ada. . . . >   A I didn't mean to suggest that C was more portable because of the E  language definition, it's not.    E >         Now when you have to get down and dirty with the O/S or themI > Graphics Hardware, C looks a lot more attractive.  It's also a lot mores > risky!  J Exactly.  I think I said that the reason C is so portable is that, becauseF it is used so heavily for OS development, quality implementations are M universally available.  Quality implementations are not universally availablef- for Ada, COBOL or any of the various Pascals.h   > K >         DEC Fortran allows you to turn on array bounds checking which canrA > catch a lot of problems in the bud.  C has no equivalent.   TheeJ > WATFOR/WATFIV compilers (is my age showing?)  even had run time checkingI > for uninitialized variables and strict run time checking for the types,VK > modes, etc, of function and subroutine arguments;  if a subroutine took a J > REAL and you passed an INTEGER, you got a run time error!  An enviromentH > like this makes it downright difficult to make some of the most commonL > coding errors!  C has no equivalent nor can I see how such things could beB > built in without turning it into Fortran or Ada or something!!!! > J >         One of the underlying assumptions in C and Unix is that you knowL > what you are doing.  Ninety percent of the time this is a damned dangerous > assumption to make!!!! > J >         The tools can be built to make coding errors damned difficult toK > make but some people feel that the restrictions stiffle their creativity!t >  > <snip>I > >You see. Here's the problem.  People use C for tasks it wasn't created.E > for.  Why?  Because portability is so important, because quality of E > implementation (including standard libraries) is important.  These S0 > factors have outweighed other considerations.< > <snip> >    -Jordan Hendersono jordan@greenapple.com    ------------------------------   End of INFO-VAX 2000.564 ************************