0 INFO-VAX	Wed, 14 Jan 2004	Volume 2004 : Issue 26      Contents: Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: 500.000 AMD64's shipped... Re: Can't boot. Please help.! Re: Could OpenVMS-Athlon ever be? P Re: Faced with flagging PC sales,  HPand others are pushing aggressively into di Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: file and record formats  Re: HP FUDBusting P HP Integrity Superdome server sets world record for 10-terabyte, single-system T- Re: HP will license its intellectual property - Re: HP will license its intellectual property 3 Re: Modify the logical SYS$STARTUP, why or why not? 3 Re: Modify the logical SYS$STARTUP, why or why not? 3 Re: Modify the logical SYS$STARTUP, why or why not? 3 Re: Modify the logical SYS$STARTUP, why or why not?  NOSUCHID with Oracle 9i install # Re: NOSUCHID with Oracle 9i install 3 Re: Personal Workstation 500a/au hardware questions  PowerStorm 300 & DS25 + Re: Problem Porting Unix X Apps To OpenVMS. 1 Re: Problem using bind() function in 'C' program. 1 Re: Problem using bind() function in 'C' program. 1 Re: Problem using bind() function in 'C' program.  Purging NTP log @ Re: The Register: OpenVMS among most-secure of operating systems@ Re: The Register: OpenVMS among most-secure of operating systems@ Re: The Register: OpenVMS among most-secure of operating systems@ Re: The Register: OpenVMS among most-secure of operating systems@ Re: The Register: OpenVMS among most-secure of operating systems@ Re: The Register: OpenVMS among most-secure of operating systems# Re: To all "patriotic" americans... ! Re: VMS runs well on HP Superdome ! Re: VMS runs well on HP Superdome ! Re: VMS runs well on HP Superdome   F ----------------------------------------------------------------------  % Date: Tue, 13 Jan 2004 18:48:04 +0000 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> ' Subject: Re: 500.000 AMD64's shipped... 0 Message-ID: <bu1eh4$1ct$1@new-usenet.uk.sun.com>   David Svensson wrote:  > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bu1359$qra$1@new-usenet.uk.sun.com>...  >  >>David Svensson wrote:  >>v >>>mistdragon@zdnetonebox.com (mist dragon) wrote in message news:<7500353b.0401122133.3e66f991@posting.google.com>... >>>  >>> O >>>>>So they claim to have shipped half a million chips. But shipped to whom ?   >>>>I >>>>Take a walk on local computer shops and malls and look at labels from A >>>>ibm, hp etc. And then take the same walk looking for itanium.  >>>>F >>>>I could ask the same question - where did those 10k itaniums go ?  >>>> >>>>M  >>>  >>> ? >>>As usual when you post. Nothing new and what everyone knows.  >>> I >>>Itanium and the mainstream AMD64 CPU's target different markets, so it G >>>is hardly surprising that there are more AMD64 shipped than Itanium. - >>>Anyone believing something else is stupid.  >>> 	 >>>/David  >>? >>Humm Opteron is a mainstream AMD64 CPU and it very definitely ? >>targets the same market as Itanium. Anyone thinking otherwise  >>is stupid. >>	 >>Regards  >>Andrew Harrison  >  > E > No, Opteron is not a mainstream CPU, Athlon 64 is, and it is pretty E > expected that they have shipped that many Athlon 64's. On the other A > hand, the figures I have seen for Opteron are much lower than I D > expected, considering that they are much easier to sell because of3 > it's x86 compatibility and it's pretty low price.  >   ? What on earth are you talking about Opteron is a mainstream CPU E just as Xeon or Xeon MP, you can use it to build 1-8 way workstations . and servers which is very much the mainstream.  < Would you say that Xeon and Xeon MP arn't mainstream CPU's ?  F > The point about stupid was for for those who might have thought thatA > Itanium would sell better than AMD64. Which I hardly think many  > believed and expected. >   ? Well the problem with that theory such that it is is that AMD64 @ does compete directly with Itanium heard of Deerfield. Its aimedB at the 1-2 way 64 bit server market and 64 bit workstation market.  ? I agree with you in one sense it was stupid for anyone to think > that Itanium would sell better than AMD64, its more expensive,# hotter and slower where it matters.    regards  Andrew Harrison    ------------------------------  # Date: Tue, 13 Jan 2004 19:05:55 GMT 9 From: "Fred Kleinsorge" <my-last-name@stardotzko.dec.com> ' Subject: Re: 500.000 AMD64's shipped... 3 Message-ID: <nkXMb.12208$Ix1.3236@news.cpqcorp.net>   K 498,000 of them are running in x86 32-bit mode.  1,800 are running a 64 bit % windows beta.  200 are running Linux.    1/2 ;-)   ; "mist dragon" <mistdragon@zdnetonebox.com> wrote in message 7 news:7500353b.0401122133.3e66f991@posting.google.com... K > > So they claim to have shipped half a million chips. But shipped to whom  ?  > G > Take a walk on local computer shops and malls and look at labels from ? > ibm, hp etc. And then take the same walk looking for itanium.  > C > I could ask the same question - where did those 10k itaniums go ?  >  > M    ------------------------------  # Date: Tue, 13 Jan 2004 20:19:30 GMT & From: Rick Jones <foo@bar.baz.invalid>' Subject: Re: 500.000 AMD64's shipped... 3 Message-ID: <mpYMb.12226$9E1.7228@news.cpqcorp.net>   P Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote:  F > What on earth are you talking about Opteron is a mainstream CPU justF > as Xeon or Xeon MP, you can use it to build 1-8 way workstations and, > servers which is very much the mainstream.  A While the Opteron 8XX CPU claims support for up to 8 CPUs without A further logic, to date, I do not think anyone has announced 8 CPU 	 systems.    
 rick jones --  ? Process shall set you free from the need for rational thought.  F these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...    ------------------------------  % Date: Tue, 13 Jan 2004 16:44:37 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>' Subject: Re: 500.000 AMD64's shipped... ) Message-ID: <400466C2.56B377D0@istop.com>    David Svensson wrote: H > Itanium and the mainstream AMD64 CPU's target different markets, so itF > is hardly surprising that there are more AMD64 shipped than Itanium., > Anyone believing something else is stupid.  L The problem is that Carly/Curly killed Alpha in order to adopt a commodity ,K industry standard high volume, low price chip that would also allow them to J reduce the number of architectures That was the whole rationale behind the( publicly released excuse for the murder.  M But if IA64 will be restricted to low volume, high cost, niche market, single J source, and only a few customers (HP , SGI), then none of the reasons thatK justified the killing of Alpha to move to IA64 are valid. Alpha was just as J viable as IA64 and already in the market, tested, and years ahead of IA64.   ------------------------------  % Date: Tue, 13 Jan 2004 16:47:50 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>' Subject: Re: 500.000 AMD64's shipped... ) Message-ID: <40046783.8D4B44EC@istop.com>   ( Andrew Harrison SUNUK Consultancy wrote:? > Humm Opteron is a mainstream AMD64 CPU and it very definitely ? > targets the same market as Itanium. Anyone thinking otherwise  > is stupid.  I I disagree. AMD's 64 bit 8086s are aimed at low and mid range servers and I workstations. IA64 is essentially limited to proprietary high end servers 3 running proprietary OS (HP-UX, VMS and Tandem NSK).   N IA64 can forget about Linux and Windows since the market prefers commodity and: industry standard hardware to run those operating systems.   ------------------------------  % Date: Tue, 13 Jan 2004 23:08:12 +0100  From: Dirk Munk <munk@home.nl>' Subject: Re: 500.000 AMD64's shipped... 2 Message-ID: <bu1qtt$69i$1@news4.tilbu1.nb.home.nl>   Fred Kleinsorge wrote:M > 498,000 of them are running in x86 32-bit mode.  1,800 are running a 64 bit ' > windows beta.  200 are running Linux.  > 	 > 1/2 ;-)   Q I suppose a few more are running 32 or 64 bit Linux, and yes the rest is running  P very fast in 32 bit mode. That's one of the selling points Fred, in half a year : or so they are running very fast and cheap in 64 bit mode.  Q Itanium can only run fast (and expensive and hot) in 64 bit mode. In 32 bit mode  M it is a disaster. AMD was clever enough to continue with a basic Intel rule:   *Backwards* *Compatibility*.  Q If you can use the same binaries on a 64 bit home computer, a 64 bit workstation  N and a 64 bit server, then that is a selling point! Or do you really think the O Windows & Linux software manufacturers want to maintain a 32 bit 8086 software  M version (because Intel claims we don't need 64 bit on the desktop), a 64 bit  P 8086 software version (because AMD and many consumers think otherwise) and a 64 P bit Itanium software version? I doubt it, and believe me Intel knows better too.  O When (and in my view not if) the AMD64 will be sold in large quantaties, Intel  O will be forced to step into this market. They certainly will have their 64 bit    8086 ready, they are not stupid.  N The only question is what this means for the Itanium. In my view it will be a M niche processor for high-end servers, and not the industry standard CPU that  J Intel and HP started out to develop. The 64 bit 8086 will be the industry O standard. Not because it is the best cpu, but because it is the cpu the market   wants.  P But then I don't believe the Itanium is the best cpu either, it is the cpu that  the managers wanted.   ------------------------------  # Date: Tue, 13 Jan 2004 22:35:03 GMT   From: CJT <abujlehc@prodigy.net>' Subject: Re: 500.000 AMD64's shipped... * Message-ID: <40047376.6030805@prodigy.net>   JF Mezei wrote:    > David Svensson wrote:  > H >>Itanium and the mainstream AMD64 CPU's target different markets, so itF >>is hardly surprising that there are more AMD64 shipped than Itanium., >>Anyone believing something else is stupid. >  > N > The problem is that Carly/Curly killed Alpha in order to adopt a commodity ,M > industry standard high volume, low price chip that would also allow them to L > reduce the number of architectures That was the whole rationale behind the* > publicly released excuse for the murder. > O > But if IA64 will be restricted to low volume, high cost, niche market, single L > source, and only a few customers (HP , SGI), then none of the reasons thatM > justified the killing of Alpha to move to IA64 are valid. Alpha was just as L > viable as IA64 and already in the market, tested, and years ahead of IA64.  ? Just as in Iraq, there's no way to turn back time even once the 0 rationale for an action is revealed to be false.   --  G After being targeted with gigabytes of trash by the "SWEN" worm, I have F concluded we must conceal our e-mail address.  Our true address is theF mirror image of what you see before the "@" symbol.  It's a shame such( steps are necessary.          ...Charlie   ------------------------------  % Date: Wed, 14 Jan 2004 00:16:20 +0100  From: Dirk Munk <munk@home.nl>' Subject: Re: 500.000 AMD64's shipped... 2 Message-ID: <bu1utl$7ar$1@news4.tilbu1.nb.home.nl>   mist dragon wrote:M >>Ironic that HP, the company that wants so much to enter the digital imaging N >>market, would align itself with a chip vendor which doesn't have a commodity >>64 bit desktops. >  > 7 > Afaik, HP does have AMD64 desktops, but not Opterons.   P Don't know if they are on the market yet, but the specs were (hidden) on the HP  web site some time ago.    >  > N >>Had intel produced a 64 bit 8086 last year, I am not sure AMD would have had >>much chance of survival. >  > C > According rumours, they are already doing it and about to release H > 2005. Of course by that time, they propably have to comply with AMD if  > they want to keep the share :) >   > Even SGI is looking for AMD64. > F > But I doubt that HP will abandon IA64 - they will just keep multiple > 64 bit processors around.   7 They can't even if they wanted to, and for two reasons.   J The first reason is that HP UX and VMS have been ported to IA64, and if I O understand correctly it will be a hell of a job to find another cpu for HP UX.  O Since VMS already was designed to be portable to another cpu, it will be a lot   easier to do that.  P The second reason is that HP has committed itself to sell IA64 cpu's, come what M may. The contract with Intel seems to be so tight, that HP can't do anything  N else then to go on with the IA64 even if it would proof to be almost suicidal.   >  > M    ------------------------------  # Date: Tue, 13 Jan 2004 23:36:50 GMT & From: Rick Jones <foo@bar.baz.invalid>' Subject: Re: 500.000 AMD64's shipped... 2 Message-ID: <mi%Mb.12249$Yi3.126@news.cpqcorp.net>   Dirk Munk <munk@home.nl> wrote: F > Itanium can only run fast (and expensive and hot) in 64 bit mode. In  > 32 bit mode it is a disaster.   F It is a bit of a hair splitting, but that should perhaps be "In 32-bitD x86 mode..." - a hair worth splitting because people take that as inF the "telephone game" and arrive at statements like IPF doesn't supportC 32-bit execution.  It is quite possible to have rather fast running % 32-bit IPF code on HP-UX for example.   
 rick jones --  = denial, anger, bargaining, depression, acceptance, rebirth... C                                      where do you want to be today? F these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...    ------------------------------  % Date: Tue, 13 Jan 2004 19:27:02 -0500 ( From: David Froble <davef@tsoft-inc.com>' Subject: Re: 500.000 AMD64's shipped... , Message-ID: <40048CD6.7050707@tsoft-inc.com>   Dirk Munk wrote:   > mist dragon wrote: > H >>> Ironic that HP, the company that wants so much to enter the digital  >>> imaging G >>> market, would align itself with a chip vendor which doesn't have a  
 >>> commodity  >>> 64 bit desktops. >> >> >>8 >> Afaik, HP does have AMD64 desktops, but not Opterons. >  > K > Don't know if they are on the market yet, but the specs were (hidden) on    > the HP web site some time ago. >  >> >>H >>> Had intel produced a 64 bit 8086 last year, I am not sure AMD would  >>> have had >>> much chance of survival. >> >> >>D >> According rumours, they are already doing it and about to releaseI >> 2005. Of course by that time, they propably have to comply with AMD if ! >> they want to keep the share :)  >>! >> Even SGI is looking for AMD64.  >>G >> But I doubt that HP will abandon IA64 - they will just keep multiple  >> 64 bit processors around. >  > 9 > They can't even if they wanted to, and for two reasons.  > J > The first reason is that HP UX and VMS have been ported to IA64, and if H > I understand correctly it will be a hell of a job to find another cpu K > for HP UX. Since VMS already was designed to be portable to another cpu,  % > it will be a lot easier to do that.     F This is HP's problem, if and only if its Intel that gives up on IA-64.    H > The second reason is that HP has committed itself to sell IA64 cpu's, G > come what may. The contract with Intel seems to be so tight, that HP  F > can't do anything else then to go on with the IA64 even if it would  > proof to be almost suicidal.    N This isn't a problem for HP, as long as Intel continues to develope IA-64 and P sells them (or gives them away) to HP at a reasonable price.  VMS is running on O IA-64, and it will be successful on IA-64.  I never doubted that.  The problem  I for HP, and what I've worried about, is whether there will be IA-64 CPUs  = available to run VMS and whatever other uses HP wants it for.   L The subject line says it all.  If AMD starts eating into Intel's profitable K products, Intel will have to react, and they may do so to the exclusion of  N secondary interests.  At the reported rate IA-64 is selling, I may be kind to " just call it a secondary interest.  O Similar to the exporting of jobs from the US, how long can Intel support their  N expensive 'ego' CPU if they're not pulling in bundles of cash from the 'bread M and butter' desktop market?  I don't care how much money they have available  Q now.  If they start losing money, the well will run dry at some time.  Dry wells    won't continue to produce IA-64.     Dave     --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------    Date: 13 Jan 2004 16:47:21 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) ' Subject: Re: 500.000 AMD64's shipped... = Message-ID: <cf15391e.0401131647.1e60cdc5@posting.google.com>   [ JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<4003A02E.23A91A6F@istop.com>...  > the inferior iA64 thing.  E Do you realize the AMD x86-64 architecture isn't even a true 64 bits? F Yes, it has 64-bit registers, but its virtual address space is only 48F bits. That is actually smaller than Madison's _physical_ address spaceE of 50 bits. (Opteron has a 40-bit physical address space.) IA64 has a ? full 64-bit virtual address space. While x86-64 contains 64-bit F extensions to x86-32, Itanium was designed for 64 bits from the ground up.   2 Integer registers: 16 for x86-64; 128 for Itanium.  < Floating-point registers 8 for x86-64 (it's saddled with the> now-ancient x87 floating-point architecture); 128 for Itanium.  E If you look at the SPECfp2000 results at http://spec.org, all the top ? results, from HP, SGI, Bull, Dell, Supermicro, and ION Computer @ Systems, are with the Itanium 2.  The #1 top result is for an HPA system with Itanium 2 (1.5 Ghz) at 2119. Power4+ (1.7 Ghz) is way C behind at 1699, Pentium 4EE (3.2 Ghz) at 1516, Opteron (2.2 Ghz) at E 1514, EV7 (1.15 Ghz in GS1280) at 1482, and SPARC (1.32 Ghz) at 1350.    ------------------------------  % Date: Tue, 13 Jan 2004 19:55:40 -0500 * From: JF Mezei <jfmezei.spamnot@istop.com>' Subject: Re: 500.000 AMD64's shipped... ) Message-ID: <4004937D.1C2D8E2E@istop.com>    Rick Jones wrote: E > 32-bit execution.  It is quite possible to have rather fast running ' > 32-bit IPF code on HP-UX for example.   K Ok, I am puzzled. How would one generate 32 bit mode IA64 native code ? (or ! why would one do such a thing ?).    ------------------------------   Date: 14 Jan 2004 06:40:31 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>' Subject: Re: 500.000 AMD64's shipped... 5 Message-ID: <DTiotGxQ0bj6-pn2-DGGNdHGKeOnC@localhost>   F On Mon, 12 Jan 2004 21:43:23 UTC, JF Mezei <jfmezei.spamnot@istop.com> wrote:   > mist dragon wrote:- > > http://www.theinquirer.net/?article=13566 J > > Opteron has also outsaled the whole history if itanium processors last
 > > year.. > N > So they claim to have shipped half a million chips. But shipped to whom ? IfN > they were shipped from AMD Taiwan to AMD's wharehouse in the USA, it doesn'tT > automatically mean that there are 500,000  64 bit 8086s running at customer sites. > M > The fab was probably given some order to make 500k chips, so it did and now # > brags about its accomplishements.   F There's one in my dealer's shop, in downtown Munich, just waiting for D me to go and buy. I need to recover from Xmas and my wife's expired ( (Fire GL) graphics card first though :-)   --   Cheers - Dave.   ------------------------------  # Date: Tue, 13 Jan 2004 23:12:09 GMT - From: "Gary Cooper" <gwcooper@mindspring.com> % Subject: Re: Can't boot. Please help. A Message-ID: <dX_Mb.7210$i4.2924@newsread1.news.atl.earthlink.net>   K I had this problem once. It was due to a motherboard rev mis-match with the  BIOS. G It require me making a low level boot floppy from the appropriate Alpha I Firmware CD. I can't remember all the details, however once booted at the G BIOS level, a complete re-setup of everything including the OS type was 	 required.   5 "Alder" <PGDEHMKOKIMD@spammotel.com> wrote in message $ news:r2hMb.8595$wf1.7724@edtnps89...- > AlphaServer AS1000A 4/266 (hobbyist system) 
 > OpenVMS 7.3  > C > I seem to have offended the gods, somewhat.  I rebooted my system D > yesterday and received a blank screen for my efforts.  No hardwareH > detection messages, nothing; only the little blurb about BIOS versionsH > appears before the screen goes black.  This was after several attemptsE > to resolve a strange network problem by rebooting.  Anyone have any 1 > clues on how to revive or diagnose the problem?  >  > Here's what led up to it:  > I > Yesterday, I shut down my system in preparation for an upgrade from VMS I > 7.3 to 7.3.1.  I booted from the newly burned CD-ROM, but it apparently E > lacked a proper boot sector.  I had just used burner software I was J > unfamiliar with, so the results were a little disappointing, but perhaps > not unexpected.  > I > I was surprised, however, that once I rebooted normally from the system H > disk, that I could not connect to this system with TELNET.  I did haveH > an SSH server running, so I then tried that.  Strangely, I was able toE > connect using a normal user account, but connections as SYSTEM were E > refused because of an invalid password (yes, I know the password)!?  > I > I tried several conversational boots, but TELNET remained broken.  This J > despite the fact I could PING the VMS system and could also see from theI > user account, i.e. with "$ TCPIP SHOW SERVICE TELNET/FULL", that Telnet , > was enabled and listening for connections. > 
 > Regards, > Alder  >    ------------------------------   Date: 14 Jan 2004 06:46:05 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>* Subject: Re: Could OpenVMS-Athlon ever be?5 Message-ID: <DTiotGxQ0bj6-pn2-QqS36vfVynG0@localhost>   C On Mon, 12 Jan 2004 14:38:31 UTC, John Reagan <john.reagan@hp.com>   wrote:   > Dave Weatherall wrote: > C > > Thanks again John. I must find the time to read up on GEM. Any  
 > > pointers?  > >  > @ > There was an article in a Digital Technical Journal years ago. > D > http://research.compaq.com/wrl/DECarchives/DTJ/DTJ808/DTJ808SC.TXT   Thanks again John    --   Cheers - Dave.   ------------------------------  # Date: Tue, 13 Jan 2004 19:42:11 GMT " From:   VAXman-  @SendSpamHere.ORGY Subject: Re: Faced with flagging PC sales,  HPand others are pushing aggressively into di 0 Message-ID: <00A2BD6B.F0AB3425@SendSpamHere.ORG>  q In article <nB21ceC0oeiu@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: q >In article <OFC9A649A6.0504AABE-ON85256E1A.00517BDB-85256E1A.0052495F@metso.com>, norm.raphael@metso.com writes:  > I >> A succession of surprise celebrity appearances and a spate of dramatic L >> announcements kept Consumer Electronics Show (CES) attendees on the edgesD >> of their seats during Hewlett-Packard CEO Carly Fiorina's keynoteN >> presentation late last week. Speaking before a packed Hilton Theater in LasI >> Vegas, Fiorina unveiled a new digital music player, revealed an online I >> music alliance with Apple Computer, and weighed in on the online music  >> piracy debate.  > I >   HP should license more technology from Apple.  Like a better GUI than  >   any they're selling now.    K ... and instead of dicking with Itanicum, they should port VMS to the Power K PC.  Then Apple could license it for OS 11 (Oh, the return of the 11) and a K great GUI atop and we'd finally have VMS running on a quality bit of laptop 
 hardware.  --B http://www.legacy-2000.com  for the *best* OpenVMS system securityC                             solutions that others only claim to be.  --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------    Date: 13 Jan 2004 13:27:32 -0600 From: briggs@encompasserve.org$ Subject: Re: file and record formats3 Message-ID: <9ZNzQcf2SLnu@eisner.encompasserve.org>   b In article <200401130958.44076.kpederson@ewu.edu>, Kaleb Pederson <kpederson@mail.ewu.edu> writes:O > By default, when I create a file under OpenVMS it ends up with the following  	 > format:  >  > Directory $1$DGA110:[TST]  > 7 > TESTFILE.TXT;1                File ID:  (16114,631,0) / > Size:            1/137        Owner:    [TST] $ > Created:   13-JAN-2004 09:48:01.86( > Revised:   13-JAN-2004 09:48:01.87 (1) > Expires:   <None specified> ! > Backup:    <No backup recorded>  > Effective: <None specified>  > Recording: <None specified>   > File organization:  Sequential > Shelved state:      Online" > Caching attribute:  WritethroughJ > File attributes:    Allocation: 137, Extend: 0, Global buffer count: 0,  > Version limit: 3H > Record format:      Variable length, maximum 0 bytes, longest 60 bytes6 > Record attributes:  Carriage return carriage control > RMS attributes:     None > Journaling enabled: NoneA > File protection:    System:RWED, Owner:RWED, Group:RWED, World:  > Access Cntrl List:  None > Client attributes:  None >   > Total of 1 file, 1/137 blocks. > M > So, according to the record attributes this file is "Variable Length" with  O > "Carriage return carriage control." Yet, if I dump the file, I never see any  4 > carriage returns (nor line feeds for that matter).  G The "carriage return carriage control" indicates that applications that D wish to display this file should treat each record as a line of text2 followed by an implicit carriage return/line feed.  ? The reason you don't see the CRLF in the file data is that it's 	 implicit.    > This ends up a problem  I > for some of EXEs that we have around here.  For example, the UNIX diff  N > command expects either CRLF or CR but reports completely inaccurate results P > when it sees files of this type (and I am aware that I could use OpenVMS diff  > and edit...).   G How do you use a UNIX diff command on a VMS text file?  If you transfer H the text file to a UNIX box, the issue is how your file transfer program@ treats it.  If you run the diff command on the VMS platform, the? issue is how your file access layer (the CRTL) treats the file.    > Here's a sample dump:  > 5 > Virtual block number 1 (00000001), 512 (0200) bytes  > J >  74 73 65 74 20 61 20 73 69 20 73 69 68 74 00 3C <.this is a test 000000  G 00 3C is the record length (two byte little-endian encoding decimal 60)   J >  63 20 6D 27 49 20 74 61 68 74 20 65 6C 69 66 20  file that I'm c 000010J >  6C 20 74 69 20 64 6E 61 20 67 6E 69 74 61 65 72 reating and it l 000020J >  00 3C 65 6B 69 6C 20 74 73 75 6A 20 73 6B 6F 6F ooks just like<. 000030 >  ^^ ^^J >  61 68 20 74 61 68 74 20 65 6C 69 66 20 79 6E 61 any file that ha 000040J >  61 65 72 63 20 6E 65 65 62 20 72 65 76 65 20 73 s ever been crea 000050J >  74 73 79 73 20 73 69 68 74 20 6E 6F 20 64 65 74 ted on this syst 000060J >  69 64 00 24 74 69 20 73 73 65 6C 6E 75 20 6D 65 em unless it$.di 000070 >        ^^ ^^J >  6B 6F 6F 6C 20 79 6C 6C 61 65 72 20 74 27 6E 64 dn't really look 000080J >  20 65 6C 69 66 20 73 69 68 74 20 65 6B 69 6C 20  like this file  000090J >  00 00 00 00 00 00 00 00 00 00 00 00 FF FF 29 3A :).............. 0000A0 > ! > I guess I have three questions:  > N > 1) I know that I can use an FDL and convert the file to Stream-LF, which is N > what I do now on an as-needed basis.  But, is there a way that I can change L > the default so that all files created are Stream-LF?  Is there any reason  > that I shouldn't?   D No.  There is no OS level default for file format.  Depending on theL application that creates the file there may be an application level default.   > Q > 2) How can I understand this file type?  What are its specs and where are they  J > located?  I tried creating different files, and none of them seem to be M > exactly the same.  They all use the '00' as a line terminator somehow, but  L > they also all seem to have a different preceding character that I haven't  > been able to isolate.   E Each record is preceded with a two byte little-endian byte count.  If 8 the record length is odd, it is padded with a null byte.  D If your Unix diff utility is looking at raw file data then your Unix diff utility is broken.    	John Briggs   ------------------------------  % Date: Tue, 13 Jan 2004 11:52:13 -0800 - From: Kaleb Pederson <kpederson@mail.ewu.edu> $ Subject: Re: file and record formats2 Message-ID: <200401131152.13558.kpederson@ewu.edu>  F > > By default, when I create a file under OpenVMS it ends up with the > > following format:  > B >    This depends completely on the software you use to create it.  O I had tried various OpenVMS editors.  I just tried the OpenVMS VIM port and it  H creates the StreamLF format, which I would expect based on your comment.  L > > 1) I know that I can use an FDL and convert the file to Stream-LF, whichK > > is what I do now on an as-needed basis.  But, is there a way that I can M > > change the default so that all files created are Stream-LF?  Is there any  > > reason that I shouldn't? > B >    No, you can't change this short of getting the source for the! >    application and changing it.   N I thought I heard somebody say something about setting default file type on a 5 filesystem, but I must have been imagining things....   E >    If you take an application that requires some other format, like J >    keyed-indexed, and somehow forced it to be Stream-LF, the application >    would be broken.   O Yeah, I know better than that.  My directory structure would allow me to do it  K on certain directories and be safe, which is what I was really refering to.   M > > 2) How can I understand this file type?  What are its specs and where are M > > they located?  I tried creating different files, and none of them seem to M > > be exactly the same.  They all use the '00' as a line terminator somehow, I > > but they also all seem to have a different preceding character that I ! > > haven't been able to isolate.  > E >    They don't use 00 as a line terminator.  That's just part of the J >    record length for the short records you're writing.  There is a guide? >    to file applications and an RMS manual.  Start with those.   L Ahh.  That makes perfect sense now.  In regards to the files, somebody else L just replied with a link to the RMS reference manual, how about the other.  I I'm still trying to figure out the OpenVMS terminology, help system, and  M documentation locations.  Unfortunately, I haven't been able to find as much  L documentation as I would like (although I've now bookmarked more than a few  pages on the HP site).  ; > > 3) Is there something else that I should be looking at?  > G >    From other stuff you've written I think you're working with broken I >    ports of GNU software.  Most GNU tools should be reading these files H >    via the C library, which sits on top of RMS, which makes the choice$ >    of file attributes transparent.  L I come from a Unix background and prefer the toolset that it provides.  The M following test program confirmed that if accessed via the standard C library   things read correctly:   #include <iostream.h>  #include <fstream.h> int main(void) {         char tmp[1024]; $         ifstream in("eve_file.txt");         if (!in) {7                 cout << "Unable to open file." << endl;                  return 1; 	         }          in.getline(tmp,1024); ,         cout << "line read:" << tmp << endl;         in.close();          return 0;  }   L Apparently I'll just have to change the port if I want it to work... or use  Stream-LF when possible.   Thanks for the info.   --Kaleb    ------------------------------  % Date: Tue, 13 Jan 2004 12:20:36 -0800 - From: Kaleb Pederson <kpederson@mail.ewu.edu> $ Subject: Re: file and record formats2 Message-ID: <200401131220.36588.kpederson@ewu.edu>  ; On Tuesday 13 January 2004 11:27 am, briggs@enc*.org wrote:. [snip]I > > So, according to the record attributes this file is "Variable Length".M > > with "Carriage return carriage control." Yet, if I dump the file, I never > > > see any carriage returns (nor line feeds for that matter). >6I > The "carriage return carriage control" indicates that applications that.F > wish to display this file should treat each record as a line of text4 > followed by an implicit carriage return/line feed. >eA > The reason you don't see the CRLF in the file data is that it'sr > implicit.   M I didn't realize that it was giving me the record length and that those just rN happened to be zero nor did I realize that it was describing how the file was 1 to be output, not the way in which it was stored.   I > How do you use a UNIX diff command on a VMS text file?  If you transferrJ > the text file to a UNIX box, the issue is how your file transfer programB > treats it.  If you run the diff command on the VMS platform, theA > issue is how your file access layer (the CRTL) treats the file.   + I have a port of Unix Diff that I got from:r   http://www.polarhome.com/vim/   N I tried running it on the files named above, and it failed miserably.  I also N have some other programs that are expecting StreamLF files, that I would have N hoped would work with the Variable Length files.  Apparently I'll either have M to modify the port or use an FDL to make it an understandable format for the P misc. programs I'm using.D  M > > 2) How can I understand this file type?  What are its specs and where aresM > > they located?  I tried creating different files, and none of them seem to M > > be exactly the same.  They all use the '00' as a line terminator somehow,eI > > but they also all seem to have a different preceding character that If! > > haven't been able to isolate.T >oG > Each record is preceded with a two byte little-endian byte count.  If : > the record length is odd, it is padded with a null byte. >eF > If your Unix diff utility is looking at raw file data then your Unix > diff utility is broken.t  F I checked the C++ implementation on OpenVMS and it handles everything M correctly, but how would you do that in Standard C?  In my limited knowledge sM of C, there is no such thing as a getline function?  You have to explicitely  M tell it what the end of the line is, or how many bytes to read, or ... don't - you?  I It is definitely broken, even if it is only because the didn't take into  . account all the standard OpenVMS file formats.   --Kalebl   ------------------------------  % Date: Tue, 13 Jan 2004 12:22:38 -0800s- From: Kaleb Pederson <kpederson@mail.ewu.edu>n$ Subject: Re: file and record formats2 Message-ID: <200401131222.38203.kpederson@ewu.edu>  9 On Tuesday 13 January 2004 10:32 am, David M Smith wrote:$M > On Tue, 13 Jan 2004 09:58:44 -0800, Kaleb Pederson <kpederson@mail.ewu.edu>d >o [snip]K > This default format for VMS text files is described, at least, in the RMSo% > reference manual. See, for example:c >m > M > http://h71000.www7.hp.com/doc/731FINAL/4523/4523pro_006.html#rms_record_attl
 >ribute_fieldu  = Thanks for the link.  It would have taken me a while to find.d  F > (unwrap if wrapped). The record attribute you show, sometimes calledE > "implied carriage control" (IIRC) implies that each record is to be K > preceeded by a line feed and followed by a carriage return when output -- I > programs which use RMS for output have it happen automatically. I don't K > know why this format was chosen historically, but that is just the way itA > works.  O I guess it doesn't really matter, although it would be interesting to know.  I eE don't really need to change it for the OpenVMS utilities as they all iK understand the various record formats.  It's the programs ported from Unix o, that don't understand the formats correctly.   --Kaleb    ------------------------------   Date: 13 Jan 2004 20:55:28 GMT, From: bill@gw5.cs.uofs.edu (Bill Gunshannon)$ Subject: Re: file and record formats9 Message-ID: <bu1lvv$cia25$1@ID-135708.news.uni-berlin.de>s  3 In article <9ZNzQcf2SLnu@eisner.encompasserve.org>,'! 	briggs@encompasserve.org writes: d > In article <200401130958.44076.kpederson@ewu.edu>, Kaleb Pederson <kpederson@mail.ewu.edu> writes: >> >> This ends up a problem J >> for some of EXEs that we have around here.  For example, the UNIX diff O >> command expects either CRLF or CR but reports completely inaccurate results mQ >> when it sees files of this type (and I am aware that I could use OpenVMS diff k >> and edit...).  G Sorry, wrong.  Unix text file recordds are seperated by a newline whichtG is the ASCII Line Feed character.  CR is just another character with nom special meaning at all.   F > If your Unix diff utility is looking at raw file data then your Unix > diff utility is broken.e  H No more than your hammer is broken if it doesn't properly put in screws.I All files in Unix are raw.  diff is just the wrong tool if you are trying-  to compare none Unix Text Files.   bill   -- 4J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   a   ------------------------------  % Date: Tue, 13 Jan 2004 13:11:32 -0800n- From: Kaleb Pederson <kpederson@mail.ewu.edu> $ Subject: Re: file and record formats2 Message-ID: <200401131311.32466.kpederson@ewu.edu>  ; On Tuesday 13 January 2004 12:55 pm, Bill Gunshannon wrote:a5 > In article <9ZNzQcf2SLnu@eisner.encompasserve.org>,e >f# > 	briggs@encompasserve.org writes:eF > > In article <200401130958.44076.kpederson@ewu.edu>, Kaleb Pederson   <kpederson@mail.ewu.edu> writes: > >> This ends up a problem-K > >> for some of EXEs that we have around here.  For example, the UNIX diff@H > >> command expects either CRLF or CR but reports completely inaccurateM > >> results when it sees files of this type (and I am aware that I could used > >> OpenVMS diff and edit...).m >rI > Sorry, wrong.  Unix text file recordds are seperated by a newline whichuI > is the ASCII Line Feed character.  CR is just another character with no  > special meaning at all.h  D Yeah, that was a typo on my part.  It should have read CRLF or LF.  H Techinically, it's only looking for the LF but it has the capability of I ignoring whitespace, which therefore ignores the CR (among other things).h   --Kalebr   ------------------------------  + Date: Tue, 13 Jan 2004 16:13:28 -0500 (EST)w+ From: Lord Isildur <isildur@andrew.cmu.edu>i$ Subject: Re: file and record formatsI Message-ID: <Pine.LNX.4.58-035.0401131612460.11143@unix44.andrew.cmu.edu>   F the well-known commands dos2unix and unix2dos are handy for performing' the cr/cr-lf conversion and vice versa.p   Isildura    + On Tue, 13 Jan 2004, Bill Gunshannon wrote:m  5 > In article <9ZNzQcf2SLnu@eisner.encompasserve.org>, # > 	briggs@encompasserve.org writes:sf > > In article <200401130958.44076.kpederson@ewu.edu>, Kaleb Pederson <kpederson@mail.ewu.edu> writes: > >> > >> This ends up a problemoK > >> for some of EXEs that we have around here.  For example, the UNIX diffhP > >> command expects either CRLF or CR but reports completely inaccurate resultsR > >> when it sees files of this type (and I am aware that I could use OpenVMS diff > >> and edit...). >eI > Sorry, wrong.  Unix text file recordds are seperated by a newline which I > is the ASCII Line Feed character.  CR is just another character with no  > special meaning at all.a >AH > > If your Unix diff utility is looking at raw file data then your Unix > > diff utility is broken.e >uJ > No more than your hammer is broken if it doesn't properly put in screws.K > All files in Unix are raw.  diff is just the wrong tool if you are tryingp" > to compare none Unix Text Files. >a > bill >  > --L > Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesF > bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. > University of Scranton   |@ > Scranton, Pennsylvania   |         #include <std.disclaimer.h> >l   ------------------------------    Date: 13 Jan 2004 15:30:36 -0600 From: briggs@encompasserve.org$ Subject: Re: file and record formats3 Message-ID: <WniIZV2ar6I6@eisner.encompasserve.org>d  b In article <200401131222.38203.kpederson@ewu.edu>, Kaleb Pederson <kpederson@mail.ewu.edu> writes:Q > I guess it doesn't really matter, although it would be interesting to know.  I iG > don't really need to change it for the OpenVMS utilities as they all nM > understand the various record formats.  It's the programs ported from Unix i. > that don't understand the formats correctly.  @ While the VMS file model is a poor fit for C's byte stream style5 I/O, it is a good fit for Fortran's record style I/O.3  > And conversely, while the Unix file model is a natural fit for@ C's byte stream style I/O, it is a poor fit for Fortran's record style I/O model.  F It's not that you can't do C on VMS or Fortran on Unix.  But there are5 some compromises that need to be made in either case.I   	John Briggs   ------------------------------    Date: 13 Jan 2004 15:17:15 -0600 From: briggs@encompasserve.org$ Subject: Re: file and record formats3 Message-ID: <BKTDkWJn9tHK@eisner.encompasserve.org>   b In article <200401131220.36588.kpederson@ewu.edu>, Kaleb Pederson <kpederson@mail.ewu.edu> writes:H > I checked the C++ implementation on OpenVMS and it handles everything O > correctly, but how would you do that in Standard C?  In my limited knowledge oO > of C, there is no such thing as a getline function?  You have to explicitely nO > tell it what the end of the line is, or how many bytes to read, or ... don't c > you?  B In Unix, files are streams of bytes.  In VMS, files are streams of? records.  The C library routines tend to present a picture thatf# is consistent with the Unix notion.   B This is one of the central challenges facing the developers of the= C run time library under VMS.  How do you take a file that is0C natively composed of a bunch of records and present it to a programu/ that assumes that it will be a stream of bytes.u  ; The simple answer is that you convert the record boundariess? into line feeds on input.  And you convert line feeds to record-; boundaries on output.  And you tweak this behavior a littlel= for each possible VMS file format so that you'll be doing thes= right thing with both binary data and text files.  And you do E this all in the run time library so that your application programmersr, don't usually need to know or care about it.  K > It is definitely broken, even if it is only because the didn't take into t0 > account all the standard OpenVMS file formats.  I If it had used the VMS C run time library instead of (apparently) rollingr2 its own then this problem might not have occurred.   	John Briggs   ------------------------------    Date: 13 Jan 2004 15:41:02 -0600 From: briggs@encompasserve.org$ Subject: Re: file and record formats3 Message-ID: <EkUjtH96u9K+@eisner.encompasserve.org>s  h In article <bu1lvv$cia25$1@ID-135708.news.uni-berlin.de>, bill@gw5.cs.uofs.edu (Bill Gunshannon) writes:5 > In article <9ZNzQcf2SLnu@eisner.encompasserve.org>, # > 	briggs@encompasserve.org writes: G >> If your Unix diff utility is looking at raw file data then your Unixa >> diff utility is broken. > J > No more than your hammer is broken if it doesn't properly put in screws.K > All files in Unix are raw.  diff is just the wrong tool if you are trying " > to compare none Unix Text Files.  I No.  His diff utility is broken.  It's treating a VMS file as a stream of I bytes.  VMS files are not streams of bytes.  They are streams of records.-  D The Unix diff utility is an interesting example of a record oriented, utility acting on a byte stream data source.  J On VMS, a properly written diff utility should look for record boundaries,E not newline characters.  But if you're going for a simple port, usingiA the C RTL and examining the virtual data stream for newlines is a G reasonable way to proceed.  Examining the raw on-disk data for newlinesv is flat wrong.   	John Briggs   ------------------------------  % Date: Tue, 13 Jan 2004 17:02:53 -0500n* From: JF Mezei <jfmezei.spamnot@istop.com>$ Subject: Re: file and record formats) Message-ID: <40046B09.995C7A44@istop.com>h   Kaleb Pederson wrote:i& >         ifstream in("eve_file.txt"); >         if (!in) {9 >                 cout << "Unable to open file." << endl;p >                 return 1;M >         }w >         in.getline(tmp,1024);v  Q Ok, first of all, this doesn't look like C to me. This looks a lot more like C++.   M If your program sees the "raw" file data, it is because it has used  routinesdL which bypass the "transparent layer" and look at the raw data.  For normal CN program, one has to work very hard to access the data at such a low level, andN normal programs are just fed unix-like data no matter what the reald low levelM file format is because RMS automatically converts the data to the format that J C expects. So no matter what type of line terminator you might have in theA file, when you use the C RTL, you'll see a \n as line terminator.e  H You may wish to obtain the VMS documentation on C++ since it may provideG insights on why the above code acesses the raw data. There may be extra5K arguments to supply to the "open" command that will tell RMS to provide thes default C translation.  K For instance, on VMS, the standard fopen and open routines can be augmentedtN with many arguments descriobing the RMS attributes and type of access you want (for instance, shared access)    ------------------------------  % Date: Tue, 13 Jan 2004 16:05:38 -0800"- From: Kaleb Pederson <kpederson@mail.ewu.edu>h$ Subject: Re: file and record formats2 Message-ID: <200401131605.38759.kpederson@ewu.edu>  4 On Tuesday 13 January 2004 02:02 pm, JF Mezei wrote: [snip]I > Ok, first of all, this doesn't look like C to me. This looks a lot morer > like C++.e  O Yeah, I meant C++.  I was going to do it in C and then realized that it had no aL getline equivalent, so switched to C++ and forgot to change my message back.  F > If your program sees the "raw" file data, it is because it has used J > routines which bypass the "transparent layer" and look at the raw data. L > For normal C program, one has to work very hard to access the data at suchM > a low level, and normal programs are just fed unix-like data no matter what K > the reald low level file format is because RMS automatically converts thetC > data to the format that C expects. So no matter what type of linetM > terminator you might have in the file, when you use the C RTL, you'll see a> > \n as line terminator. >aJ > You may wish to obtain the VMS documentation on C++ since it may provideI > insights on why the above code acesses the raw data. There may be extra M > arguments to supply to the "open" command that will tell RMS to provide the  > default C translation.  O I'm hoping to find some good documentation on all the C/C++ OpenVMS specifics, iN but have had a hard time trying to find anything really useful, although I've 2 only been looking a couple of days now off-and-on.  
 help says:  B             int open(const char *file_spec, int flags,...); (CEC C             Extension)  H        where the ...  represents optional file attribute arguments.  TheG        file attribute arguments are the same as those used in the creat         function.   Looking up creat:d  H             int creat(const char *file_spec, mode_t mode,...); (Compaq C             Extension)  J        where the ...  is an optional argument list of character strings of        the following form:  .        "keyword = value",...,"keyword = value"  3        Or in the case of "acc" or "err", this form:s          "keyword"  D        The keyword is an RMS field in the file access block (FAB) orJ        record access block (RAB), and the value is valid for assignment toJ        that field.  Some fields permit you to specify more than one value.:        In these cases, the values are separated by commas.  3 I found the following once I knew what to look for:c  4 http://h71000.www7.hp.com/commercial/c/5763p021.html  F Hopefully I can find more on the OpenVMS specifics.  I'm particularly C interested in some of the ACL functions but didn't see any in help.V  
 Thanks again.t   --Kaleb    ------------------------------  % Date: Tue, 13 Jan 2004 20:21:37 -0500I* From: JF Mezei <jfmezei.spamnot@istop.com>$ Subject: Re: file and record formats) Message-ID: <40049990.F96AE900@istop.com>e   Kaleb Pederson wrote:eP > Yeah, I meant C++.  I was going to do it in C and then realized that it had noN > getline equivalent, so switched to C++ and forgot to change my message back.  G fgets is the c function to read a line. Used in conjunction with fopen.   G I am not familiar with the "instream in("filename");  construct in C++.f  N I suggest you lookup its documentation it may reveal that it bypasses standard1 RMS processing unless you specify something else.i  D >             int open(const char *file_spec, int flags,...); (CEC C >             Extension)    The flags might looks like this:  ; 	fopen(filenam,"w","ctx=rec","shr=get","rfm=var","rat=cr");t  K When you open an existing file, the "rfm=var" would not be needed since the8I system knows what the file is internally. Note that by default, C creates0< streamlf files unless you specify you want VMS native files.  K One generally has to work hard to get the C RTL to bypass the RMS layer and J access the raw data. Seems that you managed to get there without lifting a
 finger :-)   ------------------------------   Date: 14 Jan 2004 02:49:15 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)$ Subject: Re: file and record formats9 Message-ID: <bu2anb$cp7h5$1@ID-135708.news.uni-berlin.de>s  I In article <Pine.LNX.4.58-035.0401131612460.11143@unix44.andrew.cmu.edu>,t. 	Lord Isildur <isildur@andrew.cmu.edu> writes: > H > the well-known commands dos2unix and unix2dos are handy for performing) > the cr/cr-lf conversion and vice versa.o  F True, but apparently the files he is trying to "diff" don't have thoseD characters either.  They are "implied" by the file type and provided) by VMS utilities when the file is output.:   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   2   ------------------------------   Date: 14 Jan 2004 02:57:23 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)$ Subject: Re: file and record formats9 Message-ID: <bu2b6j$cp7h5$2@ID-135708.news.uni-berlin.de>e  3 In article <EkUjtH96u9K+@eisner.encompasserve.org>,p! 	briggs@encompasserve.org writes:Ej > In article <bu1lvv$cia25$1@ID-135708.news.uni-berlin.de>, bill@gw5.cs.uofs.edu (Bill Gunshannon) writes:6 >> In article <9ZNzQcf2SLnu@eisner.encompasserve.org>,$ >> 	briggs@encompasserve.org writes:H >>> If your Unix diff utility is looking at raw file data then your Unix >>> diff utility is broken.a >> 4K >> No more than your hammer is broken if it doesn't properly put in screws.uL >> All files in Unix are raw.  diff is just the wrong tool if you are trying# >> to compare none Unix Text Files.0 > K > No.  His diff utility is broken.  It's treating a VMS file as a stream oftK > bytes.  VMS files are not streams of bytes.  They are streams of records.a  I I agree, his diff is broken but that has nothing to do with the Unix difflJ command which was designed to compare records in a text file in the formatG Unix uses for text files, that means lines of ASCII text termnated with I the newline character.  Anything else is not a text file and not food for  the diff command.a   > F > The Unix diff utility is an interesting example of a record oriented. > utility acting on a byte stream data source.  D All Unix files are byte streams.  The only thing that makes a file aG text file is the character set it contains and (assuming it is supposeds: to be record oriented) the presence of newline characters.   > L > On VMS, a properly written diff utility should look for record boundaries,G > not newline characters.  But if you're going for a simple port, using-C > the C RTL and examining the virtual data stream for newlines is ahI > reasonable way to proceed.  Examining the raw on-disk data for newlines  > is flat wrong.  F True also, but that isn't Unix's fault.  Using diff to try and compareD binary (non-ASCII) files on a Unix box will also provide some ratherE strange results, especially depending on the terminal your using. :-)d    bill   -- DJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   v   ------------------------------  % Date: Tue, 13 Jan 2004 21:38:31 -0600n@ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>$ Subject: Re: file and record formats6 Message-ID: <4004B9B7.9D2935E4@NeOaSrPtAhMlNiOnWk.net>   Kaleb Pederson wrote:  > N > By default, when I create a file under OpenVMS it ends up with the following	 > format:  >  > Directory $1$DGA110:[TST]m > 7 > TESTFILE.TXT;1                File ID:  (16114,631,0)b/ > Size:            1/137        Owner:    [TST]n$ > Created:   13-JAN-2004 09:48:01.86( > Revised:   13-JAN-2004 09:48:01.87 (1) > Expires:   <None specified> ! > Backup:    <No backup recorded>r > Effective: <None specified>  > Recording: <None specified>o  > File organization:  Sequential > Shelved state:      Online" > Caching attribute:  WritethroughI > File attributes:    Allocation: 137, Extend: 0, Global buffer count: 0,I > Version limit: 3H > Record format:      Variable length, maximum 0 bytes, longest 60 bytes6 > Record attributes:  Carriage return carriage control > RMS attributes:     None > Journaling enabled: NoneA > File protection:    System:RWED, Owner:RWED, Group:RWED, World:  > Access Cntrl List:  None > Client attributes:  None >   > Total of 1 file, 1/137 blocks. > L > So, according to the record attributes this file is "Variable Length" withN > "Carriage return carriage control." Yet, if I dump the file, I never see anyL > carriage returns (nor line feeds for that matter).  This ends up a problemH > for some of EXEs that we have around here.  For example, the UNIX diffM > command expects either CRLF or CR but reports completely inaccurate resultstO > when it sees files of this type (and I am aware that I could use OpenVMS diffo > and edit...).A >  > Here's a sample dump:  > 5 > Virtual block number 1 (00000001), 512 (0200) bytesr > J >  74 73 65 74 20 61 20 73 69 20 73 69 68 74 00 3C <.this is a test 000000J >  63 20 6D 27 49 20 74 61 68 74 20 65 6C 69 66 20  file that I'm c 000010J >  6C 20 74 69 20 64 6E 61 20 67 6E 69 74 61 65 72 reating and it l 000020J >  00 3C 65 6B 69 6C 20 74 73 75 6A 20 73 6B 6F 6F ooks just like<. 000030 >  ^^ ^^J >  61 68 20 74 61 68 74 20 65 6C 69 66 20 79 6E 61 any file that ha 000040J >  61 65 72 63 20 6E 65 65 62 20 72 65 76 65 20 73 s ever been crea 000050J >  74 73 79 73 20 73 69 68 74 20 6E 6F 20 64 65 74 ted on this syst 000060J >  69 64 00 24 74 69 20 73 73 65 6C 6E 75 20 6D 65 em unless it$.di 000070 >        ^^ ^^J >  6B 6F 6F 6C 20 79 6C 6C 61 65 72 20 74 27 6E 64 dn't really look 000080J >  20 65 6C 69 66 20 73 69 68 74 20 65 6B 69 6C 20  like this file  000090J >  00 00 00 00 00 00 00 00 00 00 00 00 FF FF 29 3A :).............. 0000A0 > ! > I guess I have three questions:  > M > 1) I know that I can use an FDL and convert the file to Stream-LF, which isoM > what I do now on an as-needed basis.  But, is there a way that I can changeiK > the default so that all files created are Stream-LF?  Is there any reason  > that I shouldn't?a > P > 2) How can I understand this file type?  What are its specs and where are theyI > located?  I tried creating different files, and none of them seem to behH > exactly the same.  They all use the '00' as a line terminator somehow,  D No, they don't use *ANY* line terminators at all, except for Stream,G Stream_CR and Stream_LF which use <CR><LF>, <CR> and <LF> respectively.    > butnK > they also all seem to have a different preceding character that I haven'tM > been able to isolate.E  D What you're seeing is two bytes (word integer, little endian) at theF beginning of the record which indicate the length of each record. It'sD called a "length attribute". Use DUMP/RECORD (See HELP DUMP) and you7 won't see them at all, just the content of each record.o  G The carriage control indicated by the record attributes is generated atk) run time, it is *NOT* stored in the file.t  9 > 3) Is there something else that I should be looking at?o   Take a look at:   ) http://www.djesys.com/vms/freevms/#mentorI1 http://www.djesys.com/vms/freevms/mentor/rms.html    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/s   ------------------------------  % Date: Tue, 13 Jan 2004 23:20:45 -0500d3 From: "Richard B. Gilbert" <rgilbert88@comcast.net>o$ Subject: Re: file and record formats0 Message-ID: <KIudnZH0xcY9XpndRVn-tw@comcast.com>  I "Carriage return carriage control" means that terminal drivers and print oI symbionts will append a carriage return/linefeed to each record when the  H file is typed or printed.  Neither the CR nor the LF is included in the G file; just a flag bit somewhere in the file header that indicates that t< the file has the carriage return carriage control attribute.  E If you need to process a file with Unix tools, Stream LF is a better  H choice.  The alternatives are to convert the files before processing or > to fix the I/O in  the Unix software to read VMS file formats.   Kaleb Pederson wrote:   N >By default, when I create a file under OpenVMS it ends up with the following  >format: >" >Directory $1$DGA110:[TST] >a6 >TESTFILE.TXT;1                File ID:  (16114,631,0). >Size:            1/137        Owner:    [TST]# >Created:   13-JAN-2004 09:48:01.86u' >Revised:   13-JAN-2004 09:48:01.87 (1)h >Expires:   <None specified>  >Backup:    <No backup recorded> >Effective: <None specified> >Recording: <None specified> >File organization:  Sequential  >Shelved state:      Onlinei! >Caching attribute:  WritethroughfI >File attributes:    Allocation: 137, Extend: 0, Global buffer count: 0, t >Version limit: 3uG >Record format:      Variable length, maximum 0 bytes, longest 60 bytesf5 >Record attributes:  Carriage return carriage controlp >RMS attributes:     Noner >Journaling enabled: NoneV@ >File protection:    System:RWED, Owner:RWED, Group:RWED, World: >Access Cntrl List:  Nones >Client attributes:  Nonet >  >Total of 1 file, 1/137 blocks.2 > L >So, according to the record attributes this file is "Variable Length" with N >"Carriage return carriage control." Yet, if I dump the file, I never see any L >carriage returns (nor line feeds for that matter).  This ends up a problem H >for some of EXEs that we have around here.  For example, the UNIX diff M >command expects either CRLF or CR but reports completely inaccurate results 0O >when it sees files of this type (and I am aware that I could use OpenVMS diff   >and edit...). >  >Here's a sample dump: >74 >Virtual block number 1 (00000001), 512 (0200) bytes >6I > 74 73 65 74 20 61 20 73 69 20 73 69 68 74 00 3C <.this is a test 0000002I > 63 20 6D 27 49 20 74 61 68 74 20 65 6C 69 66 20  file that I'm c 000010 I > 6C 20 74 69 20 64 6E 61 20 67 6E 69 74 61 65 72 reating and it l 000020 I > 00 3C 65 6B 69 6C 20 74 73 75 6A 20 73 6B 6F 6F ooks just like<. 000030  > ^^ ^^0I > 61 68 20 74 61 68 74 20 65 6C 69 66 20 79 6E 61 any file that ha 000040 I > 61 65 72 63 20 6E 65 65 62 20 72 65 76 65 20 73 s ever been crea 000050hI > 74 73 79 73 20 73 69 68 74 20 6E 6F 20 64 65 74 ted on this syst 000060,I > 69 64 00 24 74 69 20 73 73 65 6C 6E 75 20 6D 65 em unless it$.di 000070 
 >       ^^ ^^ I > 6B 6F 6F 6C 20 79 6C 6C 61 65 72 20 74 27 6E 64 dn't really look 000080aI > 20 65 6C 69 66 20 73 69 68 74 20 65 6B 69 6C 20  like this file  000090hI > 00 00 00 00 00 00 00 00 00 00 00 00 FF FF 29 3A :).............. 0000A0h >f  >I guess I have three questions: >rM >1) I know that I can use an FDL and convert the file to Stream-LF, which is eM >what I do now on an as-needed basis.  But, is there a way that I can change oK >the default so that all files created are Stream-LF?  Is there any reason   >that I shouldn't? >oP >2) How can I understand this file type?  What are its specs and where are they I >located?  I tried creating different files, and none of them seem to be xL >exactly the same.  They all use the '00' as a line terminator somehow, but K >they also all seem to have a different preceding character that I haven't 0 >been able to isolate. >a8 >3) Is there something else that I should be looking at? >  >Thanks for the help.- >  >--Kaleb >. >  . >d   ------------------------------  % Date: Tue, 13 Jan 2004 20:55:09 -0600s@ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net> Subject: Re: HP FUDBusting6 Message-ID: <4004AF8D.CF1F3B80@NeOaSrPtAhMlNiOnWk.net>   Keith Parris wrote:S >  > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<btub0a$pil$1@new-usenet.uk.sun.com>... ; > > Ohh dear. The Tru64 extensions to HP-UX have apparentlye: > > slipped to 2005 and HP are apparently stopping selling > > Tru64 in 2004. > / > According to http://www.hp.com/go/tru64unix :s2 > "HP is fully committed to the Tru64 UNIX roadmapD > (http://h30097.www3.hp.com/unix/downloads/tru64_unix_roadmap.ppt),G > delivering updates until at least 2006 with standard support at leastf
 > until 2011.   E I know I've seen Sun's opportunistic ads in print. I think I even sawe. one in a Time magazine in my dentist's office.  . ...and hp's response is only available online.  , I wonder which one(s) got the most exposure?   --   David J. Dachterat dba DJE Systemsy http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/a   ------------------------------    Date: 13 Jan 2004 15:49:39 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)dY Subject: HP Integrity Superdome server sets world record for 10-terabyte, single-system Th= Message-ID: <cf15391e.0401131549.5f473517@posting.google.com>y  @ HP Integrity Superdome server sets world record for 10-terabyte, single-system TPC-H benchmark 
 by Jim Lofinko  A What a way to start the year -- on January 5, HP posted the firstr= 10-terabyte benchmark result ever on HP hardware (a single HPe@ Integrity Superdome server with HP Surestore storage). At 49,108E queries per hour (QphH), the Integrity Superdome server's results areoC very close to those of the five-system IBM p690 cluster-at half thebA price. In fact, when you look at it from a performance-per-systemoE perspective, the Integrity Superdome outperformed the IBM p690 serverv= by more than four times. Following are the benchmark details:    HP Integrity Superdome server & HP Surestore Disk System 2100 storage  49,108 QphH 
 $118/QphH  HP-UX 11i v2 operating system  Oracle10g database e 64 CPUs  Available March 25, 2004; This is another proof point, along with the three-terabyte, : single-system TPC-H, TPC-C (over 1 million tpmC), SPECjbb,B SPECJappServer, and LINPACK benchmarks, showing that the IntegrityD Superdome is the top-performing high-end server on the market today.C The Integrity Superdome server provides balanced performance across 6 multiple benchmarks, workloads, and operating systems.  ' The official results can be found here:e5 http://www.tpc.org/tpch/results/tpch_perf_results.asp    ------------------------------  # Date: Tue, 13 Jan 2004 18:47:09 GMT & From: Rick Jones <foo@bar.baz.invalid>6 Subject: Re: HP will license its intellectual property3 Message-ID: <N2XMb.12196$mw1.9459@news.cpqcorp.net>m  " John Smith <a@nonymous.com> wrote:% > Are there any intellectuals at HP? u   Yes.  
 rick jones -- iG oxymoron n, commuter in a gas-guzzling luxury SUV with an American flageF these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...u   ------------------------------  % Date: Wed, 14 Jan 2004 00:24:51 -0500p* From: JF Mezei <jfmezei.spamnot@istop.com>6 Subject: Re: HP will license its intellectual property) Message-ID: <4004D285.C8788981@istop.com>i   Bob Koehler wrote: > > OpenVMS code too ?I >    DEC often said they would, but got no buyers.  Of course, those weree >    in the Palmer days.  L Why buy the code when you could get Palmer, Curly and probably Carly to giveN it to you for free? Look at what Intel and Microsoft got from Palmer. And look at what Intel got from Curly.i  M Maybe Carly has realised that she now has to start charging for the remainingl% IP before it is all been given away ?     F Seriously, I think that Carly's statement is more about declaring "oil4 reserves" than declaring intentions to sell the oil.   ------------------------------    Date: 13 Jan 2004 12:23:22 -08007 From: jones.computer.srv@worldnet.att.net (Daryl Jones) < Subject: Re: Modify the logical SYS$STARTUP, why or why not?= Message-ID: <8a646952.0401131223.5053ff40@posting.google.com>u  h Kilgallen@SpamCop.net (Larry Kilgallen) wrote in message news:<W8r1MdGpivNT@eisner.encompasserve.org>...^ > In article <04011217214316@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:M > > I just came across some changes to the logicals SYS$STARTUP in one of oureF > > remote sites.  I would tend to think that this is not a good idea. > > , > > Anyone else doing the same type changes? > >  > > ANy other comments?  > A > As someone occasionally called upon to rescue sites in trouble,r > I am of mixed emotions.u > / > 	On one hand, it makes problems hard to find.  > , > 	On the other hand, I bill by the hour :-)   Larry,   Excellant response!!!e   Regards, Daryl Jonest   ------------------------------  # Date: Tue, 13 Jan 2004 23:07:10 GMTs- From: "Gary Cooper" <gwcooper@mindspring.com>d< Subject: Re: Modify the logical SYS$STARTUP, why or why not?A Message-ID: <yS_Mb.7207$i4.6566@newsread1.news.atl.earthlink.net>d  9 Stuff lie that will surely come back to haunt the sysmgr.d  5 "John Brandon" <brandon@dalsemi.com> wrote in messageo+ news:04011217214316@dscis6-0.dalsemi.com...uK > I just came across some changes to the logicals SYS$STARTUP in one of ourlD > remote sites.  I would tend to think that this is not a good idea. >y* > Anyone else doing the same type changes? >i > ANy other comments?n >Y >h > " > $ show logical /full sys$startup: >    "SYS$STARTUP" [exec] = "MYSTARTUP" (LNM$SYSTEM_TABLE)' >         = "SYS$SYSROOT:[SYS$STARTUP]"f >         = "SYS$MANAGER"n@ > 1  "MYSTARTUP" [exec] = "DISK2:[SYSCOMMON]" (LNM$SYSTEM_TABLE)E > 1  "SYS$MANAGER" [exec] = "SYS$SYSROOT:[SYSMGR]" (LNM$SYSTEM_TABLE): >3 >0 >s > J*o*h*n B*r*a*n*d*o*np > VMS Systems Administratorc, > firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  % Date: Tue, 13 Jan 2004 19:32:56 -0500,( From: David Froble <davef@tsoft-inc.com>< Subject: Re: Modify the logical SYS$STARTUP, why or why not?, Message-ID: <40048E38.7050000@tsoft-inc.com>   Daryl Jones wrote:  j > Kilgallen@SpamCop.net (Larry Kilgallen) wrote in message news:<W8r1MdGpivNT@eisner.encompasserve.org>... > ^ >>In article <04011217214316@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes: >>L >>>I just came across some changes to the logicals SYS$STARTUP in one of ourE >>>remote sites.  I would tend to think that this is not a good idea.o >>> + >>>Anyone else doing the same type changes?e >>>a >>>ANy other comments? >>>-A >>As someone occasionally called upon to rescue sites in trouble,y >>I am of mixed emotions.  >>/ >>	On one hand, it makes problems hard to find.w    Q Even worse, once you find some type of aberation, you then want to determine why tN it happened rather than just change it back to a more standard configuration. L Might be a reason (didn't say valid) and your changes could break something.     Dave   -- u4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Roadr Vanderbilt, PA  15486v   ------------------------------  % Date: Tue, 13 Jan 2004 21:26:14 -0600.@ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>< Subject: Re: Modify the logical SYS$STARTUP, why or why not?6 Message-ID: <4004B6D6.D6F35704@NeOaSrPtAhMlNiOnWk.net>   John Brandon wrote:E > K > I just came across some changes to the logicals SYS$STARTUP in one of ourtD > remote sites.  I would tend to think that this is not a good idea. > * > Anyone else doing the same type changes? >  > ANy other comments?> > " > $ show logical /full sys$startup: >    "SYS$STARTUP" [exec] = "MYSTARTUP" (LNM$SYSTEM_TABLE)' >         = "SYS$SYSROOT:[SYS$STARTUP]"t >         = "SYS$MANAGER"c@ > 1  "MYSTARTUP" [exec] = "DISK2:[SYSCOMMON]" (LNM$SYSTEM_TABLE)E > 1  "SYS$MANAGER" [exec] = "SYS$SYSROOT:[SYSMGR]" (LNM$SYSTEM_TABLE)n  D Well, since SYS$STARTUP is DEFINEd in STARTUP.COM, any changes to itG would be likely to occur in SYLOGICALS.COM, unless an additional sin isa& committed and STARTUP.COM is modified.  H In the example, SYLOGICALS *IS* executed during a minimal boot. If DISK2B is not MOUNTed or DEFINEd until SYSTARTUP_VMS, the change would beE relatively harmless to an upgrade. Otherwise, there is some potentiall for trouble.   -- > David J. Dachtera- dba DJE Systems- http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/R   ------------------------------   Date: 13 Jan 2004 18:26:40 EST, From: "Paul Dembry" <pade@nospam.trifox.com>( Subject: NOSUCHID with Oracle 9i install0 Message-ID: <bu1urg$4rt@dispatch.concentric.net>  D I am getting an ORA-27125 unable to create shared memory segment andF %SYSTEM-F-NOSUCHID, unknown rights identifier when trying to create anL Oracle database on OpenVMS 7.3-2. Is there any way that I can determine whatL identifier it is looking for? For example, on Unix systems there is truss orL tusc which trace all system calls and show some of the parameters.  Is there3 any similar thing that I can use on VMS to do this?. TIA, Paul   ------------------------------   Date: 14 Jan 2004 00:12:19 EST% From: "Paul Dembry" <pade@trifox.com>r, Subject: Re: NOSUCHID with Oracle 9i install0 Message-ID: <bu2j3j$4rv@dispatch.concentric.net>  0 I overlooked the ORA_SGA identifier.  Works now. Paul   ------------------------------  + Date: Tue, 13 Jan 2004 23:03:24 -0600 (CST), From: sms@antinode.org< Subject: Re: Personal Workstation 500a/au hardware questions) Message-ID: <04011323032483@antinode.org>:  & From: Tom Crabtree <tccrab@sunset.net>  I > ZLXp's do not work in PWS's.  Looks like you'll be in the market for a n > new monitor.  H    I'll say.  Even if I wanted to use a serial console, VMS just crashes( at DECwindows startup with the old card.  H > I've had _nothing but trouble_ with the SIO chipset and the IDE CDRom.E > The best advice I have is to unplug it's ribbon cable and leave it s3 > unplugged.                         [^^^^--- its.]   E    I can believe that, too.  I noticed that the console "show config"e first reports stuff like:        Bus 00  Slot 04: PCI IDE@                                   pqb0.0.1.4.0          PCI EIDEH                                   dqa0.0.0.4.0            52X24X52 CD-RWL                                   dqb0.0.1.4.0           TOSHIBA CD-ROM XM65  ; but after a "show dev", it changes its mind about the "pq":w  @                                   pqa0.0.0.4.0          PCI EIDE  , I take this as a sign of terminal confusion.  B > I have a SCSI CDRom in an external box that I plug in as needed.  G    I can add an internal SCSI CD-ROM easily enough, but faster, cheaperiH IDE CD-RW drives are also more common than SCSI.  (Also, I'd like to useE a wide SCSI hard disk, but wide CD-ROM drives are not selling for $5,oG either, and I hate to waste two ports on the SCSI card for the internall" stuff.  Everything's complicated.)    + From: Lord Isildur <isildur@andrew.cmu.edu>v  + > [...] we solved the issue by buying a fewo5 > pci ide cards and not using the onboard one at all.|  G    Is there an add-in IDE adapter supported (or even tolerated) by VMS?-  B    Thanks for confirming my disappointing conclusions, by the way.  H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode-orgo    Saint Paul  MN  55105-2547e   ------------------------------   Date: 13 Jan 2004 17:45:46 GMT< From: gartmann@non.immunbio.mpg.de.sens (Christoph Gartmann) Subject: PowerStorm 300 & DS250 Message-ID: <bu1asa$46a$1@n.ruf.uni-freiburg.de>   Hello,  K in a brand-new DS25 I installed a PowerStorm 300 graphics card. At the boot K prompt I entered "set console graphics" and then "init" but I still get the L boot prompt on the serial console. Booting into OpenVMS finally produces theL login-screen at the graphics monitor. Is it possible to have the boot-prompt there as well?   Regards,    Christoph Gartmann'   -- oE  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieI  Postfach 1169                 Internet: gartmann@immunbio dot mpg dot de2  D-79011  Freiburg, Germanyg9                http://www.immunbio.mpg.de/home/menue.html    ------------------------------  # Date: Tue, 13 Jan 2004 19:17:32 GMT19 From: "Fred Kleinsorge" <my-last-name@stardotzko.dec.com>m4 Subject: Re: Problem Porting Unix X Apps To OpenVMS.3 Message-ID: <gvXMb.12212$Up1.2240@news.cpqcorp.net>,   Need more information.  K X11 has a pretty broad and general color scheme, and it it's visible to the  X11 application.  H What depth are the displays (8, 16, 24, other)?  What is the visual typeK (PseudoColor, DirectColor, TrueColor)?  What call are you using to allocatee- a pixel?  Are you using the default colormap?   I The most common pixel depths are 8, 16 and 24.  But there are others thatiK are less common - 4, 12, 32.  Depths greater than 8 typically are TrueColortK or DirectColor, depths less than 8 are typically PseudoColor or StaticGrey.cH A PseudoColor type is an index into a colormap table, the colormap tableJ typically has 3 8 bit values (red, green, blue).  Where a TrueColor visualI type the pixel itself encodes the red, green, and blue values *and* wheredI each of the values is located in the pixel... and this *can* be different ' between architectures, cards, and OS's.       ? "Anonymous #314" <anon314@mail.ourservers.net> wrote in message 7 news:bf17a894.0401131041.435d285b@posting.google.com...uF > I'm trying to port some X apps from Unix to OpenVMS and I'm having a > bit of trouble.e >bD > The platform is an OpenVMS Alpha box, OpenVMS v7.3-1 with Compaq CH > v6.5 and Multinet TCP/IP v4.4 (and the latest version of DEC Winodws). >rH > My problem is that the colors for the display are comming out strange,F > not like they are supposed to.  I haven't had to change any of the X > code from the4H > Unix side to get it compile and run and on Unix it runs just file with > proper colors. >EE > At first I thought it was my PC X software, but when I run the same-C > program from a Unix box I get the proper colors, but not from theiE > OpenVMS box.  I have a little test program to display a window of 8@F > colors and thats working and I get proper colors there.  Any X I useF > on the PC I get the same thing, green characters where I should haveA > white, red backgrounds where I should have black ones and green - > blackgrounds where I should have blue ones.o >tH > Any idea as to where I should start looking?  Some suggested that mabyE > I set the wrong ENDIAN stuff, but the program is written for little 2 > endian and that's what the compiler defaults to. >f  > Any help would be apprecaited.   ------------------------------    Date: 13 Jan 2004 15:04:56 -0800$ From: rcbryan@hotmail.com (RC Bryan): Subject: Re: Problem using bind() function in 'C' program.= Message-ID: <fbcf38dc.0401131504.66a21b50@posting.google.com>   ( I have almost exactly the same thing on 
 OpenVMS V7.3 a Compaq TCP/IP V5.1 - ECO 4 Compaq C V6.4-005.  D and things work the same as it says in the text book (Stevens, "UnixF Network Programming")  The code I have lying around that I think I got, from the Stevens web site has the following:  8     if ( (sockfd = socket(AF_INET, SOCK_STREAM, 0)) < 0)5         err_dump("server: can't open stream socket");i2     bzero((char *) &serv_addr, sizeof(serv_addr));(     serv_addr.sin_family      = AF_INET;2     serv_addr.sin_addr.s_addr = htonl(INADDR_ANY);5     serv_addr.sin_port        = htons(SERV_TCP_PORT); 4     if (bind(sockfd, (struct sockaddr *) &serv_addr, sizeof(serv_addr)) < 0)C5         err_dump("server: can't bind local address");      listen(sockfd, 5);
 <<<snip>>>     clilen = sizeof(cli_addr);=     newsockfd = accept(sockfd, (struct sockaddr *) &cli_addr,o	 &clilen);d  F I like the suggestion of checking your byte ordering in the addresses,? etc.  It is also possible my version works since I am using thesD version in the C-Run Time Library.  I suggest that you make sure youE are using the correct version of the routines.  If you were using the  HP TCPIP product, I would do a:'   $tcpip show device m  E to make sure nobody else was using that port.  Another thing to thinkmF about is that ports below 1000 are only available to priviledged usersA (on Unix at least).  Choose a nice high port number that will noteA collide with anybody.  In our office, we commonly use our 4 digitV9 extension numbers on shared machines to avoid collisions.   ? I don't recall ever seeing the "function not implemented" errorsC message.  A TCPIP stack without bind is like a car without wheels. -& There is something else going on here.  B I just checked on an AIX system and 76 is ENOTCONN, "Socket is notB connected."  I like that a lot more.  As I recall, the TCPIP errnoD values come from a different place than other values and you have toB look them up with a Unix mindset (or in the TCPware manual).  With+ ENOTCONN, I would check your byte ordering..  : (Side note, 76 does not appear on HP/UX or Solaris.  76 isF EPROCUNAVAIL "Bad procedure for program" on Tru64, which does not show1 up on the list of error return values from bind.)a   Regards,	 /RC Bryanu    g "Tucker" <glarb.nospam.org@ix.netcom.com> wrote in message news:<NemdnQrr5IDiHJ7dRVn-ug@comcast.com>...  > Hi there.e > = > I have a user who is using this statement in a 'C' program:a > L >  if( (ret=bind(fd, (struct sockaddr *)&saddr, sizeof(struct sockaddr_in))) > <0)u > E > (I belive the program is trying to create a TCP/IP socket, and then  > listen...)J > The program is returning an error (return code -1) and sets errno to 76,N > which translates into "function not implemented".  I am at a loss as to what > might be the cause of this.t >  > Here is the environment: >  > OpenVMS V7.3-1  (alpha)d > C V5.6 > TCPware V5.6-2 > ( > Is there any other information needed? >  > Thanks, in advance,h > Tucker   ------------------------------  % Date: Tue, 13 Jan 2004 16:04:45 -0700 / From: "Tucker" <glarb.nospam.org@ix.netcom.com>i: Subject: Re: Problem using bind() function in 'C' program.0 Message-ID: <a8udnS2HaJEQ5JndRVn-hQ@comcast.com>  H I'm going to ask the programmer to take a look at this.  However, I justJ tried to install CSWS_JAVA on my system, and I noticed the same error textK ("function not implemented") in that log file, when it tries to bind to theaI port it creates.  I'm sure that HPs code is correct for this.  I'm really G confused about this, because it seems to be an error when trying to use F 'bind()', but it seems to me that that function should be implemented!   Thanks in advance for any help.f  I > Just a guess... the members of the sockaddr struct (port, address, etc) H > must be in network byte order (big endian).  However, VMS machines areK > little-endian machines.  Make sure the members of the sockaddr struct arec& > converted using htons() and htonl(). >rL > If this program wasn't ported from a big-endian machine (say, a Sun), this > detail might be overlooked.a >,$ > On Mon, 12 Jan 2004, Tucker wrote:
 > > Hi there.o > >n? > > I have a user who is using this statement in a 'C' program:m > >l? > >  if( (ret=bind(fd, (struct sockaddr *)&saddr, sizeof(structi sockaddr_in))) > > <0)  > >eG > > (I belive the program is trying to create a TCP/IP socket, and thena > > listen...)L > > The program is returning an error (return code -1) and sets errno to 76,K > > which translates into "function not implemented".  I am at a loss as tor what > > might be the cause of this.t > >m > > Here is the environment: > >h > > OpenVMS V7.3-1  (alpha)o
 > > C V5.6 > > TCPware V5.6-2 > > * > > Is there any other information needed? > >a > > Thanks, in advance,s
 > > Tucker   ------------------------------  % Date: Tue, 13 Jan 2004 19:45:44 -0600 6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler>: Subject: Re: Problem using bind() function in 'C' program.T Message-ID: <craigberry-FE3C98.19454413012004@dsl081-159-101.chi1.dsl.speakeasy.net>  0 In article <a8udnS2HaJEQ5JndRVn-hQ@comcast.com>,1  "Tucker" <glarb.nospam.org@ix.netcom.com> wrote:)  J > I'm going to ask the programmer to take a look at this.  However, I justL > tried to install CSWS_JAVA on my system, and I noticed the same error textM > ("function not implemented") in that log file, when it tries to bind to theyK > port it creates.  I'm sure that HPs code is correct for this.  I'm reallyiI > confused about this, because it seems to be an error when trying to useaH > 'bind()', but it seems to me that that function should be implemented!  E You would have a very strangely mangled system indeed if bind() were 2C not available in the CRTL, but it seems possible there could be an R> issue with the installation of your TCP/IP stack, among other D possibilities.  Just be aware that "function not implemented" often E means that a particular operation on a particular object type is not w> supported or depends on an unavailable resource.  It does not C (necessarily) mean that there is a missing routine in the run-time rA library.  It could mean that some dynamically loaded resource is eB unavailable due to any number of reasons, such as quota problems, 8 privilege problems, installation or upgrade errors, etc.   ------------------------------    Date: 13 Jan 2004 11:58:49 -0800% From: tadamsmar@yahoo.com (Tom Adams)  Subject: Purging NTP log= Message-ID: <ea44f5a1.0401131158.34a823fc@posting.google.com>n  B Is there an easy way to keep the NTP log file from growing without limits?-  E I don't see a way to start NTP using SYS$STARTUP:NTP_STARTUP.COM thato sendsj the log file to NL:o  F The only idea I have is to stop and start NTP on a daily basis so thatE it creates a new log file, then purge the old log files.  It that thee best approach?   ------------------------------  % Date: Tue, 13 Jan 2004 12:57:15 -0600o( From: Wayne Sewell <wayne@tachysoft.com>I Subject: Re: The Register: OpenVMS among most-secure of operating systemsg/ Message-ID: <00A2BD5D.3B7A2652.1@tachysoft.com>g  ( >From: icerq4a@spray.se (David Svensson) >X-Newsgroups: comp.os.vmsJ >Subject: Re: The Register: OpenVMS among most-secure of operating systems   >Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bu12iq$qi9$1@new-usenet.uk.sun.com>... >> Bob Koehler wrote:e >> > In article <bu0i0q$kn8$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes:o >> >   >> > sC >> >>Care to speculate on how different the OpenVMS distribution ispL >> >>now to the VAX distribution in the 80's in terms of code, functionality
 >> >>etc etc.  >> > r >> > w& >> >    Better.  Larger.  More secure. >> > n >> >    No, I'm not guessing.  >> > i >> Ohhh yessssss yooou are.x >>   >l! >Is this a serious Sun employee ?r >rD >Give some bad taste of Sun actually, of an otherwise great company. >lG >Shame that my good view of Sun is darkened by this embarrasing person.e >a    O That pretty well sums it up for me.  In general, I don't care about sun one wayeM or the other.  But I have a low opinion of them for allowing/encouraging thisuM harassment.  Most of us are here to get information on how to use vms, not tof4 hear a competitor's views on perceived deficiencies.   Wayne O ===============================================================================nN Wayne Sewell, Tachyon Software Consulting  (281)812-0738   wayne@tachysoft.com; http://www.tachysoft.com/www/tachyon.html and wayne.html   sO ===============================================================================nB Jed Clampett, checking into hotel: "This place got a cement pond?", 	Ellie May: "And do yuh let critters in it?"   ------------------------------  % Date: Tue, 13 Jan 2004 18:37:32 +0000rO From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> I Subject: Re: The Register: OpenVMS among most-secure of operating systemsa0 Message-ID: <bu1dtd$168$1@new-usenet.uk.sun.com>   Bob Koehler wrote: > In article <bu12iq$qi9$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes: >  >  >>Ohhh yessssss yooou are. >  > E >    Care to speculate as to which one of us has access to the source  >    listings and NDA info?I > E >    I have no reason to BS you, Andrew.  When I say I'm not guessing & >    its a fact.  Not a boast, a fact. >   A Well lets just examine the external evidence shall we rather than:/ your rahter laughable secret squirrel response.o  > Even CERT which is a highly unreliable source puts OpenVMS way: ahead of VMS for vunerabilities. There was the DECnet Worm and thats just about it.  > On the other had the current OS has SSH, Bind, and a number of other advisories posted for it.e  < Security through obscurity cuts both ways because it removes6 your ability to prove your point while I at least have collateral to support mine.M  7 So Ohhh Yesss You Arrr untill that is you can prove youp arn'to   Regardsm Andrew Harrison,   ------------------------------  % Date: Tue, 13 Jan 2004 18:41:34 +0000eO From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com>tI Subject: Re: The Register: OpenVMS among most-secure of operating systems.0 Message-ID: <bu1e4u$1a5$1@new-usenet.uk.sun.com>   David Svensson wrote:e > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bu12iq$qi9$1@new-usenet.uk.sun.com>...i >  >>Bob Koehler wrote: >> >>>In article <bu0i0q$kn8$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes:  >>>s >>>m >>>uB >>>>Care to speculate on how different the OpenVMS distribution isK >>>>now to the VAX distribution in the 80's in terms of code, functionality  >>>>etc etc. >>>0 >>>1$ >>>   Better.  Larger.  More secure. >>>0 >>>   No, I'm not guessing.  >>>  >> >>Ohhh yessssss yooou are. >> >  > " > Is this a serious Sun employee ? > E > Give some bad taste of Sun actually, of an otherwise great company.  > H > Shame that my good view of Sun is darkened by this embarrasing person. >  > /David  1 No because a point that Bob cannot prove deserves  to be treated lightly.  0 I was pre-empting the I have seen the sources so4 I know but can't tell you response that Bob supplied  3 Security through obscurity has its downsides as you  and Bob have just found out.   Regards  Andrew Harrison0   ------------------------------  % Date: Tue, 13 Jan 2004 16:26:01 -080000 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>I Subject: Re: The Register: OpenVMS among most-secure of operating systemse% Message-ID: <40041c1a@cpns1.saic.com>L  ( Andrew Harrison SUNUK Consultancy wrote: > Bob Koehler wrote: > J >> In article <bu12iq$qi9$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK > >> Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes: >> >> >>> Ohhh yessssss yooou are. >> >> >>F >>    Care to speculate as to which one of us has access to the source >>    listings and NDA info? >>F >>    I have no reason to BS you, Andrew.  When I say I'm not guessing' >>    its a fact.  Not a boast, a fact.l >> > C > Well lets just examine the external evidence shall we rather thanF1 > your rahter laughable secret squirrel response.  > @ > Even CERT which is a highly unreliable source puts OpenVMS way< > ahead of VMS for vunerabilities. There was the DECnet Worm > and thats just about it. > @ > On the other had the current OS has SSH, Bind, and a number of! > other advisories posted for it.O > > > Security through obscurity cuts both ways because it removes8 > your ability to prove your point while I at least have > collateral to support mine.a > 9 > So Ohhh Yesss You Arrr untill that is you can prove you  > arn't   H But I have proven it.  Unfortunately, for whatever reason, you have not D been able to accept (or perhaps understand) the proof.  It probably $ isn't worth it but I will try again.  G I have SSH on my VMS system.  It is NOT openssh.  Are you aware of any eI advisories against it?  I can state that my VMS system has never been in  2 any sort of security or DoS danger because of SSH.  G I have BIND on my system.  None of the vulnerabilities posted for BIND rF have ever been able to impact my VMS version of BIND and I test every I one.  So, how does the fact that I run BIND impact the level of security l of my VMS system?r  D You have tried to claim the certain patches to VMS layered products E should have been reported to CERT (if CERT was to be used as a valid hA metric) because they addressed security issues.  However, when I iA followed up on the ones you claimed, that claim turned out to be fL specious (the patches were not, in fact, fixing any kind of security issue).  G So, let's get specific.  You say the CERT advisories aren't valid as a aE metric because too many security issues in VMS don't get reported to wF CERT.  Let's start there.  Name some.  I'll follow up and verify your I accuracy.  Remember, it has to be something that impacts the security or   stability of VMS.   H Step two.  Name any opensource product that is now available on VMS for G which an exploit was discovered that could be used to compromise a VMS cH system (or even provide a means to make a DoS attack against it).  I am D aware of one, but that was in a product that no one concerned about A security would ever run.  Remember, the fact that an exploit was rG discovered does NOT necesarily mean that the exploit could actually be n used against a VMS system.  I I'm calling you on your bluff.  You claim the security concerns with VMS  9 tend to get obfuscated.  Trot 'em out and show 'em to us.n  
 Mark Berrymann   ------------------------------  # Date: Tue, 13 Jan 2004 23:35:42 GMToL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")I Subject: Re: The Register: OpenVMS among most-secure of operating systems 6 Message-ID: <00A2BD73.5E1BE74B@SSRL.SLAC.STANFORD.EDU>  g In article <734da31c.0401131018.7bcce8b4@posting.google.com>, icerq4a@spray.se (David Svensson) writes:i >Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bu12iq$qi9$1@new-usenet.uk.sun.com>... >> Bob Koehler wrote:a >> > In article <bu0i0q$kn8$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes:: >> > V >> > bC >> >>Care to speculate on how different the OpenVMS distribution is0L >> >>now to the VAX distribution in the 80's in terms of code, functionality
 >> >>etc etc.C >> > 2 >> >  & >> >    Better.  Larger.  More secure. >> > 0 >> >    No, I'm not guessing.n >> > 0 >> Ohhh yessssss yooou are.0 >> 3 > ! >Is this a serious Sun employee ?0 >3D >Give some bad taste of Sun actually, of an otherwise great company. >tG >Shame that my good view of Sun is darkened by this embarrasing person.b  J While I hold no brief for Andrew, or for Sun, and while I'm fully on boardK the "VMS is more secure" boat, I do think I should point out that Andrew istJ behaving here, very clearly, in the tradition of English Pantomimes, stageO shows which include a certain amount of audience chanting along with charactersl saying    "Oh, yes I do!"  "Oh, no I don't!"  J So this post of his is, in form at least, a bit more playful than it might appear on the surface.   -- Alana --  O ===============================================================================d0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025sO ===============================================================================f   ------------------------------  % Date: Wed, 14 Jan 2004 00:25:25 -0000m4 From: "Chris Casey" <chris.caseyNOSPAM@ntlworld.com>I Subject: Re: The Register: OpenVMS among most-secure of operating systems3@ Message-ID: <_%%Mb.2942$JL4.21343@newsfep4-glfd.server.ntli.net>  K "Andrew Harrison SUNUK Consultancy" <Andrew_No.Harrison_No@nospamn.sun.com>  wrote in message
 > >>I have@ > >>however claimed that OpenVMS isn't as secure as people think8 > >>it is and proved the claim on a number of occasions. > >> > >>RegardsP > >>Andrew Harrisone > >>  B > Your assumption that I have attacked OpenVMS's relative securityB > vs other OS's is incorrect I have not as is your assumption that@ > I have failed to support the points I have made with evidence. >aF > I have attacked and will continue to attack the OpenVMS BS merchantsF > who claim that OpenVMS must be more secure than any other OS becauseH > it gets less CERT advisories posted for it because it is a BS measure.  G > Pehaps you should read the thread a bit more carefully before jumpingl! > in it would certainly help you.t > 	 > Regardsn > Andrew Harrisonw   ANDREW,S  K you are obviously unwilling or unable to understand the English language so J I will not bother to debate the point any further after this note. It is aJ shame that you take such an objectionable stand because on some topics you( actually have some valid points to make.  J I have not assumed anything and have not asked you to defend your positionJ on VMS security vs. any other. I merely asked you to stop attacking othersK on the subject unless you have a valid point to make regarding the positionvK of VMS security vs. any other o/s. You do not appear to have any such point.A to make but merely hide behind a relentless carping on about CERTs advisories.r  I So Andrew put up or shut up. If you think that VMS is not the most secure K operating system then state your case and give evidence to prove it withouttG any reference to CERT as you obviously think that any such reference is  totally invalid.   ------------------------------  % Date: Tue, 13 Jan 2004 21:10:41 -0600e@ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>, Subject: Re: To all "patriotic" americans...6 Message-ID: <4004B331.63915548@NeOaSrPtAhMlNiOnWk.net>   Tony Reed wrote: > 8 > In article <40019B8F.F2DC14A7@NeOaSrPtAhMlNiOnWk.net>,D >  "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net> wrote: > K > :I don't think it was guilt so much as being desirous of leaving a legacyU3 > :other than that of one of great "robber barons".C > H > These old "robber barons" all made their money befoe income taxes were: > introduced. This is more important than one might think.   Explain.   > I blame Henry ; > Ford for Walmart. I blame the Romans for everything else.    I don't get the connection.s  B Henry Ford developed mass-production into a viable business model.  @ Sam Walton developed discounted mass-merchandising into a viable business model.   E Rome developed debauchery, perversion and decadence into an art form./  D Where's the link I'm not seeing? ...or is it as simple as "commodityB products" that are affordable to the masses? (Rome doesn't fit the	 pattern.)H   -- l David J. Dachterat dba DJE Systemss http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/n   ------------------------------    Date: 13 Jan 2004 15:55:08 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)f* Subject: Re: VMS runs well on HP Superdome= Message-ID: <cf15391e.0401131555.2733f529@posting.google.com>o  [ JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<3FFC8CB5.C5798739@istop.com>...4F > The owners of VMS have decided to restrict VMS to "Business CriticalJ > Computing", which is just a buzzword for "sell it those those willing to > pay a premium".  > M > The problem is thet fewer and fewer are willing to pay a premium for stuff.bM > This is visible throughout the economy, from airlines to department stores.e  A The owners of VMS have decided to make it available at lower costoE thanks to sharing hardware platforms with Linux, Windows, and HP-UX. w This makes it more affordable.  D More and more customers realize that Total Cost of Ownership is whatE counts.  VMS provides excellent TCO.  No viruses.  High reliability. a, High availability.  Even Disaster Tolerance.   ------------------------------    Date: 13 Jan 2004 16:29:25 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)-* Subject: Re: VMS runs well on HP Superdome= Message-ID: <cf15391e.0401131629.6f524fd1@posting.google.com>,   Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<btjghu$3l4$1@new-usenet.uk.sun.com>...e7 > Take OpenVMS, your quarterly revenue for Alphaserverst4 > running OpenVMS is ~65 million dollars, with ~1500 > units.  D IDC Quarterly Server Tracker data for Q3CY2003 says $75M in customer4 revenues, which I've corrected you on twice already.  3 > This is an average after discount of 43K per unit 0 > or put another way the average number of CPU's3 > in an AlphaServer running OpenVMS is 2, you arn'ta6 > even in ES45 territory which is ~110K after discount > for the ES45 with 4 CPU's.  B The average is pretty useless as an indicator of _typical_ systemsF sold -- for the average price of $50K per server ($75M/1507) you could? buy an awful lot of DS15s, and not come close to touching a 64p F GS1280.  The majority of those 1,507 estimated units are probably DS15A class uniprocessors, DS25s, and so forth, but with some number ofDB ES45s, ES47s, ES80s and even a few 64-processor GS1280s included. E Since IDC Quarterly Server Tracker data is for initial purchases onlyrD (upgrades are not included)), the majority of SMP systems bought are9 probably not fully populated, leaving room for expansion.c  7 > If it becomes a commodity it will be because Dell andn6 > IBM get behind it and sell large quantities of IA-644 > based servers and if this happens then HP will not4 > be able to charge a premium for the Itanium server% > platform on which OpenVMS will run.P  F Oh, no! You've figured out HP's secret plan -- same HW price for a box2 whether it runs Windows, Linux, HP-UX, or OpenVMS.  @ All the better for OpenVMS customers, and something they've been asking us for for many years.E  0 > HP has a track record of failing to make money4 > competing with in particular Dell in the commodity0 > server and desktop market. Replacing IA32 with2 > IA64 is a technology change not a change in HP's/ > business model and won't in the low end spaceh, > have any impact on HP's ability to compete. > with Dell who already undercut HP with their > IA64 server(s).i  " One could readily re-word this as:  B Sun has a track record of failing to make money competing with, inF particular, Linux-on-Intel from HP, IBM & Dell in the commodity serverE and desktop market.  Replacing SPARC with AMD Opteron is a technologyo? change, not a change in Sun's business model, and won't, in thetD low-end space, have any impact on Sun's ability to compete with IBM,8 HP, and Dell servers who already undercut Sun with their Linux-on-Intel servers.    > OpenVMS currently doesn'tt > scale to more than 16 CPU'so  C Oracle's recent 32-way results with Oracle Rdb on a GS1280 indicater
 otherwise.D OpenVMS had made great strides in SMP scalability in 7.3, 7.3-1, and( now 7.3-2, as many customers can attest.  ' > With OpenVMS's clustering capabilitesr$ > the most obvious platform would be" > clusters of 2-4 way IA64 servers   Which many customers will buy.  " > which if they become a commodity > HP will not make money on. >2$ > So I repeat why on earth would you% > want to sell hardware. You are much7# > more likely to make a return from8
 > selling SW.-  E HP will continue to make (small) amounts of money on the servers, andvC continue to make goodly amounts on storage, OpenVMS itself, layeredMF software products (including that excellent Clustering you mentioned),
 and Services.A  D Sun will continue to sell servers, and continue to send the majorityB of its storage reveunes to Hitachi, the majority of its clusteringB software revenues to Veritas, and continue to miss out on Services pretty much entirely.h   ------------------------------  % Date: Tue, 13 Jan 2004 21:01:38 -0600 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>* Subject: Re: VMS runs well on HP Superdome6 Message-ID: <4004B112.4A201D87@NeOaSrPtAhMlNiOnWk.net>   Keith Parris wrote:A > ] > JF Mezei <jfmezei.spamnot@istop.com> wrote in message news:<3FFC8CB5.C5798739@istop.com>...RH > > The owners of VMS have decided to restrict VMS to "Business CriticalL > > Computing", which is just a buzzword for "sell it those those willing to > > pay a premium".e > >eO > > The problem is thet fewer and fewer are willing to pay a premium for stuff.IO > > This is visible throughout the economy, from airlines to department stores.O > C > The owners of VMS have decided to make it available at lower costrF > thanks to sharing hardware platforms with Linux, Windows, and HP-UX.  > This makes it more affordable. > F > More and more customers realize that Total Cost of Ownership is what
 > counts.   D Not where I "live"! Around here, cost to acquire *VASTLY* out-weighsF TCO. I don't think anyone around here even knows how to calculate TCO.  = > VMS provides excellent TCO.  No viruses.  High reliability.S. > High availability.  Even Disaster Tolerance.  F Well, 5 out of seven isn't bad, but it won't get you "home". You still need:a  ( o Visible (advertised in the mainstream)0 o Affordable (competitive with the alternatives)  A Anyone wanna be a straight-A student? ...or do you settle for the  D-minus (5/7=71%)?   --   David J. DachteraM dba DJE SystemsI http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/d   ------------------------------   End of INFO-VAX 2004.026 ************************