INFO-VAX Sat, 11 Aug 2007 Volume 2007 : Issue 437 Contents: Re: Intel marginalizing Itanium even faster than expected? Re: Intel marginalizing Itanium even faster than expected? Re: Oldest Alpha for upgrade to Integrity Re: Oldest Alpha for upgrade to Integrity Re: Oldest Alpha for upgrade to Integrity Re: Oldest Alpha for upgrade to Integrity Re: Oldest Alpha for upgrade to Integrity Re: Oldest Alpha for upgrade to Integrity Re: OT: iWork 08 is unnacceptably slow on my iBook. Was: Re: X Window Servers Re: OT: iWork 08 is unnacceptably slow on my iBook. Was: Re: X Window Servers Re: Rexx for OpenVMS Re: Rexx for OpenVMS Re: TPU on MAC OS-X ? Re: What VMS does RIGHT! Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion RE: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion RE: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion RE: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: X Window Servers ---------------------------------------------------------------------- Date: Fri, 10 Aug 2007 15:05:01 -0400 From: "David Turner, Island Computers" Subject: Re: Intel marginalizing Itanium even faster than expected? Message-ID: This makes you wonder whether the "investment" in new Itanium systems is any more appropriate than buying new or refurbished Alpha systems. They could be as obsolete as each other in a couple of years... the moral of this story is to stick with what you know.... Alpha ! So who out there at HP is developing VMS for Intel 64? Or is it a swiss company? DT "yyyc186" wrote in message news:1186755185.762202.53090@i38g2000prf.googlegroups.com... >> http://www.intel.com/technology/architecture-silicon/intel64/ >> >> (Conniving bastards take AMD's innovations and call it "Intel 64"! >> That itself is enough to make me buy AMD64 chips.) > > I have bought only AMD chips for PC's for years and I only buy Alphas > for OpenVMS. Titanic was a chip the industry never needed. > ------------------------------ Date: Fri, 10 Aug 2007 12:29:23 -0700 From: Doug Phillips Subject: Re: Intel marginalizing Itanium even faster than expected? Message-ID: <1186774163.254423.164600@q4g2000prc.googlegroups.com> On Aug 10, 8:22 am, Ron Johnson wrote: > Guess what architecture that /Intel 64/ is... > > http://www.intel.com/technology/architecture-silicon/intel64/ > > (Conniving bastards take AMD's innovations and call it "Intel 64"! > That itself is enough to make me buy AMD64 chips.) > It looks like it's Intel architecture. There doesn't seem to be anything there that Intel hasn't at least alluded to before. back in May, here in c.o.v., I noted the following: : The Penryn road map shows 45nm 2007, 32nm 2009 and 22nm 2011. The : scuttlebutt about the Larrabee many-core project seems to point to : IA with x86 So, it's 2007 and they've announced 45nm and more fully described some of the evolutionary tweaks. Where's the news? I'm not a chip-head, but what I read about the Larrabee makes me think it'll be Itanium-type technology that can run x86-64 instructions. Will you be shocked then, too? ------------------------------ Date: 10 Aug 2007 14:06:28 -0500 From: brooks@cuebid.zko.hp.nospam (Rob Brooks) Subject: Re: Oldest Alpha for upgrade to Integrity Message-ID: tadamsmar writes: > Does the AlphaStation 400, AlphaServer 800 and the DS10 support VMS > 8.3 or some version that Integrity also supports? Yes > I was wondering if I can upgrade my older Alphas to a common version > with Integrity. > Is there a web page that shows what hardware will run what versions of > VMS? While there have been a few VAX models that have been dropped along the way, no Alpha hardware has been dropped from support. In this context, I'm talking about models that were properly sold with VMS, not oddball things like the multia, or the early EV3-based systems. I think you are making this task seem much harder than it really is. -- Rob Brooks MSL -- Nashua brooks!cuebid.zko.hp.com ------------------------------ Date: 10 Aug 2007 14:18:15 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: Oldest Alpha for upgrade to Integrity Message-ID: In article , brooks@cuebid.zko.hp.nospam (Rob Brooks) writes: > tadamsmar writes: >> Does the AlphaStation 400, AlphaServer 800 and the DS10 support VMS >> 8.3 or some version that Integrity also supports? > Yes More conventional language would be that VMS V8.3 supports the hardware, rather than the other way around. > While there have been a few VAX models that have been dropped along the way, > no Alpha hardware has been dropped from support. In this context, I'm talking > about models that were properly sold with VMS, not oddball things like the > multia, or the early EV3-based systems. I do not believe any EV3 system ever escaped from the DEC labs. They had a Unix system as a coprocessor to take care of features not present on the experimental EV3 chip. ------------------------------ Date: 10 Aug 2007 15:43:44 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Oldest Alpha for upgrade to Integrity Message-ID: In article <1186767412.536166.201930@d30g2000prg.googlegroups.com>, tadamsmar writes: > Does the AlphaStation 400, AlphaServer 800 and the DS10 support VMS > 8.3 or some version that Integrity also supports? Yep. In fact I'm running 8.3 on my DS10L. > I was wondering if I can upgrade my older Alphas to a common version > with Integrity. > > Is there a web page that shows what hardware will run what versions of > VMS? You need to check the SPD for the version. The FAQ will point you to the SPD. ------------------------------ Date: Fri, 10 Aug 2007 13:54:25 -0700 From: Doug Phillips Subject: Re: Oldest Alpha for upgrade to Integrity Message-ID: <1186779265.538338.67060@e9g2000prf.googlegroups.com> On Aug 10, 3:43 pm, koeh...@eisner.nospam.encompasserve.org (Bob Koehler) wrote: > In article <1186767412.536166.201...@d30g2000prg.googlegroups.com>, tadamsmar writes: > > > Does the AlphaStation 400, AlphaServer 800 and the DS10 support VMS > > 8.3 or some version that Integrity also supports? > > Yep. In fact I'm running 8.3 on my DS10L. > > > I was wondering if I can upgrade my older Alphas to a common version > > with Integrity. > > > Is there a web page that shows what hardware will run what versions of > > VMS? > > You need to check the SPD for the version. The FAQ will point you > to the SPD. Or use this link: Start on page 28. ------------------------------ Date: Fri, 10 Aug 2007 14:27:55 -0700 From: "AlexNOSPAMDaniels@themail.co.uk" Subject: Re: Oldest Alpha for upgrade to Integrity Message-ID: <1186781275.003056.176530@j4g2000prf.googlegroups.com> On Aug 10, 8:06 pm, bro...@cuebid.zko.hp.nospam (Rob Brooks) wrote: > While there have been a few VAX models that have been dropped along the way, > no Alpha hardware has been dropped from support. In this context, I'm talking > about models that were properly sold with VMS, not oddball things like the > multia, or the early EV3-based systems. Hey Rob! That's not quite right, some Alpha's that were sold with VMS have been dropped from support. Specifically the DEC 2000 Models 300/500, and the Tadpole ALPHAbook 1. http://h71000.www7.hp.com/openvms/hw_supportchart.html Support for both was dropped after V7.3-1, with that being the last version. While I can understand the reasons, the way it was handled was very poor, and showed zero thought for the customer. For continued support where people had already upgraded to V7.3-1, the only path would be to then downgrade back to V6.2-1HX , as that was the last version those boxes could run (supported) with Prior Version Support (PVS). http://h71000.www7.hp.com/openvms/openvms_supportchart.html Alex ------------------------------ Date: Sat, 11 Aug 2007 01:58:24 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: Oldest Alpha for upgrade to Integrity Message-ID: "AlexNOSPAMDaniels@themail.co.uk" writes: >That's not quite right, some Alpha's that were sold with VMS have been >dropped >from support. >Specifically the DEC 2000 Models 300/500, and the Tadpole ALPHAbook 1. >http://h71000.www7.hp.com/openvms/hw_supportchart.html >Support for both was dropped after V7.3-1, with that being the last >version. Interesting that some of the VAXes have a maximum supported version of V7.3, others have no maximum version. So that would seem to mean that not all VAXes supported on V7.3 will be supported on a later version, but that would imply that there actually _will be_ a later version. Really? ------------------------------ Date: Fri, 10 Aug 2007 21:36:23 +0100 From: Elliott Roper Subject: Re: OT: iWork 08 is unnacceptably slow on my iBook. Was: Re: X Window Servers Message-ID: <100820072136238160%nospam@yrl.co.uk> In article , < @SendSpamHere.ORG> wrote: > In article , "P. Sture" > writes: > >A couple of days ago I received the offer of a download of a 30 day demo > >of iWork 08. The speculation earlier this year that it would include a > >spreadsheet application was indeed correct. > > I didn't receive such an offer. I'd be happy to check it out to see how > it performs on a faster G4. I don't see any download offer on the Apple > site though. http://www.apple.com/iwork/trial/ It ain't rocket science VAXman. ;-) Goes like *stink* in a Mac Pro. Rather tasty. I spent today getting my grey wall ready to go to the tip. It was sad. I have a few vaxstations and alpha 166 workstations plus untold CRT monitors in various stages of dereliction. About 25 Km SE of Manchester UK. Free to anyone who wants to come and get 'em. Philip Helbig was here yesterday. He got the best stuff. But be quick. My missus wants the Portakabin for her gym. If you can't decode the teco in my sig. Somebody else needs 'em more than you do. ;-) -- To de-mung my e-mail address:- fsnospam$elliott$$ PGP Fingerprint: 1A96 3CF7 637F 896B C810 E199 7E5C A9E4 8E59 E248 ------------------------------ Date: Fri, 10 Aug 2007 21:29:41 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: OT: iWork 08 is unnacceptably slow on my iBook. Was: Re: X Window Servers Message-ID: <9N4vi.39$GQ4.14@newsfe12.lga> In article <100820072136238160%nospam@yrl.co.uk>, Elliott Roper writes: > > >In article , < @SendSpamHere.ORG> wrote: > >> In article , "P. Sture" >> writes: > >> >A couple of days ago I received the offer of a download of a 30 day demo >> >of iWork 08. The speculation earlier this year that it would include a >> >spreadsheet application was indeed correct. >> >> I didn't receive such an offer. I'd be happy to check it out to see how >> it performs on a faster G4. I don't see any download offer on the Apple >> site though. >http://www.apple.com/iwork/trial/ > >It ain't rocket science VAXman. ;-) Well, I went to the /iwork page and did a search for "30 Day". No hit. >Goes like *stink* in a Mac Pro. Rather tasty. ??? >I spent today getting my grey wall ready to go to the tip. >It was sad. >I have a few vaxstations and alpha 166 workstations plus untold CRT >monitors in various stages of dereliction. About 25 Km SE of Manchester >UK. Free to anyone who wants to come and get 'em. >Philip Helbig was here yesterday. He got the best stuff. >But be quick. My missus wants the Portakabin for her gym. > >If you can't decode the teco in my sig. Somebody else needs 'em more >than you do. >;-) I don't even need a machine for that one! -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" http://tmesis.com/drat.html ------------------------------ Date: Fri, 10 Aug 2007 17:38:02 -0300 From: "Shmuel (Seymour J.) Metz" Subject: Re: Rexx for OpenVMS Message-ID: <46bcdaba$1$fuzhry+tra$mr2ice@news.patriot.net> In <8730409.ys5V6uXsJ9@linux1.krischik.com>, on 08/08/2007 at 07:08 PM, krischik@users.sourceforge.net said: >It it worked REXX would have been a cool extension to DCL as Rexx has >this unique feature of passing on unused strings return values to the >command prompt or the editor - or whereever Rexx was embedded into. Actually, it does something even more useful; it passes the value of an expression to the current environment, which might not be constant within the REXX code. I've seen plenty of REXX code that sends commands to two or more different environments within a few lines. I agree that without the ability REXX would be a far weaker language for scripting. -- Shmuel (Seymour J.) Metz, SysProg and JOAT Unsolicited bulk E-mail subject to legal action. I reserve the right to publicly post or ridicule any abusive E-mail. Reply to domain Patriot dot net user shmuel+news to contact me. Do not reply to spamtrap@library.lspace.org ------------------------------ Date: Fri, 10 Aug 2007 17:46:06 -0500 From: "Craig A. Berry" Subject: Re: Rexx for OpenVMS Message-ID: In article <8730409.ys5V6uXsJ9@linux1.krischik.com>, krischik@users.sourceforge.net wrote: > > I got following reply from mark Hessling in private mail > > > > Hi Tom, > > > > I did the port many years ago. I don't have access to a VMS machine > > anymore so can't tell you if it still compiles. But here are the contents > > of the README.VMS that comes with Regina: > > > > This is a preliminary version of Regina for OpenVMS. It has been tested > > on OpenVMS 6.1, 6.2 on VAX and AXP systems. Those are something like 12-to-14-year-old releases of OpenVMS. If someone has old Rexx binaries lying around, they probably still work fine, but this news is too old to be of much value in assessing current build prospects. > > To build Regina, you need MMK. This can be obtained from: > > ftp://ftp.wku.edu/ Well, now you'd most likely be able to find it at: http://vms.process.com/scripts/fileserv/fileserv.com?MMK > > Run the build.com DCL script to compile and link Regina. This is a > > front-end to the MMK description file; descrip.mms. > > > > Things that don't work: > > > > - Stream BIF when the target is a directory. This is due to the C runtime > > funtion stat() not behaving the same as on other operating systems. That's not a highly specific description, but there have been changes in stat() including the fact that it has been possible for some time now to define the use of a POSIX-compliant stat structure at compile time (/define=_USE_STD_STAT). > > - Time('O') does not work. The C runtime function gmtime() does not > > return valid values. ISTR there was significant work done with gmtime() in the v7.x era. We're now well into the v8.x era, so whatever the problem there was has very likely long since been fixed. > > - Uname BIF doesn't return anything. I have no idea what that means or what underlying features it depends on. > > - Execution of external commands crash Regina. Since there are lots of other applications that can execute external commands on OpenVMS, it sounds like something that could be made to work if someone were willing to look into it. -- Posted via a free Usenet account from http://www.teranews.com ------------------------------ Date: Sat, 11 Aug 2007 15:55:36 +1000 From: Jim Duff Subject: Re: TPU on MAC OS-X ? Message-ID: <46bd4f59$1@dnews.tpgi.com.au> Paul Raulerson wrote: > >>> [snip] >> erm, see TPU procedure CALL_USER? It's trivial to implement logical >> name and symbol functions using a user written program. >> >> > _12> >> >> Jim. > > LOL! Whatever else TPU may be- "trivial" is definitely the wrong word. > "Easy" would not work there either --- "arcane" comes to mind. :) > > Rexx and Perl and so forth can be just as arcane though. > > -Paul > > Arcane as opposed to say, lisp? That's what people are forced to program in to extend emacs. And perhaps I should have said "trivial for a programmer that can spend 20 minutes reading the manual". Having never programmed in TPU before today, that's about how long it took me to figure out how to do this. (Although my original statement re: trivial was referring to writing the user routine to perform the logical name translation, the TPU code was even simpler). All up, 130 lines of C code (not including comments) and 27 lines of TPU would be considered trivial for any accomplished programmer. Conversion of the C code to your preferred language is left as an exercise for the student. $create log.c $deck #include #include #include #include #include #include #include #pragma environment save #pragma extern_model globalvalue extern unsigned long int tpu$_success; #pragma environment restore extern int decc$crtl_init (void); static int rtl_init = 0; /******************************************************************************/ extern int tpu$calluser (int *int_param, struct dsc$descriptor *str_param, struct dsc$descriptor *result_param) { /* ** Translate, delete, or set a logical name via TPU's CALL_USER ** interface. ** ** If the int_param is zero, the logical name is deleted. If it's one, ** the logical name is set. If it's two, the logical name is translated ** and the equivalence name is returned in result_param. ** ** The str_param contains the logical name to delete or translate, or a ** logical name / equivalence name pair separated by a tilde (~). ** ** Example TPU code to call this routine would be: ** ** ! Get the physical terminal name ** result_string := CALL_USER (2, "TT"); ** ** The result string can return three strings other than the equivalence ** name: ** ** SUCCESS indicates the routine deleted or set the logical correctly, ** PARAM_ERROR indicates a parameter error (you didn't set the integer ** parameter to 0, 1, or 2; you passed in a zero length ** string parameter; you passed in a string parameter that ** either did not contain a tilde when you specified "set ** logical"; or the logical or equivalence names exceeded ** 31 ot 255 characters respectively. ** ERROR indicates an unexpected error occurred. */ static char *success = "SUCCESS"; static char *error = "ERROR"; static char *param_error = "PARAM_ERROR"; static char *result_p; static char *p; static unsigned long int r0_status; static unsigned long int flags = LNM$M_CASE_BLIND; static int result_len; static short int i; static short int j; static char log[31]; static char eqv[255]; static char seen_tilde; static struct dsc$descriptor_s log_d = { sizeof (log), DSC$K_DTYPE_T, DSC$K_CLASS_S, log }; static struct dsc$descriptor_d eqv_d = { sizeof (eqv), DSC$K_DTYPE_T, DSC$K_CLASS_S, eqv }; /* ** Because we are not being called by a C main(), initialize the ** C RTL. */ (void)decc$crtl_init (); rtl_init = 1; /* ** Assume success status. */ result_len = strlen (success); result_p = success; if (str_param->dsc$w_length == 0) { /* ** Zero length string. Parameter error. */ result_len = strlen (param_error); result_p = param_error; } else { log_d.dsc$w_length = sizeof (log); eqv_d.dsc$w_length = sizeof (eqv); switch (*int_param) { case 0: /* ** Delete the logical name */ r0_status = lib$delete_logical (str_param); if (r0_status != SS$_NORMAL) { result_len = strlen (error); result_p = error; } break; case 1: /* ** Set the logical name */ /* ** Break the string up into the logical and equiv names. */ seen_tilde = FALSE; p = str_param->dsc$a_pointer; for (i = j = 0; i < str_param->dsc$w_length; i++) { if (p[i] == '~') { seen_tilde = TRUE; log_d.dsc$w_length = j; j = 0; continue; } else { if (seen_tilde) { if (j > sizeof (eqv)) { result_len = strlen (param_error); result_p = param_error; break; } else { eqv[j] = p[i]; } } else { if (j > sizeof (log)) { result_len = strlen (param_error); result_p = param_error; break; } else { log[j] = p[i]; } } j++; } } eqv_d.dsc$w_length = j; if (i == str_param->dsc$w_length && seen_tilde) { /* ** Actually set the logical. */ r0_status = lib$set_logical (&log_d, &eqv_d); if (r0_status != SS$_NORMAL) { result_len = strlen (error); result_p = error; } } if (!seen_tilde) { result_len = strlen (param_error); result_p = param_error; } break; case 2: /* ** Translate the logical name */ r0_status = lib$get_logical (str_param, &eqv_d, &eqv_d.dsc$w_length, 0, 0, 0, 0, &flags); if (r0_status != SS$_NORMAL) { result_len = strlen (error); result_p = error; } else { result_len = eqv_d.dsc$w_length; result_p = eqv_d.dsc$a_pointer; } break; default: /* ** Parameter error */ result_len = strlen (param_error); result_p = param_error; break; } } /* ** Set up the results string descriptor and copy the result to it. */ r0_status = lib$sget1_dd (&result_len, result_param); if (r0_status != SS$_NORMAL) { return r0_status; } (void)memcpy (result_param->dsc$a_pointer, result_p, result_len); return tpu$_success; } $eod $cc log $link/share log,sys$input/opt $deck symbol_vector=(tpu$calluser=procedure) $eod $loc = f$environment ("default") $define tpu$calluser 'loc'log.exe $create x.x $deck PROCEDURE set_logical LOCAL result, log, eqv; log := READ_LINE ("Logical name to define> ", 31); eqv := READ_LINE ("Value> "); EDIT (log, TRIM); EDIT (eqv, TRIM); result := CALL_USER (1, log + "~" + eqv); MESSAGE (result); ENDPROCEDURE; PROCEDURE del_logical LOCAL result, log; log := READ_LINE ("logical name to delete> ", 31); EDIT (log, TRIM); result := CALL_USER (0, log); MESSAGE (result); ENDPROCEDURE; PROCEDURE show_logical LOCAL result, log; log := READ_LINE ("show logical> ", 31); EDIT (log, TRIM); result := CALL_USER (2, log); MESSAGE (result); ENDPROCEDURE; $eod To use, edit X.X and at the command prompt, type "extend all". To demo the functionality, you can then type "tpu show_logical", "tpu set_logical", or "tpu del_logical" at the command prompt. Jim. -- www.eight-cubed.com ------------------------------ Date: Sat, 11 Aug 2007 01:06:21 GMT From: John Santos Subject: Re: What VMS does RIGHT! Message-ID: John Reagan wrote: [snip] > > As for getting it into BASIC, COBOL, and Macro-32, I'll add it to the list. > Thank you! > If I could just get all of you to use the LSE language templates, you > wouldn't be having all these syntax errors, no? :-) > Nah, it's much easier to ask HP to change something than to make any sort of change ourselves. :-) > And before anybody starts THAT thread, we did update several of the > language LSE templates for the latest DECset release. > -- John Santos Evans Griffiths & Hart, Inc. 781-861-0670 ext 539 ------------------------------ Date: Fri, 10 Aug 2007 11:09:12 -0700 From: Sue Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <1186769352.494943.10670@i38g2000prf.googlegroups.com> On Aug 7, 10:55 am, "Richard B. Gilbert" wrote: > John Smith wrote: > >http://tinyurl.com/39ptaq > > > and > > >http://tinyurl.com/2r2w7x > I try. sue > > and > > >http://tinyurl.com/2nekbx > > > Too bad VMS doesn't have one @ HP. > > > -- > > OpenVMS - The never-advertised operating system with the dwindling ISV base. > > I thought that wasSue!- Hide quoted text - > > - Show quoted text - ------------------------------ Date: Fri, 10 Aug 2007 11:16:40 -0700 From: Sue Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <1186769800.899447.53260@d30g2000prg.googlegroups.com> On Aug 7, 12:40 pm, JF Mezei wrote: > Richard B. Gilbert wrote: > > I thought that wasSue! > > Suemay be our champion and the only one within HP able to give us any > hope, but she still isn't "HP" and still isn't allowed to send out a > press release to the news wires, and still not able to do real marketing > outside the installed base. Her hands are tied by Stallard/Livermore. I am not the only one for darn sure. I just happen to be be visable in some places and I happen to really like to talk about VMS, even to people that don't want to hear it. Speaking of, I am working on the 30th Anniversary presentation. There is so much cool information. I do not really want to include what was in the earlier ones. But VMS has really done some amazing things in its life. No wonder we can do what we do. sorry got a little excited there. sue ------------------------------ Date: 10 Aug 2007 14:09:57 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: In article <1186753503.685906.257590@i38g2000prf.googlegroups.com>, AEF writes: > What is the motivations for having a minimum password lifetime? Please > feel free to give a detailed answer. One can only guess what the NIST motivations might be, as they do not explain. One guess is that some operating systems do not have the protection VMS affords against a user repeatedly changing their own password to overflow the history buffer and thus be able to choose the same password again, effectively not changing their password. Note that VMS had that defect in the Field Test version at which password history was introduced (4.X, about 20 years ago) but it was remedied for the production release by throwing the user into generated password mode if their history buffer was full. ------------------------------ Date: Fri, 10 Aug 2007 13:03:07 -0700 From: yyyc186 Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <1186776187.465658.269550@x40g2000prg.googlegroups.com> > > If you define recently as 1977, yeah... otherwise no, CICS has handled > distributed transactions and distributed processing for 30 years or so. No it hasn't. Big Blue was "mono-iron" until the 90's. It doesn't count when you virtualize one box into 6, then "distribute" transactions across it. > > >You mean the broker that is currently up for sale? > > That would be the London Stock Exchange- the Schwab "Dutch Auction" was > pretty cool. Not even close. Schwab is/was one of the brokerage houses up for sale this year. Waterhouse was looking at buying them before they bought Ameritrade. ------------------------------ Date: Fri, 10 Aug 2007 23:57:51 +0200 From: "Dr. Dweeb" Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <46bcdf63$0$7606$157c6196@dreader2.cybercity.dk> yyyc186 wrote: > On Aug 10, 11:19 am, "Paul Raulerson" wrote: >> No, I do not mean replicated, as you would have to do under VMS. The >> data for each location resides in whatever geographic location it >> has been assigned to, and is not replicated to each other location. >> (It is replicated via storage systems for DR and such, but that has >> *nothing* to do with what we are talking about.) >> > > It is what you mean because the "modern" feature you are talking about > is ancient when implemented on a great OS. OpenVMS clusters mounted > all of the DASD cluster wide (with very few exceptions). When you > laid out your RDB storage and tables you placed the whatever tables > needed to be in whatever location for access speed. We've been geo- > locating relational database pieces since the early 1980's. Of > course, when your OS never managed to actually cluster, being able to > do that must seem like a gift from God. > > > What actually got implemented for the "Federated Database" products is > a rash of cheap and dirty data replication. It is truly pathetic when > you look under the hood. 6 months or torture with SQLServer under my belt - when I ask myself why I bother, I remind myself why there are no jobs in the Rdb/VMS world, and why the world is ignorant of just how much better Rdb+VMS is compared to other solutions. Dweeb. ------------------------------ Date: Fri, 10 Aug 2007 17:21:17 -0500 From: "Paul Raulerson" Subject: RE: Wonderful things happen to an OS when it has an internal champion Message-ID: <010d01c7db9c$c5fa6110$51ef2330$@com> > -----Original Message----- > From: yyyc186 [mailto:yyyc186@hughes.net] > Sent: Friday, August 10, 2007 3:03 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Wonderful things happen to an OS when it has an internal > champion > > > > > If you define recently as 1977, yeah... otherwise no, CICS has > handled > > distributed transactions and distributed processing for 30 years or > so. > > No it hasn't. Big Blue was "mono-iron" until the 90's. It doesn't > count when you virtualize one box into 6, then "distribute" > transactions across it. > Where are you getting this information? It is simply wrong. CICS distributed transactions have been around since the 1970's, with mainframes connected over all sorts (by today's standards) ridiculously low bandwidth connections. > > > > >You mean the broker that is currently up for sale? > > > > That would be the London Stock Exchange- the Schwab "Dutch Auction" > was > > pretty cool. > > Not even close. Schwab is/was one of the brokerage houses up for sale > this year. Waterhouse was looking at buying them before they bought > Ameritrade. Last I looked Schwab was buying back all their own stock. ------------------------------ Date: Fri, 10 Aug 2007 23:07:15 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: In article <46bcdf63$0$7606$157c6196@dreader2.cybercity.dk>, "Dr. Dweeb" writes: > > >yyyc186 wrote: >> On Aug 10, 11:19 am, "Paul Raulerson" wrote: >>> No, I do not mean replicated, as you would have to do under VMS. The >>> data for each location resides in whatever geographic location it >>> has been assigned to, and is not replicated to each other location. >>> (It is replicated via storage systems for DR and such, but that has >>> *nothing* to do with what we are talking about.) >>> >> >> It is what you mean because the "modern" feature you are talking about >> is ancient when implemented on a great OS. OpenVMS clusters mounted >> all of the DASD cluster wide (with very few exceptions). When you >> laid out your RDB storage and tables you placed the whatever tables >> needed to be in whatever location for access speed. We've been geo- >> locating relational database pieces since the early 1980's. Of >> course, when your OS never managed to actually cluster, being able to >> do that must seem like a gift from God. >> >> >> What actually got implemented for the "Federated Database" products is >> a rash of cheap and dirty data replication. It is truly pathetic when >> you look under the hood. > >6 months or torture with SQLServer under my belt - when I ask myself why I >bother, I remind myself why there are no jobs in the Rdb/VMS world, and why >the world is ignorant of just how much better Rdb+VMS is compared to other >solutions. Dweeb, do not dwell on it for too long lest you wind up in some catatonic state of blithering idiocy never to be awakened. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" http://tmesis.com/drat.html ------------------------------ Date: Fri, 10 Aug 2007 17:09:01 -0600 From: Keith Parris Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: Doug Phillips wrote: > Looks like it would take a rather large (and patient) backing and some > creative middle-ware acquisitions for that company to ever see black > ink. If VMS were to be spun off, it would have to include both the OpenVMS hardware and software business as well as the OpenVMS Services business. Both portions would be very profitable, but the Services revenues would be about 3-4 times that of the HW/SW side. But it would all add up to a very-profitable business of about $1B annual revenues. ------------------------------ Date: Fri, 10 Aug 2007 17:13:29 -0600 From: Keith Parris Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: David J Dachtera wrote: > Keith Parris wrote: >>Profitability has never been a problem for OpenVMS. So IBM couldn't >>achieve "a return to profitability" for OpenVMS. > > If so, why the cutbacks / early retirements / off-shoring / lack of marketing / > etc.? Doesn't seem justifiable. OpenVMS is only about 6-7 percent of BCS (although its share is growing as, for example, Tru64 has declined). And OpenVMS Services revenues are only about 6-7 percent of HP Services. So it's just small enough within each of those realms that it doesn't get seem to get much attention, and we've seen it contribute more than its share to cost-cutting measures lately, IMHO. ------------------------------ Date: Fri, 10 Aug 2007 17:18:00 -0600 From: Keith Parris Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: Ron Johnson wrote: > On 08/09/07 15:13, Keith Parris wrote: >>Another factor affecting product availability, in addition to the need >>or demand, is the relative difficulty of implementation. Scalability >>clusters are relatively easy to create. High-availability, >>single-system-image clusters, with shared simultaneous read/write access >>to file systems from different nodes, shared root disk, etc. aren't so >>simple to create. > > But you'd think that *somebody* else would have done it in the past > 30 years! Somebody has. TruClusters. (May they Rest in Peace. Actually, I'd rather see them rest in open source. There is precedent for that -- NonStop Clusters for UNIXware was open-sourced by Compaq.) > Or did DEC have all the important concepts bound up in patents? US patents last 17 years. So most of the pivotal ones have expired now. Recent stuff like Host-Based Mini-Merges would be another story, of course. ------------------------------ Date: Fri, 10 Aug 2007 18:31:13 -0500 From: "Paul Raulerson" Subject: RE: Wonderful things happen to an OS when it has an internal champion Message-ID: <010f01c7dba6$8ad94f60$a08bee20$@com> I'm not sure that a spin off would be a "good thing" at all for VMS. It has been disassociated somewhat from any particular hardware platform, and it isn't "open" enough, or really has a planned future on any hardware but Itanium. That of course, may be enough. I'm off to see a "Superdome" installation on Tuesday, and I am prepared to be impressed. :) -Paul > -----Original Message----- > From: Keith Parris [mailto:keithparris_nospam@yahoo.com] > Sent: Friday, August 10, 2007 6:09 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Wonderful things happen to an OS when it has an internal > champion > > Doug Phillips wrote: > > Looks like it would take a rather large (and patient) backing and > some > > creative middle-ware acquisitions for that company to ever see black > > ink. > > If VMS were to be spun off, it would have to include both the OpenVMS > hardware and software business as well as the OpenVMS Services > business. > Both portions would be very profitable, but the Services revenues would > be about 3-4 times that of the HW/SW side. But it would all add up to a > very-profitable business of about $1B annual revenues. ------------------------------ Date: Fri, 10 Aug 2007 20:30:50 -0400 From: Glenn Everhart Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <46BD033A.9000008@gce.com> VAXman- @SendSpamHere.ORG wrote: > In article <46bcdf63$0$7606$157c6196@dreader2.cybercity.dk>, "Dr. Dweeb" writes: >> >> yyyc186 wrote: >>> On Aug 10, 11:19 am, "Paul Raulerson" wrote: >>>> No, I do not mean replicated, as you would have to do under VMS. The >>>> data for each location resides in whatever geographic location it >>>> has been assigned to, and is not replicated to each other location. >>>> (It is replicated via storage systems for DR and such, but that has >>>> *nothing* to do with what we are talking about.) >>>> >>> It is what you mean because the "modern" feature you are talking about >>> is ancient when implemented on a great OS. OpenVMS clusters mounted >>> all of the DASD cluster wide (with very few exceptions). When you >>> laid out your RDB storage and tables you placed the whatever tables >>> needed to be in whatever location for access speed. We've been geo- >>> locating relational database pieces since the early 1980's. Of >>> course, when your OS never managed to actually cluster, being able to >>> do that must seem like a gift from God. >>> >>> >>> What actually got implemented for the "Federated Database" products is >>> a rash of cheap and dirty data replication. It is truly pathetic when >>> you look under the hood. >> 6 months or torture with SQLServer under my belt - when I ask myself why I >> bother, I remind myself why there are no jobs in the Rdb/VMS world, and why >> the world is ignorant of just how much better Rdb+VMS is compared to other >> solutions. > > Dweeb, do not dwell on it for too long lest you wind up in some catatonic > state of blithering idiocy never to be awakened. > There was an interesting paper given at the recent Usenix (news reports on slashdot and the like covered it and have references) in which it was pointed out that modern unix kernels - several of them plus Linux - have security holes that stem from their system kernel calls pulling arguments in from their callers in whatever order they find convenient, at different times. This allows interposition systems to be thwarted with classic time of test vs time of use attacks (race conditions) but points to serious issues whose impact is much wider spread. Compare with VMS behavior which (influenced by the pdp11 heritage where kernel really could not easily get at user space arguments once APRs changed, perhaps) generally packs arguments up and pulls them into kernel space, then uses them. (This is pretty clear for I/O calls where things go into IRPs; less obvious for others without checking them). I believe this inhibits many such attacks. (I ran into this one when writing Safety, and like everyone else in the area was concerned lest something asynchronous screw up arguments. Generally I pulled anything in to read ACLs and get file IDs and did all other access by file ID, ignoring what happened to things like filenames.) Point to make is that VMS has historically considered this problem and is able to check arguments before use, and has always been rather careful to do so. This is a kind of quality which evidently has been unconsidered in the unix community till now, yet is important. Also the recently published U. Wis. thesis of V. Prabhakaran "Iron File Systems" goes into depth about NTFS and several modern Unix filesystems. The kinds of data corruption errors he discusses are precisely the kinds of errors DKdriver has had extensive code to recover from for ages. To be sure, the error recovery in VMS is done in drivers primarily, not in filesystems, but the point to be made again is that it is done in VMS with considerable effort and care. It is also to be noted that consumers of "popular" systems have on the whole been unaware that the issue even exists. (Arguably they are too inured to having to power cycle their machines and lose files to care. There is something to be said for just not expecting much...) It is in part in issues like these where the quality with which VMS was built and rebuilt over the years shows forth. I suppose it is barely conceivable that bringing these up in publicity could have benefit, to the extent other OS users pay attention. Considering that some of the competition for VMS in situations where data integrity and security are wanted is in "hardened" Unix - like systems where precisely these issues remain, might it not behoove HP to allow its minions to bring these and other advantages to the attention of customers? I recommend the papers mentioned though: good reads. Glenn Everhart ------------------------------ Date: Fri, 10 Aug 2007 20:08:16 -0500 From: "Paul Raulerson" Subject: RE: Wonderful things happen to an OS when it has an internal champion Message-ID: <000001c7dbb4$1956bea0$4c043be0$@com> > -----Original Message----- > From: Glenn Everhart [mailto:everhart@gce.com] > Sent: Friday, August 10, 2007 7:31 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Wonderful things happen to an OS when it has an internal > champion > > > > There was an interesting paper given at the recent Usenix (news reports > on slashdot > and the like covered it and have references) in which it was pointed > out that modern > unix kernels - several of them plus Linux - have security holes that > stem from > their system kernel calls pulling arguments in from their callers in > whatever > order they find convenient, at different times. This allows > interposition systems > to be thwarted with classic time of test vs time of use attacks (race > conditions) but > points to serious issues whose impact is much wider spread. There are operators that come into play to negate these issues on most data center systems. For example, I/O is converted to channel programs on a mainframe, which do NOT have this problem. The same is true when for example, accessing data from a fibre based SAN, but... it is an issue from iSCSI, which is what it looks like brought the whole issue up. I believe you when you say VMS handles this very well- to the point it is not even an issue. But it does point out a very large culture difference between VMS and the "rest of the world." (Note the quotes!) VMS systems still tend to think of DASD as belonging to a particular system, while the more mainstream world has not gone down that path for several years. In fact, when I first started working with VMS (oh- 8 weeks ago now? :) I was rather startled but the aggressive insistence that shadowing volumes was a great idea. Well, it is a good idea - but I was surprised that in the VMS world, that kind of stuff is handled by - well - VMS. On a *host*. That is, on other platforms, considered distinctly low end. I have a little BA array here configured as RAID-5, and people, whose opinions I respect, tell me it should be shadowed. :) That mentality comes from having to deal directly with individual DASD devices, rather than virtualized devices from a SAN or some other kind of storage array. It bespeaks to me that VMS could possibly be a used as a dynamite SAN controller system, which is a flip flop from what I first assumed. To come back to the point, the issues are not issues in application systems in most data centers, because the use of "local drives", driven directly by the system, is not something that fits into most storage plans these days. And using systems like AIX for SAN controllers - the system has been modified so that the problem just isn't a problem. In other words, while you are precisely correct, it is a non-issue in most installations. Sort of a "cry wolf." > > This is a kind of quality which evidently has been unconsidered in the > unix community > till now, yet is important. > It is not unconsidered - merely solved differently. The solution of which must seem rather strange to some VMS people, I suppose. I do like the way VMS solves it, but it places VMS into a different fit in my mind. VMS is utterly superior to Windows servers for applications and such, and very much superior in most areas to Linux in the same situation. Meaning primarily Linux on x86 boxes. > Also the recently published U. Wis. thesis of V. Prabhakaran "Iron File > Systems" goes > into depth about NTFS and several modern Unix filesystems. The kinds of > data corruption > errors he discusses are precisely the kinds of errors DKdriver has had > extensive > code to recover from for ages. To be sure, the error recovery in VMS is > done in > drivers primarily, not in filesystems, but the point to be made again > is that it > is done in VMS with considerable effort and care. It is also to be > noted that > consumers of "popular" systems have on the whole been unaware that the > issue > even exists. (Arguably they are too inured to having to power cycle > their machines > and lose files to care. There is something to be said for just not > expecting > much...) Well, if by "popular" systems you mean the general run of Dell PC's - then I can do nothing but agree with you. If however, you mean enterprise class systems, then I do not. I must blushingly admit to forgetting how to IPL the mainframe once, since it was not IPLed for over 2.5 years. And it was only IPL'ed then because I upgraded the OS and added a bunch of new hardware. (It did not really have to power-on IPL, but I planned it that way so I only had to do it once.) I understand there are VMS systems with that kind of reliability too. What impresses me is that the reliability is not so much because of the hardware, but because of the OS. That is, by the way, also true of Mainframe, iSeries, pSeries, Tandem, and other systems and OS combinations. Indeed, it is shocking how much of the stability exhibited by any computer system is dependent upon the OS. VMS does shine here without any question. -Paul ------------------------------ Date: Fri, 10 Aug 2007 20:59:09 -0700 From: "Tom Linden" Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: On Fri, 10 Aug 2007 16:09:01 -0700, Keith Parris wrote: > Doug Phillips wrote: >> Looks like it would take a rather large (and patient) backing and some >> creative middle-ware acquisitions for that company to ever see black >> ink. > > If VMS were to be spun off, it would have to include both the OpenVMS > hardware and software business as well as the OpenVMS Services business. > Both portions would be very profitable, but the Services revenues would > be about 3-4 times that of the HW/SW side. But it would all add up to a > very-profitable business of about $1B annual revenues. Well, if I were running it, it would be a software company, there are plenty of companies around capable of building good hardware. Likely senior management would not (mercifully) be brought along. Are you sure about the revenue figures? The numbers of which I had been given two years ago were 4 times that amount with a 20% margin -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Fri, 10 Aug 2007 21:37:15 -0700 From: "Tom Linden" Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: On Fri, 10 Aug 2007 18:08:16 -0700, Paul Raulerson wrote: > Indeed, it is shocking how much of the stability exhibited by any > computer system is dependent upon the OS. VMS does shine here without > any question. Well you should get a good chuckle out of the following: EE Times: Latest News VMware predicts death to operating systems Antone Gonsalves InformationWeek (08/09/2007 4:01 PM EDT) In the view of Mendel Rosenblum, chief scientist and co-founder of virtualization vendor VMware, today's modern operating system is destined for the dustbin, a scenario unlikely to please Microsoft or any of the Linux vendors. Rosenblum's keynote on Thursday wrapped up the LinuxWorld conference in San Francisco, preaching the virtues of virtualization, which he believes will eventually make today's complex, some would say bloated, operating systems obsolete. "It's just going to go away," Rosenblum said. Not surprisingly, Rosenblum favors a world in which a virtualization layer is tied directly to the microprocessor and other related hardware of a computer. Running on top of this layer would be virtual machines, or mini-operating systems, that would be designed to run specific applications. Merging the OS and software would create a module that would be more reliable and secure, easier to manage and offer higher performance. The reason is in the simplicity of the architecture, Rosenblum said. Operating systems, particularly Windows, have become behemoths comprising millions of lines of code expected to share resources among multiple applications. Such complexity degrades performance and makes the software less reliable and harder to manage. "If you have something this complex, it's also really hard to innovate," Rosenblum said. "This is the position that Microsoft is in, even with its team of engineers, which is huge. It's hard to do anything." The better architecture is to build software for what Rosenblum called a virtualization appliance. Software makers could package within a virtual machine only those components needed to run a particular application. "I can start simplifying these things," Rosenblum said. "I can take out parts of the OS I don't need for this application, and build an OS that's highly optimized for the application." The self-contained module would inherently have higher performance and would be less complex in terms of security, since any flaws would be the problem of the ISV, he said. There would be no separate operating system to point to as the culprit. VMware offers a variety of virtual machine management services such as dynamic balancing and resource allocation and centralized backup software. More services are coming, Rosenblum said. A new feature in beta that's available on VMware software for workstations is technology that would record the execution of a server on a virtual machine for replay later. The technology is still at the very early stages. "Our motto has always been to release it first, and let the most adventurous play with it," Rosenblum said. While it may look like VMware is building its own OS, Rosenblum insisted that isn't the case. "We don't support applications directly," he said. Linux would be the best option for building software-specific VMs because its open-source licensing terms are more flexible than those of proprietary OS vendors. Essentially, there are no royalties to pay with Linux, and users can mold the software to fit their purposes. Rosenblum sees a future in which virtualization companies like VMware are battling with OS builders like Microsoft in partnering with hardware vendors to create the best environment for their products. "If I was in their shoes, I would try to hang on to the hardware as much as I can," Rosenblum said during a question and answer session that followed his speech. "There's going to be a big battle over who is going to own this layer. The battle is going to be who's going to do the hypervisor that everybody uses." The hypervisor is the system program that provides the virtual machine environment. The term came from the IBM mainframe world. IBM was the first to introduce the concept of a virtual machine. "I give IBM credit for inventing this, but they also in some sense killed it (by turning toward operating systems)," Rosenblum said. On the hardware side, Dell Chief Technology Officer Kevin Kettler said Tuesday at LinuxWorld that the computer maker was experimenting with a hypervisor that would ship with its servers in order to provide customers with a high-performing virtualization environment out of the box. Similarly, chipmakers Intel and Advanced Micro Devices ship microprocessors with embedded virtualization technology, which both companies have targeted as a key area for future developme -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Fri, 10 Aug 2007 20:33:07 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: X Window Servers Message-ID: <7Y3vi.34$GQ4.20@newsfe12.lga> In article , koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: > > >In article <_aZui.27225$RX.17930@newssvr11.news.prodigy.net>, "Brian Tillman" writes: >> >> DECwindows 101. > > Odd, then, that you get the error you posted. I haven't seen > that on properly configured systems. > > Is it possible that something is blocking X11 on port 600? Just to set the record straight, the port is 6000, not 600. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" http://tmesis.com/drat.html ------------------------------ End of INFO-VAX 2007.437 ************************