1 INFO-VAX	Mon, 03 Sep 2001	Volume 2001 : Issue 489       Contents: Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64 Re: alpha - ia64( Compaq's Spin on IBM's AIX Monterey drop, Re: Compaq's Spin on IBM's AIX Monterey drop, Re: Compaq's Spin on IBM's AIX Monterey drop; Re: Compiled Languages (Was: e: My VMS Wish List (features)  Re: I hate Compaq  RE: I hate Compaq  Re: SMTP Que Stalls  Re: Tunneling DECnet over IP  Re: VMS 7.3 compatibility issues VMS NT/win2000 similarities  Re: VMS NT/win2000 similarities  RE: VMS NT/win2000 similarities  Re: VMS NT/win2000 similarities   F ----------------------------------------------------------------------  % Date: Sun, 02 Sep 2001 12:44:59 -0700 $ From: Alex Johnson <spectre@jhu.edu> Subject: Re: alpha - ia64 ' Message-ID: <3B928C3B.48D85801@jhu.edu>    Yousuf Khan wrote: > 3 > "Alex Johnson" <spectre@jhu.edu> wrote in message # > news:3B901219.194A55FB@jhu.edu...  > > Yousuf Khan wrote: > > > M > > > One will note that when the Hammer line comes out from AMD, it's 64-bit  > modeM > > > will default to 64-bit registers for addressing operands, but to 32-bit K > > > registers for data operands. Just proves that 64-bit is only good (so  > far)+ > > > for addressing and nothing much more.  > > K > > Well, except when you keep your data as 32 bit and your addresses as 64 L > > bit, then some bad programmer startes manipulating addresses as data andG > > you end up forced into 32 bit addresses again.  Never trust the end H > > user.  Make all integers 64 bit by default.  It isn't much slower toL > > compute 64 bit adds than 32 bit adds and space is not an issue any more. > D > It's not possible to manipulate addresses as 32-bit unless you useM > address-size-overrides, or code the whole program in the legacy 32-bit mode 3 > in which case it will be run in a 32-bit segment.  >  >     Yousuf Khan   G You are saying I cannot have an array of data.  The start of which is P E (an address).  To access the 2048th element I would compute 2048*4 (a G data value) and add this to P.  We are now manipulating an address like G data which would default to 32-bits and some information is lost.  That D is, unless you can compute 64-bit data, but then what's the point ofH saying there is a "default" since each opcode has to be specified by theH programmer/compiler and that means you will declare every data as eitherG 32 or 64 bit loads/stores/arithmatic.  Now that I look at it again, the @ very idea of saying pointers and data DEFAULT to some size seems
 ludicrous.   Alex --  G My words are my own.  They represent no other; they belong to no other. E Don't read anything into them or you may be required to compensate me % for violation of copyright and/or IP.    ------------------------------  % Date: Mon, 03 Sep 2001 00:05:21 +0200 ( From: Bernd Paysan <bernd.paysan@gmx.de> Subject: Re: alpha - ia64 & Message-ID: <3B92AD21.269DA353@gmx.de>   eddie wrote:= > I agree with Australia's decision to keep the refugees out.   D They are not refugees anymore, because their boat sank. Now they areH shipwrecked, and international sea law requires the next save harbour toB let them in. The next save harbour is the one of Christmas Island,F Australia. What happens to them after they are back on soil is another@ story, Australia doesn't have the duty to let them stay forever.  H International sea law is old and established, it is the only law againstH piracy, slavery and other bad things that happend on the open sea beforeG it was defined. Outside the national borders, it's the only common law. G Australia may have reasons not to let in refugees, but it has no reason G to refuse to help shipwrecked people; no matter how they came to become H shipwrecked. Or do Australian people prefer to be shanghaied, pirated orD such like on sea again? Do Australians even care if some other boatsD come to help them in case of distress at sea? Yes? Then help others.   --   Bernd Paysan7 "If you want it done right, you have to do it yourself"  http://www.jwdt.com/~paysan/   ------------------------------  % Date: Mon, 03 Sep 2001 09:29:54 +0930 / From: Mark Daniel <Mark.Daniel@wasd.vsm.com.au>  Subject: Re: alpha - ia64 / Message-ID: <3B92C7FA.3B2BB216@wasd.vsm.com.au>   @ I'm tempted to take exception to this throw-away comment but I'mG buggered if I can think of anything civilised, witty or educated to say 	 in reply.    Nick Maclaren wrote: > ? > In article <B4kj7.54329$n75.13608013@news4.rdc1.on.home.com>, 3 > Yousuf Khan <bbbl67@nospam.yahoo.com.spam> wrote: : > >"Sarah Warner" <sarahamiee7@yahoo.com> wrote in message= > >news:5vgj7.162589$Xr6.910102@news-server.bigpond.net.au... $ > >> EPIC is better than RISC, mate. > >> > >> Do a bit of research. > > N > >You must be the same person as Smitty, right? Both from Australia, what areO > >the chances that there are two complete morons out of the millions that come  > >from country? > A > The UK is full of civilised, witty and educated Australians who E > left because they couldn't stand the density of the morons in their  > native country ....  > 
 > Regards, > Nick Maclaren,, > University of Cambridge Computing Service,@ > New Museums Site, Pembroke Street, Cambridge CB2 3QG, England. > Email:  nmm1@cam.ac.uk1 > Tel.:  +44 1223 334761    Fax:  +44 1223 334679    --  # Non sinere illegitamus carborundum.    ------------------------------  * Date: Mon, 3 Sep 2001 02:44:30 +0000 (UTC)7 From: dsiebert@excisethis.khamsin.net (Douglas Siebert)  Subject: Re: alpha - ia64 + Message-ID: <9muqqe$1ga$1@sword.avalon.net>   I Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> writes:   1 >"aaron spink" <aaronspink@earthlink.net> writes:   M >> Over the past 7 years, Alpha was in the top spot for SpecIntXXX except for M >> about 6 non contiguous months, for SpecFpXXX the number is pretty much the 2 >> same.  Kind of an interesting graph to look at.   >Cute.    K >IIRC, a lot of times these 'lapses in leadership' were the result of DEC's I >or Compaq's marketing efforts, i.e., not announcing a given processor or K >upgrade until they thought the time was ripe - with more delays than usual  >in the market.     E Which was at least balanced by their retaking the lead from HP on two = occasions with hardware that didn't ship for quite some time.    --H Douglas Siebert                          dsiebert@excisethis.khamsin.net  M I have discovered a remarkable proof which this .sig is too small to contain!    ------------------------------  # Date: Mon, 03 Sep 2001 02:48:34 GMT   From: "eddie" <NullVoid@att.net> Subject: Re: alpha - ia64 F Message-ID: <6cCk7.9434$KV3.714178@bgtnsc04-news.ops.worldnet.att.net>  5 "Bernd Paysan" <bernd.paysan@gmx.de> wrote in message   news:3B92AD21.269DA353@gmx.de... > eddie wrote:? > > I agree with Australia's decision to keep the refugees out.  > F > They are not refugees anymore, because their boat sank. Now they areJ > shipwrecked, and international sea law requires the next save harbour toD > let them in. The next save harbour is the one of Christmas Island,H > Australia. What happens to them after they are back on soil is anotherB > story, Australia doesn't have the duty to let them stay forever.   Thats not true!   K The US does the same with Haitian and Cuban refugees when their boats sink. I They get sent back to their own country unless they reach US shore first.    > J > International sea law is old and established, it is the only law againstJ > piracy, slavery and other bad things that happend on the open sea beforeI > it was defined. Outside the national borders, it's the only common law. I > Australia may have reasons not to let in refugees, but it has no reason I > to refuse to help shipwrecked people; no matter how they came to become J > shipwrecked. Or do Australian people prefer to be shanghaied, pirated orF > such like on sea again? Do Australians even care if some other boatsF > come to help them in case of distress at sea? Yes? Then help others. >  > -- > Bernd Paysan9 > "If you want it done right, you have to do it yourself"  > http://www.jwdt.com/~paysan/   ------------------------------  # Date: Mon, 03 Sep 2001 03:41:05 GMT 2 From: "Yousuf Khan" <bbbl67@nospam.yahoo.com.spam> Subject: Re: alpha - ia64 = Message-ID: <lZCk7.82535$n75.20396264@news4.rdc1.on.home.com>   1 "Alex Johnson" <spectre@jhu.edu> wrote in message ! news:3B928C3B.48D85801@jhu.edu... J > > > Well, except when you keep your data as 32 bit and your addresses as 64J > > > bit, then some bad programmer startes manipulating addresses as data and I > > > you end up forced into 32 bit addresses again.  Never trust the end J > > > user.  Make all integers 64 bit by default.  It isn't much slower toH > > > compute 64 bit adds than 32 bit adds and space is not an issue any more.  > > F > > It's not possible to manipulate addresses as 32-bit unless you useJ > > address-size-overrides, or code the whole program in the legacy 32-bit mode5 > > in which case it will be run in a 32-bit segment.  > I > You are saying I cannot have an array of data.  The start of which is P G > (an address).  To access the 2048th element I would compute 2048*4 (a I > data value) and add this to P.  We are now manipulating an address like I > data which would default to 32-bits and some information is lost.  That F > is, unless you can compute 64-bit data, but then what's the point ofJ > saying there is a "default" since each opcode has to be specified by theJ > programmer/compiler and that means you will declare every data as eitherI > 32 or 64 bit loads/stores/arithmatic.  Now that I look at it again, the B > very idea of saying pointers and data DEFAULT to some size seems > ludicrous.  D I'm not sure what you see as a problem here. All memory accesses areI defaulted to 64-bit, and any operands within it would be 32-bit defaults. / Any relative addressing is also done as 64-bit.        Yousuf Khan    ------------------------------  # Date: Mon, 03 Sep 2001 04:09:27 GMT / From: andrew@gurney.reilly.home (Andrew Reilly)  Subject: Re: alpha - ia64 7 Message-ID: <slrn9p60jo.170h.andrew@gurney.reilly.home>   4 On Mon, 03 Sep 2001 03:41:05 GMT, Yousuf Khan wrote:F > I'm not sure what you see as a problem here. All memory accesses areK > defaulted to 64-bit, and any operands within it would be 32-bit defaults. 1 > Any relative addressing is also done as 64-bit.   @ The problem arises as soon as you want to manipulate an array of? something with more than 2G _elements_.  Sure, the start of the @ array is a 64-bit address, and address arithmetic has to be done@ with address-sized integers, but C (at least) insists that array; indices are "int".  If int is only 32-bits, then you have a  problem.  ? Other languages are probably OK to allow two classes of arrays: = "small" (<2G elements) and "large" (>2G elements), but C only  knows about one style of array.   7 Question: how _do_ other languages handle this problem?    --   Andrew   ------------------------------  # Date: Mon, 03 Sep 2001 04:20:55 GMT 2 From: "Yousuf Khan" <bbbl67@nospam.yahoo.com.spam> Subject: Re: alpha - ia64 = Message-ID: <HyDk7.82828$n75.20464393@news4.rdc1.on.home.com>   < "Andrew Reilly" <andrew@gurney.reilly.home> wrote in message1 news:slrn9p60jo.170h.andrew@gurney.reilly.home... B > The problem arises as soon as you want to manipulate an array ofA > something with more than 2G _elements_.  Sure, the start of the B > array is a 64-bit address, and address arithmetic has to be doneB > with address-sized integers, but C (at least) insists that array= > indices are "int".  If int is only 32-bits, then you have a 
 > problem.  L If "C" insists on the index variables being 32-bit, then you have a 32-bit C" compiler, not a 64-bit C compiler.       Yousuf Khan    ------------------------------    Date: 03 Sep 2001 01:01:28 -0400# From: Chris Morgan <cm@mihalis.net>  Subject: Re: alpha - ia64 / Message-ID: <87g0a4sw9j.fsf@tweety.mihalis.net>   1 andrew@gurney.reilly.home (Andrew Reilly) writes:   A > Other languages are probably OK to allow two classes of arrays: ? > "small" (<2G elements) and "large" (>2G elements), but C only ! > knows about one style of array.  > 9 > Question: how _do_ other languages handle this problem?    In Ada it's something like  $ type foo is range -2**32 .. 2**32-1;, type large_array is array (foo) of integer;    Chris 1 -- didn't compile the above, left Ada 7 years ago  --  H Chris Morgan <cm at mihalis.net>                  http://www.mihalis.net        Temp sig. - Enquire within   ------------------------------  $ Date: Sun, 2 Sep 2001 22:10:56 -0700+ From: "Dennis O'Connor" <dmoc@primenet.com>  Subject: Re: alpha - ia64 - Message-ID: <9mv3eb$pob$1@nnrp3.phx.gblx.net>   . "Bernd Paysan" <bernd.paysan@gmx.de> wrote ...( > ... international sea law requires ...  ; Since their is no international police force, your deluding : yourself to think that "international law" is anything but) a matter of convenience for most nations.   7 Of course, the major powers are happy enough to enforce 3 "international law" ... when it's convenient to ...   5 That's the way it is; it's also the way it should be.  --3 Dennis O'Connor                   dmoc@primenet.com . Vanity Web Page http://www.primenet.com/~dmoc/   ------------------------------  $ Date: Sun, 2 Sep 2001 20:48:41 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> 1 Subject: Compaq's Spin on IBM's AIX Monterey drop 9 Message-ID: <PmAk7.3748$A24.512933@news20.bellglobal.com>   & So IBM is going to focus on Power4 eh?  ' http://www.theinquirer.net/02090110.htm   6 What was that colloquialism about strange bed fellows?  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------  $ Date: Sun, 2 Sep 2001 22:55:10 -0400' From: "Bill Todd" <billtodd@foo.mv.com> 5 Subject: Re: Compaq's Spin on IBM's AIX Monterey drop ( Message-ID: <9murcq$5e0$1@pyrite.mv.net>  4 "Neil Rieck" <n.rieck@sympatico.ca> wrote in message3 news:PmAk7.3748$A24.512933@news20.bellglobal.com... ( > So IBM is going to focus on Power4 eh?  J Yet another area in which Capellas' attempt to emulate IBM fails dismally.   ------------------------------  $ Date: Mon, 3 Sep 2001 00:00:13 -0400' From: "Bill Todd" <billtodd@foo.mv.com> 5 Subject: Re: Compaq's Spin on IBM's AIX Monterey drop ( Message-ID: <9muv6q$7hr$1@pyrite.mv.net>  F Sorry 'bout that first uncharacteristically-brief post:  a slip of the finger.   4 "Neil Rieck" <n.rieck@sympatico.ca> wrote in message3 news:PmAk7.3748$A24.512933@news20.bellglobal.com... ( > So IBM is going to focus on Power4 eh?  J Yet another area in which Capellas' attempt to emulate IBM fails dismally:    F "IBM feels that AIX no longer had a chance to be the predominate (sic)J operating system on Intel.  Linux and Windows are now their picks for that spot."  H And why should Compaq believe that Tru64 has any better chance on ItanicC than AIX would?  Especially with  a) Tru64/Linux application-binary L compatibility as an announced goal and  b) all the Tru64 goodies that Compaq is giving to Linux.   J The difference is that IBM is retaining a hardware platform with which AIXI may be able to enjoy some real synergy, while Compaq just fatally wounded H theirs (and possibly Tru64's - and VMS's - survival prospects along with it).    I "AIX 5L will still remain on the Power architecture.  IBM feels the Power H architecture will give them more opportunity to differentiate their UNIXL solutions.  This move simplifies IBM's platform architectures and marketing.J AIX and OS/400 will be on Power, and Linux plus Windows will be on Intel."  K And that, folks, is strategic thinking, rather than what passes for such at J Compaq.  Of course, it's likely that Linux will also be available on Power7 in the future for those who want it, just as it is now.     K "IBM has made a tremendous investment in the Power architecture in the last G few years.  Power4 is due to be announced in September and will ship in E volume by December.  This move allows IBM to refocus their enterprise 9 marketing on a single architecture that is owned by IBM."   K Yup.  Whereas Compaq now gets to wait from two to four years before its own K enterprise systems (first Tru64, then VMS, finally NSK) have been ported to ; a far more questionable, but now the only viable, platform.e    L "They will tout Power as a great chip, much better than IPF for at least the next 5 years."  J And they'll have the distinct advantage of being able to do so truthfully.: As Compaq could have done with Alpha even more truthfully.    J "IBM was planning to use the Sequent NUMA technology on the Intel platformK to form big NUMA clusters.  Now it appears as if they are deploying NUMA onl Power4-based Regatta instead."  K While with no proprietary processor advantage Compaq fights it out with SGI @ (a competitior with far greater experience in making even modestB microprocessors do amazing things in large NUMA servers) and HP (aE competitor with an already-designed high-end Itanic NUMA server and a8I two-year head start before Compaq's first enterprise system is running ondH Itanic).  Either SGI and/or HP succeed and gobble up the high end of theG market before Compaq even shows up, or the high end decides that Itanic G isn't the ship it wants to sail on at all.  Either way, Compaq's toast.r    & And then we get to the suggested spin:    L "IBM did not want to compete with high volume Intel platform reseller CompaqH for UNIX on Intel.  When Compaq committed to port Tru64 UNIX to IPF, IBMC felt it was too much competition for the platform and dropped out."o  L Someone seems to have confused Capellas' absence of balls with the situationK at IBM:  while Compaq may have dropped Alpha because of completely vaporousnF Itanic competition, IBM would hardly have given up two years' worth ofG revenue before Tru64 even shows up if it thought that revenue was worthp	 pursuing.m    I "This move was forced by AIX ISV's who told IBM they would support either  Power or Intel, but not both."  I Wonder if this was what Compaq's own ISVs told it about Alpha and Itanic? J But the real question is how many Alpha ISVs will bother to port to ItanicF even if it's the only platform, especially given that they have a fair? amount of time to wait and see how many customers will migrate.t    L "IBM does not have a good track record of laying out a strategy and stickingK with it.  IBM programs typically last two years or less before the strategynI changes.  This makes it difficult for IBM's customers to plan long-term."t  E Especially at this point in time, making such a statement to an AlphahE customer is likely to result in grievous bodily harm.  Not to mention C instant loss of business.  Is it truly possible that Compaq doesn'ts understand this?    G "With IBM decommitting AIX on Intel and HP's UNIX business fading fast,hL events are coming together to make Compaq's Tru64 UNIX the future leader for UNIX on Intel."c  I And, as I said before, monkeys may come flying out of my butt.  But don't0 count on it.    J "Compaq welcomes IBM's Intel platform customers who were 'dropped' by this move."  
 Both of them?3   - bill   ------------------------------  $ Date: Sun, 2 Sep 2001 13:29:35 -0500C From: "Craig A. Berry" <craig.berry@nospam.SignalTreeSolutions.com>hD Subject: Re: Compiled Languages (Was: e: My VMS Wish List (features)1 Message-ID: <a05101002b7b725b14b37@[172.16.52.1]>h   Nick Maclaren wrote:  J >In article <craig.berry-E1365B.09564730082001@newsrump.sjc.telocity.net>,F >"Craig A. Berry" <craig.berry@nospam.SignalTreeSolutions.com> writes: >|>a; >|> >	I'm not interested in Perl.  The syntax is insanity.   >|>yH >|> The basic syntax is pretty much identical to any other language thatK >|> has curly braces and semicolons.  There are plenty of weird features, I K >|> grant you, but you don't have to use any more of them than you want to.h > @ >This is commonly stated, but is completely untrue.  The lack of@ >a proper syntactic specifation and thorough checking means that@ >you can drop into a fancy syntax by accident, with no hint that? >it is not doing what you intended.  This is one of the reasonsh= >that unused features are NOT harmless in any language unlesse? >they are cleanly separated from the parts of the language that- >you are intending to use. >,< >My first Perl program was 20 lines long, and I checked that> >every branch was taken correctly in both paths.  It then gave> >wrong answers as soon as I ran it on real data, because I had? >used an incorrect but undiagnosed syntax that just happened toy< >do what I was expecting in my tests.  Another language with# >similarly evil properties was SAS.4 >.; >Now, that was Perl 4, and Perl 5 is better.  But, from allb$ >accounts, not all that much better.    ? I have no quarrel with those who prefer Python to Perl, and I'mvC pleased to see there is an up-to-date Python implementation on VMS.tB But I can't let stand the suggestion that Perl's syntax is somehowC indeterminate.  Perl's grammar certainly can be and indeed has beenaD specified; for example, another Cambridge lad has put up a graphical representation of it here:E <http://simon-cozens.org/hacks/grammar.pdf>.  Plus if you are ever in D doubt about what the parser is doing with your script, you can get aB human-readable dump of opcodes, analogous to what you get from the= slightly-misnamed /MACHINE_CODE qualifier with the native VMS D compilers.  As far as checking and warning, this has been one of the? biggest growth areas in the 5.x releases of Perl (Perl 4 died a $ natural death about five years ago).  C It is true that the Perl grammar is sometimes said to involve smokeT= and mirrors, but this is a statement of frustration about theuE obscurity of the code in which the language is implemented, not about D the language itself.  That's why Perl 6, which is well under way, is> a complete rewrite from scratch.  The guy heading up internals@ development for Perl 6 is a VMS guru who appears to be doing his? development work under VMS; see last week's report of a running  interpreter prototype here:>G <http://archive.develooper.com/perl6-internals@perl.org/msg03459.html>.l? So the suggestion elsewhere in this thread that Perl is somehow E antithetical to the VMS way of doing things, if it was ever true (andrE I don't think it was), is rapidly becoming less so.  From the traffictE on the vmsperl mailing list, it appears that Perl's use on VMS is notr@ only lively and widespread but pretty well spans the government, business and education sectors.    Cheers,  Craig A. Berry   ------------------------------   Date: 2 Sep 2001 22:30:22 GMTe( From: nmm1@cus.cam.ac.uk (Nick Maclaren) Subject: Re: I hate Compaq0 Message-ID: <9mubtu$bu5$1@pegasus.csx.cam.ac.uk>  3 In article <aBwmRflV9nEZ@eisner.encompasserve.org>,j. Larry Kilgallen <Kilgallen@SpamCop.net> wrote: >.@ >>>Personally I have no desire to pay higher prices so operating? >>>systems will include a debugger that robust, since those are. >>>not problems I encounter. >> hA >> No higher prices needed - a debugger that robust would pay fortA >> itself before first release, by reducing the development costscB >> and time to release.  There is a lot of evidence for this, too.? >> So what you are actually paying is higher prices for a worse, >> product :-( > > >While that may be your "magic bullet", mine is strongly typed? >programming languages.  Different strokes for different folks.t  = You are missing the point.  SOMEONE has to debug the languagee@ run-time system that you are relying on and, if the language you= are using supports signals properly, that includes the signal ? handling.  And then there is the nasty problem of locating code > generation errors in complex, highly-optimised, often parallel> codes.  Whether produced by a strongly typed language compiler3 or by C - it makes little difference at that level.   > I favour strongly typed languages (or would, if there were any? decent ones around) when I get the choice.  But there are a LOTo< of circumstances where a lower level of debugging is needed,> though I agree that end-users like you will rarely need to get> that dirty.  Unfortunately, I am the expert of last resort (in& addition to my other tasks), and I do.  = In my long life of crime, I have implemented several languagew= run-time systems, worked for vendors and so on.  My viewpoint = is at least as much that of an implementor as that of a user.p< And I am telling you from actual experience that getting the> low-level debugging environment right speeds up and, even more@ importantly, improves the quality of operating systems, language8 environments and so on.  Such as the ones you are using.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QG, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679    ------------------------------  $ Date: Sun, 2 Sep 2001 15:40:35 -0700# From: "Tom Linden" <tom@kednos.com>o Subject: RE: I hate Compaq9 Message-ID: <CIEJLCMNHNNDLLOOGNJIKEJODDAA.tom@kednos.com>n   > -----Original Message-----1 > From: Nick Maclaren [mailto:nmm1@cus.cam.ac.uk]2* > Sent: Sunday, September 02, 2001 3:30 PM > To: Info-VAX@Mvb.Saic.Com  > Subject: Re: I hate Compaq >  > 5 > In article <aBwmRflV9nEZ@eisner.encompasserve.org>, 0 > Larry Kilgallen <Kilgallen@SpamCop.net> wrote: > > B > >>>Personally I have no desire to pay higher prices so operatingA > >>>systems will include a debugger that robust, since those are  > >>>not problems I encounter. > >> uC > >> No higher prices needed - a debugger that robust would pay forfC > >> itself before first release, by reducing the development costsiD > >> and time to release.  There is a lot of evidence for this, too.A > >> So what you are actually paying is higher prices for a worsee > >> product :-( > >e@ > >While that may be your "magic bullet", mine is strongly typedA > >programming languages.  Different strokes for different folks.e > ? > You are missing the point.  SOMEONE has to debug the languagenB > run-time system that you are relying on and, if the language you? > are using supports signals properly, that includes the signal2A > handling.  And then there is the nasty problem of locating codeR@ > generation errors in complex, highly-optimised, often parallel@ > codes.  Whether produced by a strongly typed language compiler5 > or by C - it makes little difference at that level.r > @ > I favour strongly typed languages (or would, if there were anyA > decent ones around) when I get the choice.  But there are a LOT4> > of circumstances where a lower level of debugging is needed,@ > though I agree that end-users like you will rarely need to get@ > that dirty.  Unfortunately, I am the expert of last resort (in( > addition to my other tasks), and I do.  < There is.  You can use PL/I.  You can download the kit from  freja.kednos.com   > ? > In my long life of crime, I have implemented several languageo? > run-time systems, worked for vendors and so on.  My viewpointm? > is at least as much that of an implementor as that of a user. > > And I am telling you from actual experience that getting the@ > low-level debugging environment right speeds up and, even moreB > importantly, improves the quality of operating systems, language: > environments and so on.  Such as the ones you are using. >  > 
 > Regards, > Nick Maclaren,, > University of Cambridge Computing Service,@ > New Museums Site, Pembroke Street, Cambridge CB2 3QG, England. > Email:  nmm1@cam.ac.uk1 > Tel.:  +44 1223 334761    Fax:  +44 1223 334679r >    ------------------------------  % Date: Sun, 02 Sep 2001 14:34:23 -0700d! From: Don Sykes <don@alphase.com>  Subject: Re: SMTP Que Stalls+ Message-ID: <3B92A5DF.275A6B48@alphase.com>e   Hoff,tH I downloaded the right self-extracting ECO, I think, and it extracts ok,E but when I try to upgrade it says I don't have TCPIP.   Here's what Ir get:   $ product install *   B %PCSI-E-APPLYTOERR, product DEC AXPVMS TCPIP, to which maintenance
 product DEC A0/ XPVMS TCPIP_ECO V5.0-113 applies, was not foundRE Terminating is strongly recommended.  Do you want to terminate? [YES]n/ %PCSI-E-S_OPCAN, operation cancelled by request C %PCSIUI-E-ABORT, operation terminated due to an unrecoverable error 	 condition<     $ PRODUCT SHOW PRODUCT TCPIP< ----------------------------------- ----------- ------------5 PRODUCT                             KIT TYPE    STATE < ----------------------------------- ----------- ------------9 DEC AXPVMS TCPIP V5.0-9             Full LP     Installedh< ----------------------------------- ----------- ------------   1 item found    > I didn't want to continue, til I was sure it was OK... is it ?   TIA  Donu     Hoff Hoffman wrote:- > Q > In article <3B7D64D9.39F504E9@alphase.com>, Don Sykes <don@alphase.com> writes:o > J >   Please turn off MIME and please avoid HTML when posting here.  Thanks! > ? > :...All message processing stops until I do the following:... E > : 1) delete the stalled message file from SYS$SPECIFIC:[TCPIP$SMTP]o$ > : 2) delete the entry from the que# > : 3) run "tcpip anal mail/repair"n > :....tE > :I'm running DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0w3 > :on a AlphaStation 250 4/266 running OpenVMS V7.2  > .. > L >   This looks suspiciously like a known (and fixed) bug in TCP/IP Services.I >   Please get to V5.0A with the ECO or (better) get to V5.1 and the ECO.g > P >  ---------------------------- #include <rtfaq.h> -----------------------------L >       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.comP >  --------------------------- pure personal opinion ---------------------------N >    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Sun, 02 Sep 2001 14:16:16 -0400l2 From: rdeininger@mindspring.com (Robert Deininger)% Subject: Re: Tunneling DECnet over IP.L Message-ID: <rdeininger-0209011416160001@user-2iveb47.dialup.mindspring.com>  B In article <1010901191546.61567B-100000@Ives.egh.com>, John Santos <JOHN@egh.com> wrote:o  - > On Sat, 1 Sep 2001, Robert Deininger wrote:.  J > > I remember getting a lot of ethernet errors (not just DECnet) on a VAXH > > 4600, when I was doing a lot of tests with DECRAM disks.  Whenever II > > INITed a RAM disk, the connection to other cluster members was up andsN > > down.  So it does seem possible to choke the ethernet on these systems.  IJ > > never tried the same on another VAX, for lack of extra memory.  Alphas# > > never showed the same behavior.9 > I > We used to see lots of that and the cluster losing and regaining quorum C > when doing heavy ethernet traffic (not just from the cluster, for @ > example a PC server doing backups of other PC's), but using anE > ethernet switch seemed to cure that.  However there may have been ahF > internode backup going on at the time.  (Alpha backing up one of the. > VAX'es disks to the Alpha's 8mm tape drive.) > C > P.S.  You could also be flooding the VAX'es QBUS if it is using akD > DELQA.  (I don't know what the through-put of the onboard EthernetG > chip is on some (all?) 4000's, but I think a QBUS is limited to aboutx > 1 MByte/sec.)r  D This Vax was sharing an switched ethernet port with 2 or 3 other VMSJ systems, all of which were pretty idle.  There was essentially no ethernetB traffic except SCS on this port.  The Vax 4600 itself wasn't doingI anything except initing the RAM disks.  That was enough to interfere withdH SCS.  This was using the onboard ethernet, not the DELQA.  I don't thinkC there could have been an issue with ethernet or Qbus contention.  I J decided the RAM disk driver was simply spending too much time at high IPL,H an keeping ethernet traffic from getting through.  Why a VAX should haveE to spend several seconds initing a small RAM disk, let alone blocking < interrupts, I don't know, but those were the symptoms I saw.   -- o Robert Deininger rdeininger@mindspring.com    ------------------------------  $ Date: Mon, 3 Sep 2001 07:52:29 +0800" From: "Kenneth" <best@hotmail.com>) Subject: Re: VMS 7.3 compatibility issues 0 Message-ID: <9mugkj$qlc1@imsp212.netvigator.com>  I I have scan all the executable in the system. However, some of the systemBG binaries will return me the "Potential Alpha Violation" (I am using VMSsI 7.2-1). And what Compaq reply me that this is only a freeware only, so...yL and they have certified the compatibility of VMS 7.2-1 on EV6. So the resultK is if it returns error, it doesn't necessarily an error, and if it does notoC returns an error, it does not mean it really have no error. And the * conclusion is the tool is not much useful.    6 "Main, Kerry" <Kerry.Main@compaq.com> wrote in messageL news:BE56C50EA024184DAF48F0B9A47F5CF4D49603@kaoexc01.americas.cpqcorp.net...6 > As a fyi, the program John is referring to is called > sys$system:SRM_CHECK.EXE >c< > Additional information can be found at: (Section 8 pg.8-1) >4L http://www.openvms.compaq.com/doc/73final/documentation/pdf/OVMS_73_REL_NOTE > S.pdf  >o
 > Regards, >a > Kerry Main > Senior Consultanti > Compaq Canada Corp.e > Professional Servicesa > Voice: 613-592-4660M > Fax  :  819-772-7036 > Email: Kerry.Main@Compaq.com >i >e > -----Original Message-----) > From: John Santos [mailto:JOHN@egh.com]n! > Sent: September 1, 2001 8:05 PMn > To: Info-VAX@Mvb.Saic.Comc+ > Subject: Re: VMS 7.3 compatibility issuest >f > $ > On Sun, 2 Sep 2001, Kenneth wrote: >iK > > I am using EV5 processor and want to upgrade to EV6 (21264a) processor.pJ > > Compaq says I need to recompile my C program with the C version 6.0 orL > > above. Does anyone out there has the problem in running your application > inG > > EV6 which is compile with C below 6.0? Or do you have encounter anyh thirdi4 > > party product which is not compatiable with EV6? > E > Normally, no you don't have to recompile.  However, there was a bug B > in some old compilers (actually in the GEM optimizer) that couldD > generate incorrect code sequences.  These sequences worked fine onF > EV5 and before, but could break on an EV6.  The newer compilers haveE > this bug fixed, so recompiling guarantees it won't bite you.  TherehE > is also a program you can run which will scan an image and tell youhH > if the invalid code sequences occur.  (It can get false positives fromI > data that looks like code.)  I can't remember the name of this program,t2 > but it should be in the V7.0-V7.2 release notes. > E > If you are only intending to run your programs on your new EV6, you C > can recompile with the right optimization switches and get betternC > performance.  (If you optimize for EV6 and run the resulting .EXEe@ > on EV5, it will work, but it might be slower than it would run% > un-optimized or optimized for EV5.)l >aC > Generally 3rd party products are compiled without optimization oriC > with generic optimizations, to get performance that is acceptable C > everywhere rather than great performance on some systems and pooraA > performance on others.  Something that is specifically intendediA > for high performance might be shipped either as source (compile6B > on-site) or as multiple object libraries optimized for differentC > systems, or as multiple .EXE's.  We usually ship object libraries. > for the most flexibility.I >l > --
 > John Santosh > Evans Griffiths & Hart, Inc. > 781-861-0670 ext 539   ------------------------------   Date: 2 Sep 2001 11:12:40 -0700:! From: hemanir@yahoo.com (Anamika)r$ Subject: VMS NT/win2000 similarities= Message-ID: <5130f039.0109021012.75d09861@posting.google.com>    Hi,e0   Lots of people might know what I am about say.= I recently was looking at NT/w2k system internals and noticed.? how similar it was to vms. The way the entire OS is structured,aA IRPs, sections, page handling etc., lots and lots of similarites.o= Not surprising since solomon and other VMS techies went to MS . and used stuff from their past VMS experience.6 Are there any resources that talk about more of this ?   -RSN   ------------------------------  $ Date: Sun, 2 Sep 2001 21:22:05 +03001 From: "_jussi" <jaakonaho@nospam.juhani.decus.fi>m( Subject: Re: VMS NT/win2000 similarities* Message-ID: <9mtt43$e4$1@news.kolumbus.fi>  8 > Are there any resources that talk about more of this ?  ; http://www.win2000mag.com/Articles/Index.cfm?ArticleID=4494o; http://www.win2000mag.com/Articles/Index.cfm?ArticleID=4652e and the last one not so much...e@ http://www.win2000mag.com/Articles/Index.cfm?ArticleID=4500&pg=2   _jussi   ------------------------------  $ Date: Sun, 2 Sep 2001 14:59:38 -0400+ From: "Main, Kerry" <Kerry.Main@compaq.com>o( Subject: RE: VMS NT/win2000 similaritiesR Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4D560AE@kaoexc01.americas.cpqcorp.net>  H And another pointer which includes the attached url's, plus a whole pile5 more from Mark Russinovich on NT/W2K design articles.w  & http://www.sysinternals.com/publ.shtml   Regards,  
 Kerry Main Senior Consultantn Compaq Canada Corp.e Professional Servicese Voice: 613-592-4660e Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----6 From: _jussi [mailto:jaakonaho@nospam.juhani.decus.fi] Sent: September 2, 2001 2:22 PMm To: Info-VAX@Mvb.Saic.Coma( Subject: Re: VMS NT/win2000 similarities    8 > Are there any resources that talk about more of this ?  ; http://www.win2000mag.com/Articles/Index.cfm?ArticleID=4494m; http://www.win2000mag.com/Articles/Index.cfm?ArticleID=4652e and the last one not so much...c@ http://www.win2000mag.com/Articles/Index.cfm?ArticleID=4500&pg=2   _jussi   ------------------------------   Date: 2 Sep 2001 21:30:05 -0500p+ From: young_r@encompasserve.org (Rob Young)h( Subject: Re: VMS NT/win2000 similarities3 Message-ID: <8IEAD7FLp59S@eisner.encompasserve.org>e  a In article <5130f039.0109021012.75d09861@posting.google.com>, hemanir@yahoo.com (Anamika) writes:n > Hi,r2 >   Lots of people might know what I am about say.? > I recently was looking at NT/w2k system internals and noticed A > how similar it was to vms. The way the entire OS is structured,rC > IRPs, sections, page handling etc., lots and lots of similarites.a? > Not surprising since solomon and other VMS techies went to MSh0 > and used stuff from their past VMS experience.8 > Are there any resources that talk about more of this ? >   = 	This periodically surfaces here.  Probably the best responset? 	(in my opinion) seen here was by Jamie Hanrahan in 1997.  I'veiF 	dropped this out here before, it can be found searching www.deja.com.< 	Towards the end, you can see where Jamie recommends reading= 	"Showstopper" and "Inside Windows NT" for more backgrounder.n) 	Still... this post is a classic!  Enjoy!t   				Robl   ---t  : From: Jamie Hanrahan, Kernel Mode Systems (jeh@cmkrnl.com)" Subject: Re: FX!32, affinity etc.  Newsgroups: comp.os.vmse Date: 1997/05/28 e  r  ( In article <338A488D.6FC7@videotron.ca>,1  jfmezei <"[nospam]jfmezei"@videotron.ca> writes:e) > David Cathey (Remove MX to mail) wrote:VF >> culmulated into OpenVMS.  The sad thing is Cutler had to capitulateA >> to a Windows assimilation of his theoretically sound base O/S.oB >> You'd think he would have taken the tenets of procedure calling >> standards with him as well. > D > I think it is quite pretentious of the VMSer to think that Cutler C > had so much design authority over NT and that he would have takeno > his VMS knowledge with him.T  F Not pretentitious at all.  We're simply aware of facts you apparently  haven't heard about.    F From the internals point of view there is utterly no question that NT  is VMS re-implemented.    F > Cutler was told to replace DOS with a real kernel over which Windows: > could run, over which teh windows API could run etc etc.  D This is off-topic, but you're incorrect here also.  Originally thereC wasn't even going to be a Win32 API - 16-bit Windows had not gotten B all that popular when NT was conceived.  NT was originally to be aD follow-on to OS/2, and a cooperative effort with IBM.  But somewhereF along the way Windows got pretty popular, IBM and MS split the beast, 7 IBM getting the OS/2 parts and MS keeping the NT parts.   J > If you look at the PSION PDA operating system (called EPOC), you'll alsoA > find many many similarities with VMS. Event Flags, InterprocessnJ > Mailboxes, shared memory between processes, process priorities, and evenG > a utility (SPY) which is the equivalent to SHOW SYS. Its IO system isy? > similar to VMS (an equivalent to $ASSIGN with the device nameeA > determining which driver to use, and $QIO which is more or lesss% > independant of the device itself.).c  B That's out at the UI and API level.  We're talking about internal  similarities.  c  A > NT is an WINDOWS operating system with modern operating systemsnA > services. They were implemented with the Windows API mentality.s5 > Stop thinking that NT is VMS with WINDOWS above it.o  C No, not "VMS with Windows above it", but a VMS-derived design with p" Windows above it, most certainly.   H > NT differs from DOS in that it has real operating system features, but+ > the later are not the exclusivity of VMS.-  C The "real operating system features" you speak of are at the UI andr; API level.  They are not the reasons we consider NT to be aq0 reimplementation of VMS at the internal level.    
 How about:  C The scheduler.  (process scheduler in VMS, thread scheduler in NT) mC 32 scheduling priorities, divided into the "real-time" (16-31) and ME "variable" (0-15) priority ranges.  identical preemption at ready by f> higher-priority threads; identical quantum and priority boost E implementations; identical CPU starvation avoidance mechanism to get hB out of priority inversion situations; a null thread for each CPU;  etc., etc.    6 Memory management.  0-7FFFFFFF is per-process, mostly C user-mode-accessible only; 80000000-FFFFFFFF is systemwide, mostly sC kernel-accessible only.  Functionally identical implementations of d paging vs. swapping. e  B I/O.  I could write a book (in fact, I am), but briefly, IRPs are E IRPs, UCBs are "device objects", CRBs are "controller objects", ADPs a= are "adapter objects", FDT routines are "dispatch routines", mG EXE$QIODRVPKT is IoStartPacket, StartIO routines are StartIO routines, iC fork routines are DPC routines, ASTs are APCs... etc., etc., etc., e etc., etc.    F Interrupt handling.  32 levels of interrupts (some simulated but this E is nevertheless the way the code is written).  IPLs on VMS, IRQLs on pG NT.  In order:  Passive level, APC (AST) Level, Dispatch (fork) level, dB then the IO hardware interrupts, then some "hardware maintenance" E functions like the hardware timer, IPI, power fail notification, and  $ HIGH_LEVEL to block all interrupts.   > Face it, JF, you're wrong.  Worse, you are writing not just in? misunderstanding but in ignorance of the facts.  Please go readrE _Showstopper_ and _Inside Windows NT_ (Custer) before opining furtheru on this subject. o  6 	--- Jamie Hanrahan, Kernel Mode Systems, San Diego CAB         Internet: jeh@cmkrnl.com (JH645)  CompuServe: 74140,2055  O drivers, internals, networks, applications, and training for VMS and Windows NTmD NT driver FAQ, links, and other information:  http://www.cmkrnl.com/  8 If you post a reply in news, please don't e-mail it too.   ------------------------------   End of INFO-VAX 2001.489 ************************