1 INFO-VAX	Fri, 05 Jul 2002	Volume 2002 : Issue 367       Contents:0 %NONAME-E-NOMSG ,MESSAGE NUMBER 00000002....????4 Re: %NONAME-E-NOMSG ,MESSAGE NUMBER 00000002....????4 Re: %NONAME-E-NOMSG ,MESSAGE NUMBER 00000002....????; RE: (OT) Carl Rowan and guns (was Re: A possible shift in t 
 .scn files1 Re: .scn files (Using the VAX Debugger with SCAN) 1 Re: .scn files (Using the VAX Debugger with SCAN) 1 Re: .scn files (Using the VAX Debugger with SCAN) 1 Re: .scn files (Using the VAX Debugger with SCAN) 1 Re: .scn files (Using the VAX Debugger with SCAN)  500 bytes long parameter< Re: Accessing the DCL recall buffer in a command procedure ? Advice on SCSI options sought + Re: DS10L Feedback - more for the pickin' ! + Re: DS10L Feedback - more for the pickin' !  Expect for VMS Re: Expect for VMS Re: Expect for VMS Re: Expect for VMS Re: Expect for VMS Re: Expect for VMS Re: Expect for VMS$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ RE: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications...$ RE: Fearless IPF Prognostications...$ Re: Fearless IPF Prognostications... Re: Last 3 weeks on c.o.v." Re: LGI_BRK_DISUSER and batch job? Re: Lynx & SSL< RE: New web-page dedicated to ports of PD software to OpenVM< Re: New web-page dedicated to ports of PD software to OpenVM9 New web-page dedicated to ports of PD software to OpenVMS = Re: New web-page dedicated to ports of PD software to OpenVMS = Re: New web-page dedicated to ports of PD software to OpenVMS  OpenVMS 7.2-1H1 (Differences) ! Re: OpenVMS 7.2-1H1 (Differences)  Re: Quorum discussion/questions 8 Re: System Disk Initialization Parameter recommendations8 Re: System Disk Initialization Parameter recommendations8 Re: System Disk Initialization Parameter recommendations8 Re: System Disk Initialization Parameter recommendations8 Re: System Disk Initialization Parameter recommendations8 Re: System Disk Initialization Parameter recommendations8 Re: System Disk Initialization Parameter recommendations+ Re: Three HP Press releases (via Bloomberg)  Re: UAF questions  Re: UAF questions / Re: Using GNU C on OpenVMS FAQ (Looking for it) / Re: Using GNU C on OpenVMS FAQ (Looking for it)   Re: VMS IO up to the level of PC Re: VMS port delayed! P Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar P Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar P Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar P Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar P Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar   F ----------------------------------------------------------------------   Date: 5 Jul 2002 03:26:41 -0700 $ From: tarunm_2000@yahoo.com (tmrana)9 Subject: %NONAME-E-NOMSG ,MESSAGE NUMBER 00000002....???? = Message-ID: <e8f11ce1.0207050226.6292502f@posting.google.com>   F An application tries to open file on vax5.5-2 , but on status check ofC that file ,its  returncode is 2 i.e %NONAME-E-NOMSG ,MESSAGE NUMBER  00000002....????.   ! Any Pointer to what it is....!!!!      Thanks   Tarun Maharana   ------------------------------   Date: 5 Jul 2002 10:43:47 GMT 3 From: gartmann@immunbio.mpg.de (Christoph Gartmann) = Subject: Re: %NONAME-E-NOMSG ,MESSAGE NUMBER 00000002....???? 0 Message-ID: <ag3t93$gdj$1@n.ruf.uni-freiburg.de>  d In article <e8f11ce1.0207050226.6292502f@posting.google.com>, tarunm_2000@yahoo.com (tmrana) writes:G >An application tries to open file on vax5.5-2 , but on status check of D >that file ,its  returncode is 2 i.e %NONAME-E-NOMSG ,MESSAGE NUMBER >00000002....????. > " >Any Pointer to what it is....!!!!  G The facility is unknow. If it is the status returned to the program via J an open statement in a higher programming language, see the manual of thisN language for the meaning. In Pascal it would just mean "error during open" ;-)   Regards,    Christoph Gartmann   H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, Germany                                           |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  % Date: Fri, 05 Jul 2002 07:57:41 -0400 1 From: Michael Austin <maustin@firstdbasource.com> = Subject: Re: %NONAME-E-NOMSG ,MESSAGE NUMBER 00000002....???? 2 Message-ID: <3D2589B5.C8FD7978@firstdbasource.com>  
 tmrana wrote:  > H > An application tries to open file on vax5.5-2 , but on status check ofE > that file ,its  returncode is 2 i.e %NONAME-E-NOMSG ,MESSAGE NUMBER  > 00000002....????.  > # > Any Pointer to what it is....!!!!  >  > Thanks > Tarun Maharana    7 could be a priv problem where you do not have privs set  correctly.   --   Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163 7 Sr. Consultant            http://www.firstdbasource.com                            + http://www.firstdbasource.com/donation.html / 704-947-1089 (Office)     704-236-4377 (Mobile)    ------------------------------  $ Date: Fri, 5 Jul 2002 09:41:01 -0400* From: WILLIAM WEBB <WWEBB1@email.usps.gov>D Subject: RE: (OT) Carl Rowan and guns (was Re: A possible shift in t- Message-ID: <0033000071379291000002L012*@MHS>   H =0A> > And your missing situation number three: the interim period betw= een < > > such a law taking effect and a large reduction in crime.  " I'm calling this one out as bogus.  ? Where have crime rates dropped after handguns have been banned?   , NYC?  Washington DC?  Most cities in the UK?   Don't think so.    WWWebb   -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNET # Sent: Friday, July 05, 2002 9:32 AM B To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNETD Subject: RE: [OT] Carl Rowan and guns (was Re: A possible shift in t    5 "Bill Todd" <billtodd@metrocast.net> wrote in message H news:<EYaV8.559231$%y.36742574@bin4.nnrp.aus1.giganews.com>... > "Alan = E.2 Feldman" <spamsink2001@yahoo.com> wrote in message9 > news:b096a4ee.0207041956.7e875e46@posting.google.com... 9 > > "Bill Todd" <billtodd@metrocast.net> wrote in message @ >  news:<4_8U8.81586$Ca2.4755216@bin2.nnrp.aus1.giganews.com>...A > > > "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in message = > > > news:b096a4ee.0207011838.3539685d@posting.google.com...  >  > ...  >  > > > > Am I missing something
 > > > > here?  > > > H > > > Yes.  And since it's been called to your attention already, eithe= r  >  you're a @ > > > bit slow or not really making the effort to understand it. > >  > > # > > Let's refrain from the insults.  > H > Feel free to act as you see fit; as for me, I'll decide on a case-by-= caseH > basis.  When you ask a question, be prepared for a straight answer:  = if youH > hadn't attempted to be cute in asking it, I might have been less blun= t in > responding to it.     , I don't recall being "cute". Sorry if I was.    H > > > He advocates a change that (he believes) would eliminate the perc= eived  >  need H > > > to defend one's home with a gun.  *In the absence of that change*=  (i.e., H > > > under current conditions), he clearly himself does feel such a ne= ed. H > > > There's no contradiction (because they're two different situation= s),  >  hence > > > no hypocrisy.  > >  > > E > > Uh, perceived need is not equivalent to real need. People who are H > > paranoid may have perceived needs for owning a gun in situations wh= ere   > > one doesn't really need one. > > H > > And your missing situation number three: the interim period between=  H > > such a law taking effect and a large reduction in crime. During tha= t H > > period, it would still be prudent to own a gun. And Mr. Rowan clear= lyH > > says that when faced with such dangers he would own a gun. Therefor= e,? > > he says that he would own a gun during that interim period.  > H > You're really full of shit, you know.  Or completely ignorant of simp= leD > logic.  What Mr. Rowan said is what he said, no more, no less, andH > attempting to read into it support for your own viewpoint rather than=  stickC > to it literally is the mark of demagoguery rather than intellect.     E This argument is bogus. You rule out simple conclusions based on what A he said simply because he didn't say them. I find that illogical.    Demogoguery? From Webster.com: E 1 : a leader who makes use of popular prejudices and false claims and  promises in order to gain power B 2 : a leader championing the cause of the common people in ancient times   E I am not trying to gain power and I am not in ancient times. And just C to be clear, I am not advocating gun control nor am I trying to say @ anything directly against it. I am only referring to Mr. Rowan's> remarks and am saying that at the very list there is a hint of0 hypocrisy and that they are not self-consistent.    H > The way to find out Mr. Rowan's view on whether he feels such an inte= rim H > situation would be necessary for him to address and if so how he woul= d  > address it is to ask him.     C The interim period is not optional. There *will* be a period during D which gun ownership by all private citizens is illegal and the crimeF rate will still be unacceptably high. And since there *will* be such aE period, and Mr. Rowan clearly stated he would protect his family when D the crime rate is "too high", he'll be owning a gun illegally duringA that period. If you think that there might not be such an interim  period, then you are an idiot.    ) > Lacking that, calling your suppositions F > 'evidence' of the hypocrisy you claim exists (as opposed to multipleE > alternate interpretations ranging from simply disagreeing with your H > assessment of the 'interim' conditions to the possible naivete alread= y H > mentioned) demonstrates considerably less about him than it does abou= t you. >  > - bill     That's your opinion.  A And since you insist on uncalled for insults Mr. Know-it-all, why / should everyone, or even anyone, listen to you?      Disclaimer: JMHO Alan E. Feldman  afeldman gfigroup com=   ------------------------------  $ Date: Fri, 5 Jul 2002 15:04:38 +0530# From: "Vivek Soni" <visoni@bmc.com>  Subject: .scn files / Message-ID: <uiapu0t388oja7@corp.supernews.com>    Hi,   I Encountered one more type of files (.SCN files). These were programmed in  1988.   D These file have Token sections,  Type section,  Declaration SECTION,$ PROCEDURE SECTION and MACRO Section.  J PROCEDURE SECTIONS contians the external declarations of functions defined in other .scn files.  L The functions defined in this file ( Called MACRO's here) are defined in the MACRO section.  1 PROCEDURE PARSE( file_name: varying STRING(256));  DECLARE status .  .  . 1 These functions get called from the .C files like    parse(&argument);   > Obviously the OpenVMS debugger cannot debug in to these files.  = I need more info on these file.  Can anybody help me on this.    Thanks Vivek    ------------------------------   Date: 5 Jul 2002 05:19:32 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen) : Subject: Re: .scn files (Using the VAX Debugger with SCAN)3 Message-ID: <bWTmk+YJq$o4@eisner.encompasserve.org>   U In article <uiapu0t388oja7@corp.supernews.com>, "Vivek Soni" <visoni@bmc.com> writes:  > Hi,  > K > Encountered one more type of files (.SCN files). These were programmed in  > 1988.  > F > These file have Token sections,  Type section,  Declaration SECTION,& > PROCEDURE SECTION and MACRO Section. > L > PROCEDURE SECTIONS contians the external declarations of functions defined > in other .scn files. > N > The functions defined in this file ( Called MACRO's here) are defined in the > MACRO section. > 3 > PROCEDURE PARSE( file_name: varying STRING(256));  > DECLARE status > .  > .  > . 3 > These functions get called from the .C files like  >  > parse(&argument);  > @ > Obviously the OpenVMS debugger cannot debug in to these files.  D Certainly the VMS debugger _can_ debug these files.  Try the command& SET LANGUAGE SCAN in the VAX debugger.   ------------------------------  $ Date: Fri, 5 Jul 2002 16:44:54 +0530# From: "Vivek Soni" <visoni@bmc.com> : Subject: Re: .scn files (Using the VAX Debugger with SCAN)/ Message-ID: <uiavpvtp6141a2@corp.supernews.com>   F Where can I get some basic documents for SCAN files and how does there work....   Thanks in Advance  Vivek   , Vivek Soni <visoni@bmc.com> wrote in message) news:uiavdqnc4et174@corp.supernews.com... " > I have set the following things. >  > SET LANGUAGE SCAN  >  > SET MODULE PARSE > 1 > I have also SET MODULE  for other SCAN modules.  > J > When I try to set scan modules it gives me a mesage that source code for& > this module not currently available.5 > Where as I have the source along with the C source.  > ! > Parse Debug list says its "yes"  > / > But when I try to step into the module it say  > E > %DEBUG-I-SRCNOTCURAV, source code for line xxxx in module PARSE not ) > currently available to PARSE\%line xxxx  >  > Vivek  > : > Larry Kilgallen <Kilgallen@SpamCop.net> wrote in message/ > news:bWTmk+YJq$o4@eisner.encompasserve.org... @ > > In article <uiapu0t388oja7@corp.supernews.com>, "Vivek Soni" > <visoni@bmc.com> writes:	 > > > Hi,  > > > L > > > Encountered one more type of files (.SCN files). These were programmed > in > > > 1988.  > > > J > > > These file have Token sections,  Type section,  Declaration SECTION,* > > > PROCEDURE SECTION and MACRO Section. > > > H > > > PROCEDURE SECTIONS contians the external declarations of functions	 > defined  > > > in other .scn files. > > > K > > > The functions defined in this file ( Called MACRO's here) are defined  in > the  > > > MACRO section. > > > 7 > > > PROCEDURE PARSE( file_name: varying STRING(256));  > > > DECLARE status > > > .  > > > .  > > > . 7 > > > These functions get called from the .C files like  > > >  > > > parse(&argument);  > > > D > > > Obviously the OpenVMS debugger cannot debug in to these files. > > H > > Certainly the VMS debugger _can_ debug these files.  Try the command* > > SET LANGUAGE SCAN in the VAX debugger. >  >    ------------------------------  $ Date: Fri, 5 Jul 2002 16:38:24 +0530# From: "Vivek Soni" <visoni@bmc.com> : Subject: Re: .scn files (Using the VAX Debugger with SCAN)/ Message-ID: <uiavdqnc4et174@corp.supernews.com>     I have set the following things.   SET LANGUAGE SCAN    SET MODULE PARSE  / I have also SET MODULE  for other SCAN modules.   H When I try to set scan modules it gives me a mesage that source code for$ this module not currently available.3 Where as I have the source along with the C source.    Parse Debug list says its "yes"   - But when I try to step into the module it say   C %DEBUG-I-SRCNOTCURAV, source code for line xxxx in module PARSE not ' currently available to PARSE\%line xxxx    Vivek   8 Larry Kilgallen <Kilgallen@SpamCop.net> wrote in message- news:bWTmk+YJq$o4@eisner.encompasserve.org... > > In article <uiapu0t388oja7@corp.supernews.com>, "Vivek Soni" <visoni@bmc.com> writes: > > Hi,  > > J > > Encountered one more type of files (.SCN files). These were programmed in	 > > 1988.  > > H > > These file have Token sections,  Type section,  Declaration SECTION,( > > PROCEDURE SECTION and MACRO Section. > > F > > PROCEDURE SECTIONS contians the external declarations of functions defined  > > in other .scn files. > > L > > The functions defined in this file ( Called MACRO's here) are defined in the  > > MACRO section. > > 5 > > PROCEDURE PARSE( file_name: varying STRING(256));  > > DECLARE status > > .  > > .  > > . 5 > > These functions get called from the .C files like  > >  > > parse(&argument);J > >0B > > Obviously the OpenVMS debugger cannot debug in to these files. >MF > Certainly the VMS debugger _can_ debug these files.  Try the command( > SET LANGUAGE SCAN in the VAX debugger.   ------------------------------   Date: 5 Jul 2002 08:19:48 -0600t- From: Kilgallen@SpamCop.net (Larry Kilgallen)A: Subject: Re: .scn files (Using the VAX Debugger with SCAN)3 Message-ID: <ymdddZASiqTH@eisner.encompasserve.org>r  U In article <uiavdqnc4et174@corp.supernews.com>, "Vivek Soni" <visoni@bmc.com> writes:h" > I have set the following things. >  > SET LANGUAGE SCANe >  > SET MODULE PARSE > 1 > I have also SET MODULE  for other SCAN modules.u > J > When I try to set scan modules it gives me a mesage that source code for& > this module not currently available.5 > Where as I have the source along with the C source.e > ! > Parse Debug list says its "yes"e > / > But when I try to step into the module it say. > E > %DEBUG-I-SRCNOTCURAV, source code for line xxxx in module PARSE not.) > currently available to PARSE\%line xxxx   2 Are you sure the SCAN source was compiled /DEBUG ? Have you SET SOURCE properly ?  D Those aspects are no different for SCAN than for any other language.  ? SET EVENT_FACILITY, however, is different, since the definitionE7 of events varies according to the programming language.n   ------------------------------   Date: 5 Jul 2002 08:20:51 -0600F- From: Kilgallen@SpamCop.net (Larry Kilgallen).: Subject: Re: .scn files (Using the VAX Debugger with SCAN)3 Message-ID: <fEwBU1qBq0Iq@eisner.encompasserve.org>M  U In article <uiavpvtp6141a2@corp.supernews.com>, "Vivek Soni" <visoni@bmc.com> writes: H > Where can I get some basic documents for SCAN files and how does there
 > work....  F On the Freeware disks that come with the delivery of the VMS operating system.    ------------------------------  % Date: Fri, 05 Jul 2002 11:37:15 +0200u+ From: Goran Domazet <goran.domazet@etna.hr>i! Subject: 500 bytes long parametery8 Message-ID: <evpaiu4kd2mf2vcokr7uheb2v30u7fkdkd@4ax.com>  C Is there a way to use 500 bytes long parameter in VMS (limit is 256t bytes, right?).t  D I need to execute Unix command via rsh (its VMS equivalent), and the- parameter it expects is about 500 bytes long.r   ------------------------------   Date: 5 Jul 2002 07:00:56 -0600oB From: clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley)E Subject: Re: Accessing the DCL recall buffer in a command procedure ? 3 Message-ID: <yEAC81bJVoRJ@eisner.encompasserve.org>    In article <Pine.LNX.4.21.0207042107570.1479-100000@irys.stanpol.com.pl>, "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl> writes:% > On 4 Jul 2002, Simon Clubley wrote:p >  >  Nothing hard with PIPE:: > $ LOGOUT:==PIPE RECALL/OUT=SYS$LOGIN:RECALL.TXT ; LOGOUT > B >  You *always* must substitute 'someway' LOGOUT command, be aware! > the LOGOUT substitutes :) (EOJ)a >   G D*mn. I never considered using pipe instead of command procedures, evenSI though I use pipe for other things. I guess it's true that you can carbon-O date how long people have been using VMS by how they go about doing things. :-)j Thanks, Gotfryd.   >>+ and restore command history  > ) >  Sure, requires one pseudocommand like:t0 > $ HI*STORY:=RECALL/INPUT=SYS$LOGIN:/RECALL.TXT > ? > - but you can comment this as a "great feature, you normally i > have clean recall buffer" ;) >   $ Yes, but that's not automatic... :-)   Simon.   -- !B Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       + Microsoft: The Lada of the computing world.t   ------------------------------  % Date: Fri, 05 Jul 2002 12:43:40 +0100E4 From: John Laird <john@laird-towers.freeserve.co.uk>& Subject: Advice on SCSI options sought8 Message-ID: <62vaiuguo5cfknt2t8r93svn10167ng9os@4ax.com>  J I have a pile of salvaged equipment surrounding me here in my home office,I which it is hoped may be useful in upgrading an existing system, which weEE developed and installed many years ago for a client.  Although hardly E cutting edge stuff, it is however hardware which we have by and largehL bypassed over the years and hence are not wholly familiar with.  Any help or! advice would be much appreciated.i  I Currently, the system runs on a pair of Microvax 3100/96s (gamely runningsL 6.2 for a variety of reasons) arranged as a master/slave hot-standby pair ofH standalone nodes.  Disk capacity is about 8Gb from RZ26s and 28s mountedH either internally or in BA350s.  We do not use any RAID techniques - theI slave is a full mirror of the master system, providing perfectly adequatepJ resiliency.  At busy times, the machine is maxed out either on CPU or diskI I/O (memory is adequate at 512Mb and 256Mb).  We would like to improve on-' both.  (At minimal cost, of course ;-))   K We have acquired a pair of Alphaserver 800 5/333 machines, which apart fromaG a necessary memory upgrade (to 1Gb most probably), have *nearly* enough?K storage options and boxes.  Nearly...  They were an NT cluster - each has a L KZPSA FWD SCSI adapter, connected to a single HSZ50 and thence to two BA356sH with 8-bit personality modules and 4x RZ29L-VAs.  I assume NT clusteringK supported some style of dual-hosted or dual-pathed disks.  There is also anuJ internal RZ28 (SCA variant).  Clearly, we have adequate disk capacity, butJ the single HSZ is a problem.  (We will not SCSI cluster the two machines.)   I see three possibilities:  L 1.  Acquire another 2nd user HSZ50 (plus writeback cache) and have identicalG setups.  Is this likely to be costly, either in purchase or maintenance/ terms ?h  I 2.  Live with an asymmetric configuration where one machine has the HSZ50-K and the other will need a bus convertor for the BA box (a DWZZx ?) directly1H connected.  I assume such a device might be a little harder to find, but cheaper than an HSZ50 ?   D 3.  Decide the external narrow disks might still prove a performanceL bottleneck and look at the internal option.  The onboard Qlogic card is alsoK fast-wide and there are a total of 4 slots in the drive cage.  I can sourceaK 2nd user Seagate ST19171WC 9Gb disks for the grand sum of $30 equivalent inDF quantities of 10 or more.  My concerns would be the performance of theF Qlogic vs the KZPSA/HSZ50 (although the former is working with devicesJ capable of twice the transfer rate), and the perennial issue of non-CompaqG drives, firmware revisions, etc.  Any opinions on either worry would bep greatly appreciated.    K As an aside, any ideas on why each Alpha had 3 Ethernet cards, one supplied6H DE450 and two DE500s ?  In the bundle of cables, I have a short piece ofB thinwire, terminated at both ends, which I suspect was a node-nodeH connection for clustering.  Perhaps the machines also had a simultaneous intranet/extranet presence...C  	 Many TIA.n     	John.   ------------------------------  % Date: Fri, 05 Jul 2002 11:02:21 +0200 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>o4 Subject: Re: DS10L Feedback - more for the pickin' !' Message-ID: <3D25609D.BFB2B9AE@aaa.com>.  . What's the added cost for shipment to Sweden ?$ What whould an extra KZPBA-CA cost ?  > One of thoses would nicely replace my DEC AXP 2000 modell 300,' who is becoming more and more tired...    = B.t.w, there shouln't be any problem with the hobbyist licensn on this box, would there ?   Best Regards Jan-Erik Sderholm.d   "Island (hpaq.net)" wrote: > G > Anyone willing to offer testimonial on the DS10L's we are selling ???= > # > We still have a quantity in stock   > DS10L 466Mhz EV6 CPU 2MB cache > 512MB Memory (2 x 256) > 40GB ATA100 7200RPM Disk > CDROM and Floppy > Dual 10/100 Ethernet > Power Cord >  > $950 >  > --   ------------------------------   Date: 5 Jul 2002 05:09:34 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen)a4 Subject: Re: DS10L Feedback - more for the pickin' !3 Message-ID: <0i15v$o5LRYL@eisner.encompasserve.org>   c In article <ui9qgp9f165ea4@news.supernews.com>, "Island (hpaq.net)" <dbturner@islandco.com> writes: G > Anyone willing to offer testimonial on the DS10L's we are selling ???   . I will offer testimony regarding your math :-)   > 512MB Memory (2 x 256)   > 1GB Memory (2 x 256)   > 512MB Memory (2 x 256)   ------------------------------  % Date: Fri, 05 Jul 2002 10:42:20 +0200 % From: Dominik Madon <dominik@acm.org>  Subject: Expect for VMS % Message-ID: <o2vznx6in0z.fsf@acm.org>e  D I'm looking for an expect-like program for OpenVMS. Did anyone knows? if such a tool exists on OpenVMS (either based on Perl, Tcl, or 
 whatever)?     Thanks for help, Dominik    ------------------------------  % Date: Fri, 05 Jul 2002 11:14:46 +0200 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>a Subject: Re: Expect for VMSi' Message-ID: <3D256386.6F603019@aaa.com>   2 What would a tool look like to be "expect-like" ?? What is "Expect" ??n   Jan-Erik Sderholm.i   Dominik Madon wrote: > F > I'm looking for an expect-like program for OpenVMS. Did anyone knowsA > if such a tool exists on OpenVMS (either based on Perl, Tcl, or  > whatever)? >  > Thanks for help,	 > Dominikc   ------------------------------  % Date: Fri, 05 Jul 2002 11:01:30 +0100y From: Roy Omond <Roy@Omond.net>  Subject: Re: Expect for VMSb) Message-ID: <3D256E79.5B325F08@Omond.net>r   Dominik Madon wrote:  F > I'm looking for an expect-like program for OpenVMS. Did anyone knowsA > if such a tool exists on OpenVMS (either based on Perl, Tcl, or  > whatever)?   Use Kermit.  It's free for VMS.   	 Roy Omondr Blue Bubble Ltd.   ------------------------------   Date: 5 Jul 2002 10:16:15 GMTy From: phn@icke-reklam.ipsec.nu Subject: Re: Expect for VMSi) Message-ID: <ag3rlf$spj$2@nyheter.crt.se>o  ' Jan-Erik Sderholm <aaa@aaa.com> wrote: 4 > What would a tool look like to be "expect-like" ?? > What is "Expect" ??w   See  http://expect.nist.gov/ n  F basically it's a scriptlanguage that can emulate a user at a terminal.   Qoting from expect.nist.gov :    "Expect is a tool for automating interactive applications such as telnet, ftp, passwd, fsck, rlogin, tip, etc. Expect really makes this stuff trivial. Expect is also useful for testing these same applications. And by adding Tk, you can also wrap interactive applications in X11 GUIs. l   Expect can make easy all sorts of tasks that are prohibitively difficult with anything else. You will find that Expect is an absolutely invaluable tool - using it, you will be able to automate tasks that you've never even thought of before - and you'll be able to do this automation quicklyn and easily.  "s   > Jan-Erik Sderholm.h   > Dominik Madon wrote: >> lG >> I'm looking for an expect-like program for OpenVMS. Did anyone knows B >> if such a tool exists on OpenVMS (either based on Perl, Tcl, or
 >> whatever)?m >> v >> Thanks for help,c
 >> Dominik   --   Peter Hkanson         u7         IPSec  Sverige      ( At Gothenburg Riverside ) J            Sorry about my e-mail address, but i'm trying to keep spam out,; 	   remove "icke-reklam" if you feel for mailing me. Thanx.,   ------------------------------  % Date: Fri, 05 Jul 2002 13:48:47 +0200x% From: Dominik Madon <dominik@acm.org>d Subject: Re: Expect for VMS % Message-ID: <o2vfzyyiee8.fsf@acm.org>   ! : Use Kermit.  It's free for VMS.R  B I'd like to make application test scripts (but need something more7 flexible than DTM). Is kermit enough flexible to do so?h     Dominikk   ------------------------------  % Date: Fri, 05 Jul 2002 12:55:29 +0100t From: Roy Omond <Roy@Omond.net>o Subject: Re: Expect for VMSh) Message-ID: <3D258930.87A4A8CF@Omond.net>-   Dominik Madon wrote:  # > : Use Kermit.  It's free for VMS.r > D > I'd like to make application test scripts (but need something more9 > flexible than DTM). Is kermit enough flexible to do so?i  . Well, it's only my opinion, but I'd say "yes".  < Kermit is one of the world's best pieces of software *ever*.  	 Roy Omond  Blue Bubble Ltd.   ------------------------------  % Date: Fri, 05 Jul 2002 13:51:15 +0200 % From: Dominik Madon <dominik@acm.org>e Subject: Re: Expect for VMS % Message-ID: <o2vbs9miea4.fsf@acm.org>    : What is "Expect" ??c  > It's a tool used to run "interactive" scripts like establish aA connection (telnet, ppp, and so on). It's used also in Dejagnu (aR- framework to test applications based on Tcl).      DominikR   ------------------------------   Date: 5 Jul 2002 08:32:43 GMT ( From: nmm1@cus.cam.ac.uk (Nick Maclaren)- Subject: Re: Fearless IPF Prognostications... 0 Message-ID: <ag3ljb$t4l$1@pegasus.csx.cam.ac.uk>  B In article <Rf8V8.124345$Ca2.6693380@bin2.nnrp.aus1.giganews.com>,) Bill Todd <billtodd@metrocast.net> wrote:n= >"Terry C. Shannon" <terryshannon@attbi.com> wrote in messager9 >news:y7%U8.237862$R61.87822@rwcrnsc52.ops.asp.att.net...c >mI >> In 1988 the VAXstar did in fact undergo a price reduction from $10K to  >$5K.tJ >> VAXstations didn't start out that cheap, though... cheapest pre-VAXstar >wasM >> the VAXstation II/RC (with the epoxied backplane) which IIRC was somewhereo >> between $15K and $20K.e >eK >One should, however, place this in the proper perspective.  In 1987 when I M >shopped around for a fairly minimal 8088 (not 80286, let alone 80386) system M >it ran close to $2K; today, a fairly minimal PC system costs about $600.  SosK >if a minimal VAXstation cost $15K - $20K at about the same time (let aloneIF >the $5K it dropped to soon thereafter), a price today of $5K would beL >consistent with the falling prices across the industry over the intervening >time.  G Back in 1979, I said that workstations would take off when they droppedlF to 5,000 pounds.  And I was spot on.  Of course, I didn't predict WHENH that would happen or WHICH ones would reach the limit first - but, hey!, we can't all be perfect :-)3  B Nowadays, 5,000 dollars/pounds will generate a lot of sales IF theF buyers want the machine - e.g. it delivers a lot of oomph, catches theD zeitgeist, or people believe that it is the first of a new era.  ButC few people (even in commerce) will spend that just to try somethingC out.  B And remember that I am talking about complete system cost.  Dell'sC current Itanium server range starts at $15,000, but a viable systemcE will set you back $25,000 or $35,000 if you want to do serious serverR2 application development.  Hence the sales figures.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679)   ------------------------------   Date: 5 Jul 2002 05:17:19 -0600h- From: Kilgallen@SpamCop.net (Larry Kilgallen)r- Subject: Re: Fearless IPF Prognostications... 3 Message-ID: <MKftY4E5R3xV@eisner.encompasserve.org>-  [ In article <ag3ljb$t4l$1@pegasus.csx.cam.ac.uk>, nmm1@cus.cam.ac.uk (Nick Maclaren) writes:b  D > And remember that I am talking about complete system cost.  Dell'sE > current Itanium server range starts at $15,000, but a viable systemiG > will set you back $25,000 or $35,000 if you want to do serious serverO > application development.  @ I would think a higher price should only apply for deployment --; serious _development_ can be done on a machine much smallerh# than the one needed for deployment.e  @ Or is this a comment specific to the operating systems currently running on Itanium ?   ------------------------------   Date: 5 Jul 2002 11:05:20 GMTo( From: nmm1@cus.cam.ac.uk (Nick Maclaren)- Subject: Re: Fearless IPF Prognostications...C0 Message-ID: <ag3uhg$aep$1@pegasus.csx.cam.ac.uk>  3 In article <MKftY4E5R3xV@eisner.encompasserve.org>,r/ Kilgallen@SpamCop.net (Larry Kilgallen) writes:w3 |> In article <ag3ljb$t4l$1@pegasus.csx.cam.ac.uk>, - |> nmm1@cus.cam.ac.uk (Nick Maclaren) writes:n |>  G |> > And remember that I am talking about complete system cost.  Dell'stH |> > current Itanium server range starts at $15,000, but a viable systemJ |> > will set you back $25,000 or $35,000 if you want to do serious server |> > application development.h |> cC |> I would think a higher price should only apply for deployment -- > |> serious _development_ can be done on a machine much smaller& |> than the one needed for deployment. |>  C |> Or is this a comment specific to the operating systems currentlyk |> running on Itanium ?m  ? Partly.  That is the reason that you would need more memory and ; (especially) disk space.  But the real reason is different.e  A You CANNOT do serious server work for a multi-CPU SMP system on a B single (or even dual) system, as many critical issues are entirelyE different.  This isn't just tuning, but whether the software actuallyeA works.  And a lot of experience is that certain RAS problems grownC exponentially with the number of processors (and sometimes even theo@ memory available).  And, clearly, you can't even start doing any' useful tuning on a much smaller system.   @ This was learnt back in the 1960s and 1970s and has been largely? forgotten since, especially by the community that believes thatg? intermittent or unrepeatable problems should be ignored as just > the sort of thing that happens.  Being old-fashioned, I regardC testing and tuning as part of development, and not just an optionalo= extra that you get around sometime after release to if enoughx customers make a fuss.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679V   ------------------------------  # Date: Fri, 05 Jul 2002 11:31:09 GMTi# From: "John Smith" <a@nonymous.com>a- Subject: Re: Fearless IPF Prognostications... G Message-ID: <1sfV8.23552$zGH.9126@news01.bloor.is.net.cable.rogers.com>.  5 "Bill Todd" <billtodd@metrocast.net> wrote in messageA= news:9M8V8.277203$_j6.13748425@bin3.nnrp.aus1.giganews.com...a >c ><L > Perhaps you should have made a greater effort to understand what John saidL > rather than merely attempted to be witty:  he was not questioning the $800J > per-CPU burden you quoted for the existing run rate, but suggesting thatK > this run rate would increase (and thus the per-CPU burden decrease) under7& > the changed conditions he described. >zH > Even if the $800/Alpha CPU you quote has some basis in reality (which, given7F > the considerable variation in annual Alpha development cost overhead between K > Winkler's $300m figure and Marcello's $150m figure, plus exactly what youtG > may be including in 'platform integration cost' above that might wellr apply-K > to *any* processor used, even Itanic, where Compaq designed the platform,@ isK > subject to some question), John seemed to be assuming - I would say quiteDH > reasonably - that even just $50m in additional Alpha advertising would haveJ > increased Alpha sales, and hence the run rate, significantly, not merely duepD > to the increased ad exposure itself but due to a major increase in	 perceivedJF > platform viability had Compaq started to demonstrate any such actual  > interest in making it succeed. >.   Well put. Thank you.  < The figure bandied about c.o.v. recently for Alpha/VMS/Tru64L advertising/marketing in the recent past has been somewhere between $10-20MMH annually. Compaq/HP, I'm sure, would dispute that, saying that all theirB other 'brand-awareness' advertising contributes to Alpha/VMS/Tru64K 'success'. To which I would say bullsh*t. 'COMPAQ' on the side of an F1 carnC does nothing to promote VMS because there isn't any stand-alone VMSiI advertising to speak of that ties it back to the Compaq brand ( as taught : within the first 2 weeks of Marketing 101 at any college).  K But it isn't only advertising/marketing money that will turn the corner for-L VMS, it's getting in front of prospective customers. In the past seven yearsK I have yet to see one instance in the marketspace that I deal in, brokerageaK and banking - and I'm talking about the dozens of customers we have for oureI consulting services and software products, where a DEC/Compaq salespersonoF attempted to even get in the door to sell anything other than PC's andI Proliant servers. Some may wish to dismiss my observations, saying that InI can't possibly know that this is true, but rest assured with the range ofs@ people within my customers that I speak with (CEO/CFO/COO/CTO/IT% managers/sysadm/programmers), I know.g  K The only time VMS/Tru64/Storageworks ever got presented to my customers wasrK a direct result of our efforts in convincing our customers that they shouldoE look at these offerings, and even then it was usually a tooth-pullingaF exercise in trying to convince the local DEC/Compaq office to take theK opportunity seriously enough that they would send somebody knowledgeable to  meet with the client.   H IMHO, the names Capellas and Winkler should figure prominently among theI first 15,000 heads Carly chops, followed by *most* people from the Comapqe! advertising/marketing department.e   ------------------------------   Date: 5 Jul 2002 13:30:54 GMTt( From: nmm1@cus.cam.ac.uk (Nick Maclaren)- Subject: Re: Fearless IPF Prognostications...e0 Message-ID: <ag472e$ide$1@pegasus.csx.cam.ac.uk>  3 In article <Rm1xBSSpf2lh@eisner.encompasserve.org>,p/ Kilgallen@SpamCop.net (Larry Kilgallen) writes:t3 |> In article <ag3uhg$aep$1@pegasus.csx.cam.ac.uk>,o* nmm1@cus.cam.ac.uk (Nick Maclaren) writes: |> > tF |> > You CANNOT do serious server work for a multi-CPU SMP system on aG |> > single (or even dual) system, as many critical issues are entirelyeJ |> > different.  This isn't just tuning, but whether the software actuallyF |> > works.  And a lot of experience is that certain RAS problems growH |> > exponentially with the number of processors (and sometimes even theE |> > memory available).  And, clearly, you can't even start doing anyy, |> > useful tuning on a much smaller system. |> .D |> I would call that testing rather than development.  Certainly theG |> QA department deserves a (permanent) machine as large as the largestrD |> production machine, but careful design in a language designed forH |> multitasking should avoid following an incorrect technique that might  |> not even be discovered by QA. |> tD |> But not all servers involve multitasking within a single process.F |> With a separate (serially reusable) server process for each client,D |> many concurrency issues are taken care of by the operating system |> (or RMS/Rdb/DBMS).i  = Well, those attitudes are PRECISELY what I referred to in the ? paragraph of mine that you snipped.  In both cases, I am afraidv? that you are technically wrong - though I agree that is the wayeA that modern software is developed.  It is also why I am inflictedi> with so much software that works in the restricted environment$ of development and not in the field.  = I could explain why you are technically wrong, but it is veryc= off-group.  Let it suffice to say that it was believed in thed> 1960s that you could emulate a multi-server system on a single? server, but experience showed why it does not work in practice.r< And, no, the issues CANNOT be taken care of by the operating% system, which was another 1960s myth.2     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679d   ------------------------------   Date: 5 Jul 2002 08:16:42 -0600e- From: Kilgallen@SpamCop.net (Larry Kilgallen)b- Subject: Re: Fearless IPF Prognostications...e3 Message-ID: <Rm1xBSSpf2lh@eisner.encompasserve.org>n  [ In article <ag3uhg$aep$1@pegasus.csx.cam.ac.uk>, nmm1@cus.cam.ac.uk (Nick Maclaren) writes:  > 5 > In article <MKftY4E5R3xV@eisner.encompasserve.org>,r1 > Kilgallen@SpamCop.net (Larry Kilgallen) writes:t5 > |> In article <ag3ljb$t4l$1@pegasus.csx.cam.ac.uk>,(/ > |> nmm1@cus.cam.ac.uk (Nick Maclaren) writes:  > |>  I > |> > And remember that I am talking about complete system cost.  Dell's-J > |> > current Itanium server range starts at $15,000, but a viable systemL > |> > will set you back $25,000 or $35,000 if you want to do serious server > |> > application development.r > |> aE > |> I would think a higher price should only apply for deployment -- @ > |> serious _development_ can be done on a machine much smaller( > |> than the one needed for deployment. > |> ?E > |> Or is this a comment specific to the operating systems currently( > |> running on Itanium ?  > A > Partly.  That is the reason that you would need more memory and = > (especially) disk space.  But the real reason is different.m > C > You CANNOT do serious server work for a multi-CPU SMP system on amD > single (or even dual) system, as many critical issues are entirelyG > different.  This isn't just tuning, but whether the software actually C > works.  And a lot of experience is that certain RAS problems grow-E > exponentially with the number of processors (and sometimes even theoB > memory available).  And, clearly, you can't even start doing any) > useful tuning on a much smaller system.   A I would call that testing rather than development.  Certainly the"D QA department deserves a (permanent) machine as large as the largestA production machine, but careful design in a language designed forcE multitasking should avoid following an incorrect technique that might? not even be discovered by QA.i  A But not all servers involve multitasking within a single process.1C With a separate (serially reusable) server process for each client, A many concurrency issues are taken care of by the operating systemr (or RMS/Rdb/DBMS).   ------------------------------  % Date: Fri, 05 Jul 2002 14:50:53 +0100l( From: Nic Clews <sendspamhere@127.0.0.1>- Subject: Re: Fearless IPF Prognostications...e) Message-ID: <3D25A43D.7D2BF19D@127.0.0.1>    John Smith wrote:- >  ...-J > IMHO, the names Capellas and Winkler should figure prominently among theK > first 15,000 heads Carly chops, followed by *most* people from the ComapqS# > advertising/marketing department.t  A Providing of course they are replaced with folks prepared to pushw products properly.  F As a matter of interest, I attended an *internal* presentation for CSCA by HP given by HP-UX technical folks and HP (HP original) accountoF manager (UK). They had no idea I was going to be present, arranged for and by the UNIX guys.E  E For their server roadmaps, they presented a slide on OpenVMS, porting G and operating system release plans. VMS was mentioned a couple of otheraE times as well, but remember this was aimed at current HP-UX customersh$ and Tru64. (NSK was also mentioned).  D There was an odd comment from the (surprised) UNIX audience, some ex1 VMS. The only words of migration were to Itanium.t  H Admittedly, the guy didn't have the details, and he stated that, but didG say, here is the plan. He also wasn't familiar with the Alpha plans, heoD knew that EV7 was the last, but didn't know much about it. Offline IB spoke to him, and he said he's currently getting internal trainingF (familiarization and understanding) on the GS series of Alpha systems,C and on NSK and VMS. I also learnt a bit about the Superdome and itst/ future directions, a little deja vu experience.m  B I have to say, I was pleasantly surprised. If this presentation isG indicative of what is happening within HP, I feel very comfortable with- HP and VMS.r   -- u? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciencesu nclews at csc dot comM   ------------------------------  # Date: Fri, 05 Jul 2002 14:14:18 GMTr1 From: "Terry C. Shannon" <terryshannon@attbi.com> - Subject: Re: Fearless IPF Prognostications...s- Message-ID: <_QhV8.93173$Uu2.14651@sccrnsc03>k  5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message0* news:ag3ljb$t4l$1@pegasus.csx.cam.ac.uk... <snip> >uD > And remember that I am talking about complete system cost.  Dell'sE > current Itanium server range starts at $15,000, but a viable system G > will set you back $25,000 or $35,000 if you want to do serious serverx4 > application development.  Hence the sales figures.  K I suspect that a viable McKinley developer's system will set you back quitesJ a bit less money. Like as little as one fourth to one fifth as much as the numbers cited by Nick.   ------------------------------   Date: 5 Jul 2002 15:06:20 GMTn( From: nmm1@cus.cam.ac.uk (Nick Maclaren)- Subject: Re: Fearless IPF Prognostications...h0 Message-ID: <ag4clc$n0b$1@pegasus.csx.cam.ac.uk>  - In article <_QhV8.93173$Uu2.14651@sccrnsc03>,e3 "Terry C. Shannon" <terryshannon@attbi.com> writes: 8 |> "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message- |> news:ag3ljb$t4l$1@pegasus.csx.cam.ac.uk...  |> >G |> > And remember that I am talking about complete system cost.  Dell'sFH |> > current Itanium server range starts at $15,000, but a viable systemJ |> > will set you back $25,000 or $35,000 if you want to do serious server7 |> > application development.  Hence the sales figures.. |> nN |> I suspect that a viable McKinley developer's system will set you back quiteM |> a bit less money. Like as little as one fourth to one fifth as much as thei |> numbers cited by Nick.   B Hmm.  Maybe, though I doubt that.  Under $4,000 for an entry-levelA system is possible, but I think that c. $8,000 for a system to dolE serious server development on is another.  Even if you are developingoD an application purely for very small servers (i.e. workstations usedE as servers), I think that you will be pushed to get in under $10,000.s  A My guess is that a viable development environment for workstation @ applications (i.e. no more than 2 CPUs, one network card, and noB parallel disk I/O) will come in at $5,000-$10,000.  It will depend@ on how hard Intel and HP want to push them, so your $5,000 could well be correct.  ? But to develop serious server applications that are intended tok? work reliably and perform well on 4- to 8-way SMP systems, with < multiple network cards or with parallel disk I/O will almost< certainly cost $25,000 plus.  The point is that, if you careA about RAS or performance (and I thought that you VMS people did), ? you HAVE to develop on comparable parallelism to that which the  customer is going to use.t  > Of course, many applications will need only ONE of those types> of parallelism, but some will need two or more.  And, with the; McKinley, you won't be able to save money by using a slowero( chip - unless the reports are all wrong.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679e   ------------------------------   Date: 5 Jul 2002 10:07:49 -0600b- From: Kilgallen@SpamCop.net (Larry Kilgallen):- Subject: Re: Fearless IPF Prognostications...:3 Message-ID: <O8nzA+SZ4AIJ@eisner.encompasserve.org>   [ In article <ag472e$ide$1@pegasus.csx.cam.ac.uk>, nmm1@cus.cam.ac.uk (Nick Maclaren) writes::  ? > Well, those attitudes are PRECISELY what I referred to in therA > paragraph of mine that you snipped.  In both cases, I am afraid.A > that you are technically wrong - though I agree that is the waytC > that modern software is developed.  It is also why I am inflictedu@ > with so much software that works in the restricted environment& > of development and not in the field. > ? > I could explain why you are technically wrong, but it is very ? > off-group.  Let it suffice to say that it was believed in thep@ > 1960s that you could emulate a multi-server system on a singleA > server, but experience showed why it does not work in practice.s  B I did not suggest full emulation, but rather _development_.  ThereE is an awful lot involved in correct software besides concurrency, andcC a lot of it can be gotten into working shape in an environment withaD no chance of reproducing a production concurrency problem.  ConsiderG the defects one isolates by stepping through with the debugger.  Nobody>E would consider the timing of that environment to in the least emulateu) production, and yet defects are isolated.r   ------------------------------  $ Date: Fri, 5 Jul 2002 11:44:41 -0400' From: "Main, Kerry" <Kerry.Main@hp.com>,- Subject: RE: Fearless IPF Prognostications...5T Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4026607F6@kaoexc01.americas.cpqcorp.net>   Nick,u  B >>> The point is that, if you care about RAS or performance (and IC thought that you VMS people did), you HAVE to develop on comparablen< parallelism to that which the customer is going to use. <<<<  H I would suggest that you need to QA, tune and test (functional, load andG power users) these systems on equivalent systems before production, buts  not necessarily develop on them.  E Case in point - the same argument could be stated for cluster awaressm$ ie. needing a cluster to develop on.  F While there are no doubt pro's and con's with each approach, what manyE Customers appear to be adopting today is the "develop on cheap stuff,t6 deploy on real stuff" (with good QA/Test environments)   Regardso  
 Kerry Main Senior Consultant  Hewlett-Packard Canada! Consulting & Integration Servicesg Voice: 613-592-4660n Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----2 From: Nick Maclaren [mailto:nmm1@cus.cam.ac.uk]=20 Sent: July 5, 2002 11:06 AMn To: Info-VAX@Mvb.Saic.Comp- Subject: Re: Fearless IPF Prognostications...       - In article <_QhV8.93173$Uu2.14651@sccrnsc03>,e3 "Terry C. Shannon" <terryshannon@attbi.com> writes: ; |> "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message=20e- |> news:ag3ljb$t4l$1@pegasus.csx.cam.ac.uk...a |> >J |> > And remember that I am talking about complete system cost.  Dell's=20H |> > current Itanium server range starts at $15,000, but a viable system  F |> > will set you back $25,000 or $35,000 if you want to do serious=20> |> > server application development.  Hence the sales figures. |>=20oH |> I suspect that a viable McKinley developer's system will set you back  J |> quite a bit less money. Like as little as one fourth to one fifth as=20% |> much as the numbers cited by Nick.A  B Hmm.  Maybe, though I doubt that.  Under $4,000 for an entry-levelA system is possible, but I think that c. $8,000 for a system to dohH serious server development on is another.  Even if you are developing anD application purely for very small servers (i.e. workstations used asB servers), I think that you will be pushed to get in under $10,000.  A My guess is that a viable development environment for workstatione@ applications (i.e. no more than 2 CPUs, one network card, and noE parallel disk I/O) will come in at $5,000-$10,000.  It will depend onvE how hard Intel and HP want to push them, so your $5,000 could well be' correct.  D But to develop serious server applications that are intended to workC reliably and perform well on 4- to 8-way SMP systems, with multipleeB network cards or with parallel disk I/O will almost certainly costF $25,000 plus.  The point is that, if you care about RAS or performance? (and I thought that you VMS people did), you HAVE to develop onaB comparable parallelism to that which the customer is going to use.  A Of course, many applications will need only ONE of those types ofmE parallelism, but some will need two or more.  And, with the McKinley,oC you won't be able to save money by using a slower chip - unless the  reports are all wrong.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679r   ------------------------------   Date: 5 Jul 2002 15:33:59 GMT4( From: nmm1@cus.cam.ac.uk (Nick Maclaren)- Subject: Re: Fearless IPF Prognostications...i0 Message-ID: <ag4e97$oej$1@pegasus.csx.cam.ac.uk>  c In article <O8nzA+SZ4AIJ@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:2^ |> In article <ag472e$ide$1@pegasus.csx.cam.ac.uk>, nmm1@cus.cam.ac.uk (Nick Maclaren) writes: |> uB |> > Well, those attitudes are PRECISELY what I referred to in theD |> > paragraph of mine that you snipped.  In both cases, I am afraidD |> > that you are technically wrong - though I agree that is the wayF |> > that modern software is developed.  It is also why I am inflictedC |> > with so much software that works in the restricted environmente) |> > of development and not in the field.C |> >  B |> > I could explain why you are technically wrong, but it is veryB |> > off-group.  Let it suffice to say that it was believed in theC |> > 1960s that you could emulate a multi-server system on a single/D |> > server, but experience showed why it does not work in practice. |> 6E |> I did not suggest full emulation, but rather _development_.  There H |> is an awful lot involved in correct software besides concurrency, andF |> a lot of it can be gotten into working shape in an environment withG |> no chance of reproducing a production concurrency problem.  ConsidereJ |> the defects one isolates by stepping through with the debugger.  NobodyH |> would consider the timing of that environment to in the least emulate, |> production, and yet defects are isolated.  C That is not, however, how they are located.  And that is the issue.t  D The point about the experience that I referred to is that developing? in environment A for production in environment B means that thesD problems that don't get noticed are usually subtle and intermittent,A and often in the design rather than coding.  They as often as notg5 get through the 'testing' phase and hit the customer.i  B In the most irritating cases, they occur in the testing phase, getE misdiagnosed or run up against a release schedule, fixed incorrectly,mE fail for a FEW customers, and then the customer has the job of tryingiE to persuade the vendor's support team to 'unfix' a bug so that it cantD be fixed properly.  But, far more often, it is "the developers can't/ reproduce that, so we won't treat it as a bug."l  E I can assure you that the matter of parallelism (whether CPU, networkiA or disk) is critical, and I have located more bugs of the class I ? am referring to in supposedly tested software than I can count.i     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679a   ------------------------------   Date: 5 Jul 2002 16:01:10 GMTs( From: nmm1@cus.cam.ac.uk (Nick Maclaren)- Subject: RE: Fearless IPF Prognostications....0 Message-ID: <ag4fs6$q0q$1@pegasus.csx.cam.ac.uk>  ~ In article <BE56C50EA024184DAF48F0B9A47F5CF4026607F6@kaoexc01.americas.cpqcorp.net>, "Main, Kerry" <Kerry.Main@hp.com> writes: |> iE |> >>> The point is that, if you care about RAS or performance (and IyF |> thought that you VMS people did), you HAVE to develop on comparable? |> parallelism to that which the customer is going to use. <<<<o |>  K |> I would suggest that you need to QA, tune and test (functional, load andiJ |> power users) these systems on equivalent systems before production, but# |> not necessarily develop on them.s  @ You might.  Experience from the 1960s onwards is that it doesn't work.0  H |> Case in point - the same argument could be stated for cluster awaress' |> ie. needing a cluster to develop on.c   That is correct.  You do.A  I |> While there are no doubt pro's and con's with each approach, what many2H |> Customers appear to be adopting today is the "develop on cheap stuff,9 |> deploy on real stuff" (with good QA/Test environments)l  D Yes, I am aware of that.  Not merely am I inflicted with the results@ of developed software, I get called in to help developers out of> holes that doing the job properly would have avoided entirely.    ? Look, I am not talking about the EASY issues, but the ones that@? cost real money in the field.  If a developer cannot even run ax= quick test to check his assumptions are correct, he will make ; them anyway.  That is human nature.  Even the most paranoidl< developer will not assume the worst possible interpretation,= because his product will never ship, and would run like a dogN
 if it did.  @ There is also the problem that many such errors might be spottedB as such by the author of the code ("it's not supposed to do THAT")> but may get misdiagnosed by testers ("not what I expected, but@ consistent with the documentation").  Such errors then get fixedB by hitting the field and perhaps being relied on by customers, butD often fail some time later as systems are scaled up and requirements. changed.  By then, it is too late to fix them.  A And then there are the scaling issues.  Not just performance, but A RAS.  Have you ANY idea how often we HPC people have trouble witht@ software that is developed in the way you say?  Now, clearly, it> is too expensive to develop on large HPC systems, but the same= problem arises on smaller scales.  Such as certain operationso< serialising unnecessarily, because the developer used only a? single-CPU box and the tester never realised it wasn't intendedh
 to serialise.a  > And so on.  But I have lost the argument for high-RAS software> development in general-purpose computing, and shall be retired  by the time it is reinvented :-(     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679a   ------------------------------  # Date: Fri, 05 Jul 2002 15:25:33 GMTa1 From: "Terry C. Shannon" <terryshannon@attbi.com>n- Subject: Re: Fearless IPF Prognostications...b. Message-ID: <NTiV8.409001$352.52503@sccrnsc02>  5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in messaget* news:ag4clc$n0b$1@pegasus.csx.cam.ac.uk... >l/ > In article <_QhV8.93173$Uu2.14651@sccrnsc03>, 5 > "Terry C. Shannon" <terryshannon@attbi.com> writes:-: > |> "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message/ > |> news:ag3ljb$t4l$1@pegasus.csx.cam.ac.uk...u > |> >I > |> > And remember that I am talking about complete system cost.  Dell's0J > |> > current Itanium server range starts at $15,000, but a viable systemL > |> > will set you back $25,000 or $35,000 if you want to do serious server9 > |> > application development.  Hence the sales figures.n > |>J > |> I suspect that a viable McKinley developer's system will set you back quitedK > |> a bit less money. Like as little as one fourth to one fifth as much asi ther > |> numbers cited by Nick.c >cD > Hmm.  Maybe, though I doubt that.  Under $4,000 for an entry-levelC > system is possible, but I think that c. $8,000 for a system to dooG > serious server development on is another.  Even if you are developing F > an application purely for very small servers (i.e. workstations usedG > as servers), I think that you will be pushed to get in under $10,000.  >6C > My guess is that a viable development environment for workstationrB > applications (i.e. no more than 2 CPUs, one network card, and noD > parallel disk I/O) will come in at $5,000-$10,000.  It will dependB > on how hard Intel and HP want to push them, so your $5,000 could > well be correct.  H I'd speculate that a nicely configured uniprocessor workstation will sitJ right in the middle of the aforementioned $5K-$10K price band. A similarlyL configured dual-processor workstation could go for about twice as much. Time
 will tell.   >aA > But to develop serious server applications that are intended to A > work reliably and perform well on 4- to 8-way SMP systems, withI> > multiple network cards or with parallel disk I/O will almost> > certainly cost $25,000 plus.  The point is that, if you careC > about RAS or performance (and I thought that you VMS people did),sA > you HAVE to develop on comparable parallelism to that which the0 > customer is going to use.y  H Yep. You could easily spend $25K-$50K on a generously-configured dual or quadprocessor server.    ------------------------------   Date: 5 Jul 2002 05:05:29 -0600c- From: Kilgallen@SpamCop.net (Larry Kilgallen)u# Subject: Re: Last 3 weeks on c.o.v. 3 Message-ID: <WAsUc7TkRJcg@eisner.encompasserve.org>.  [ In article <3D24EA0F.40C3139B@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:u > "Main, Kerry" wrote: >> o	 >> Larry,o >> o: >> <<<In the US there is a VMS-on-Itanium lecture tour.>>> >>  < >> I suspect you have seen these pointers, but just in case:8 >> https://www.showexhibit.com/hp_openvmstour/agenda.cfm > I > The agenda doesn't state a keynote speaker for Chicago. I'll try not to-J > read anything into that. I'm hoping for Bob Blatz. Who ever it is, we'llI > see if they have the guts to face the music out here in the real world.a  = Bob Blatz has certainly been at the last several US Symposia..   ------------------------------   Date: 5 Jul 2002 05:13:48 -0600s- From: Kilgallen@SpamCop.net (Larry Kilgallen)t+ Subject: Re: LGI_BRK_DISUSER and batch job?x3 Message-ID: <cyHaTlpEfvDX@eisner.encompasserve.org>k  d In article <oot9iu4of6qb2ev85nvovgljh88nmjbtti@4ax.com>, Kiasu Surfer <verykiasu@hotpop.com> writes:R > If LGI_BRK_DISUSER is set to 1, and its happened that this user login failed andQ > marked as intruder, and got his/her account disabled via the earlier parameter.aR > My question is, will his/her original earlier scheduled batch job still execute?  F You could experiment to be sure, but I would strongly doubt it, havingB looked through the source listings quite a bit.  DISUSER is ratherH powerful, to the best of my knowledge only overcome by a login to SYSTEM. from the system console (due to this problem).  @ Of course emergency reboot techniques will also work to overcome% DISUSER -- physical access beats all.s   ------------------------------   Date: 5 Jul 2002 13:41:31 GMT 3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)o Subject: Re: Lynx & SSLo0 Message-ID: <ag47mb$jou$1@n.ruf.uni-freiburg.de>  N In article <afsis7$rcg$1@n.ruf.uni-freiburg.de>, I (Christoph Gartmann) wrote:Q >In article <j8u++ultWsfA@mcduck.acs.wmich.edu>, grant@rigel.cc.wmich.edu writes:oD >>It has been quite a while, but I did succeed in building lynx with2 >>openSSL.  Here are the notes I made at the time: >>K >> I was able to build LYNX 2.8.3 with OpenSSL 0.9.5A by changing the build$G >> procedure.  The required changes were to force the MULTINET build tonI >> use UCX emulation mode.  This required changing two DESCRIP.MMS files.a >> >> In [.LYNX2-8-3]DESCRIP.MMSm5 >>   Inside the 'option .eq. "MULTINET"' check, addedl >>   option = "UCX"o! >>   optfile = "MULTINET_UCX.OPT"w >>, >> and creating MULTINET_UCX.OPT containing:2 >> multinet_root:[multinet.library]ucx$ipc/library >>- >> In [.LYNX2-8-3.WWW.LIBRARY.VMS]DESCRIP.MMS,C >>   Inside the '.ifdef MULTINET', comment out "TCP = MULTINET" ande >>   and "TCP = UCX" >>   then, change CFLAGS to beN >>  "/decc/Prefix=All $(DEBUGFLAGS) $(DCFLAGS) /Define=($(EXTRADEFINES), UCX)" >>K >> Once these changes were made, the sources which had been built by adding@M >> the http://www.moxienet.com patches to the 2.8.3 distribution worked fine.  >>H >>Note that, at least at that time, there were some patches which neededD >>to be downloaded. These don't exist at that site any more, and may >>not be necessary, anyway.r >>I >>The biggest problem I had was that the build procedures and code workedI7 >>only with UCX and UCX emulation, not Multinet native.o >-J >No I succeeded in compiling and linking Lynx with SSL. What I did so far:I >1.  Modified BUILD.COM and LIBMAKE.COM to add "USE_SSL,USE_OPENSSL_INCL"i' >    to each "/DEFINE=(...)"-statement.t5 >2.  Modified the LINK command in BUILD.COM and addedi. >       ssllib:libssl/LIB,ssllib:libcrypto/LIB= >    at the end. The order of the two libraries is mandatory!u
 >3.  Added >       typedef int pid_t;! >    to the beginning of HTTP.C .s >e8 >But I cannot make any https-connections. What I get is:= >    Alert!: Unable to make secure connection to remote host.o >nL >Why? Do I have to configure OpenSSL or create any certificates or the like?A >All I did was what is mentioned in INSTALL.VMS, including tests.i  J Got it working. The trick is indeed to compile Lynx for UCX and run it viaL Multinet's UCX emulation. This is because Lynx mixes sockets between OpenSSLN and Multinet. And as OpenSSL doesn't support Multinet directly but via the UCX% emulation Lynx has to do the same :-(a    Anyway, to get it going I had to - edit WWW_TCP.H and replace      #include <ucx$inetdef.h>s   withA      #include <multinet_root:[multinet.include.vms]ucx$inetdef.h>6(   Thanks to Hunter Goatley for this tip.3 - create a file MULTINET_UCX.OPT as described aboveoH - modify the build procedure to use the above opt file instead of UCXSHR5     $!  if option .eqs. "UCX" then optfile = "UCXSHR"r:     $  if option .eqs. "UCX" then optfile = "MULTINET_UCX"1 - removed the typdef statement I applied before  )  O Now the question remains: will OpenSSL ever include support for native Multineto- or will Lynx be modified to demangle sockets?    Regards,    Christoph Gartmannm  H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, Germany                                           |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  $ Date: Fri, 5 Jul 2002 11:55:34 -0400* From: WILLIAM WEBB <WWEBB1@email.usps.gov>E Subject: RE: New web-page dedicated to ports of PD software to OpenVM7- Message-ID: <0033000071392427000002L072*@MHS>e  A =0AI'd agree except it rendered me blind before I could ascertaink the colour scheme.   WWWebb   -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNETk$ Sent: Friday, July 05, 2002 11:52 AMB To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNETE Subject: RE: New web-page dedicated to ports of PD software to OpenVMd     Agreed.    Jan-Erik S=F6derholm wrote:mA > Well, the colour settings on your page must be one of the worseM1 > I'v ever seen ! It's nearly impossible to read.q >e > Jan-Erik S=F6derholm.s >t >  >o >f > JOUKJ wrote: >a >>Hi Folks,  >>E >>I created a new web page (http://nchrem.tnw.tudelft.nl/openvms/) on$E >>which I'm planning to give a summary of all Publicdomain software I0H >>contributed ports and/or patches to in order to have it run on OpenVM= S.F >>The page consists of link to where to get it, what to patch and what; >>software it depends on (and links to these dependencies).gF >>I hope this page will help to make it easier to find/install/run the >>software on OpenVMS. >>H >>The present page is an initial version. More packages will be added t= heH >>coming weeks, but already contains some nice ports not found elsewher= et& >>on OpenVMS pages(i.e. Ted & Pfaedit) >> >>                 Joukl >=   ------------------------------  % Date: Fri, 05 Jul 2002 12:31:40 -0400w1 From: Michael Austin <maustin@firstdbasource.com>wE Subject: Re: New web-page dedicated to ports of PD software to OpenVM 2 Message-ID: <3D25C9EC.6BCB8526@firstdbasource.com>   WILLIAM WEBB wrote:g > @ > I'd agree except it rendered me blind before I could ascertain > the colour scheme. >  > WWWebb >  > -----Original Message-----1 > From: Info-VAX-Request@Mvb.Saic.Com at INTERNET & > Sent: Friday, July 05, 2002 11:52 AMD > To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNETG > Subject: RE: New web-page dedicated to ports of PD software to OpenVMk > 	 > Agreed.v >  > Jan-Erik Sderholm wrote:vC > > Well, the colour settings on your page must be one of the worse 3 > > I'v ever seen ! It's nearly impossible to read.: > >  > > Jan-Erik Sderholm.- > >- > >- > >- > >  > > JOUKJ wrote: > >5
 > >>Hi Folks,m > >>G > >>I created a new web page (http://nchrem.tnw.tudelft.nl/openvms/) oniG > >>which I'm planning to give a summary of all Publicdomain software IZK > >>contributed ports and/or patches to in order to have it run on OpenVMS.sH > >>The page consists of link to where to get it, what to patch and what= > >>software it depends on (and links to these dependencies).PH > >>I hope this page will help to make it easier to find/install/run the > >>software on OpenVMS. > >>K > >>The present page is an initial version. More packages will be added theaJ > >>coming weeks, but already contains some nice ports not found elsewhere( > >>on OpenVMS pages(i.e. Ted & Pfaedit) > >> > >>                 Joukn > >   > This just proves that anyone can create a web site, but having9 the ability to make it look nice takes a special breed ofc? developer. One with both technical and artistic skills.  I havem@ the former, my son has a little of both.  I give him content and he makes it look good.  ? The site probably has some good content, although one would nott; wish to view the entire site as it would render one blind. n -- e Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163 7 Sr. Consultant            http://www.firstdbasource.com                            + http://www.firstdbasource.com/donation.htmly/ 704-947-1089 (Office)     704-236-4377 (Mobile)t   ------------------------------  % Date: Fri, 05 Jul 2002 12:24:35 +0200h' From: JOUKJ <joukj@hrem.stm.tudelft.nl> B Subject: New web-page dedicated to ports of PD software to OpenVMS2 Message-ID: <3D2573E3.6090201@hrem.stm.tudelft.nl>  	 Hi Folks,e  D I created a new web page (http://nchrem.tnw.tudelft.nl/openvms/) on D which I'm planning to give a summary of all Publicdomain software I G contributed ports and/or patches to in order to have it run on OpenVMS.tE The page consists of link to where to get it, what to patch and what ,9 software it depends on (and links to these dependencies).xE I hope this page will help to make it easier to find/install/run the h software on OpenVMS.  H The present page is an initial version. More packages will be added the G coming weeks, but already contains some nice ports not found elsewhere h$ on OpenVMS pages(i.e. Ted & Pfaedit)                      Jouk    ------------------------------  % Date: Fri, 05 Jul 2002 12:24:40 +0200a9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>pF Subject: Re: New web-page dedicated to ports of PD software to OpenVMS' Message-ID: <3D2573E8.A2F8948A@aaa.com>c  ? Well, the colour settings on your page must be one of the worse / I'v ever seen ! It's nearly impossible to read.    Jan-Erik Sderholm.          JOUKJ wrote: >  > Hi Folks,e > E > I created a new web page (http://nchrem.tnw.tudelft.nl/openvms/) oncE > which I'm planning to give a summary of all Publicdomain software IaI > contributed ports and/or patches to in order to have it run on OpenVMS.kF > The page consists of link to where to get it, what to patch and what; > software it depends on (and links to these dependencies).rF > I hope this page will help to make it easier to find/install/run the > software on OpenVMS. > I > The present page is an initial version. More packages will be added thegH > coming weeks, but already contains some nice ports not found elsewhere& > on OpenVMS pages(i.e. Ted & Pfaedit) >  >                  Joukn   ------------------------------  # Date: Fri, 05 Jul 2002 15:53:25 GMT . From: Michael Rice <MichaelARice@adelphia.net>F Subject: Re: New web-page dedicated to ports of PD software to OpenVMS+ Message-ID: <3D25C117.8050603@adelphia.net>o   Agreed.    Jan-Erik Sderholm wrote: A > Well, the colour settings on your page must be one of the worsea1 > I'v ever seen ! It's nearly impossible to read.  >  > Jan-Erik Sderholm.t >  >  >  >  > JOUKJ wrote: >  >>Hi Folks,a >>E >>I created a new web page (http://nchrem.tnw.tudelft.nl/openvms/) onsE >>which I'm planning to give a summary of all Publicdomain software IeI >>contributed ports and/or patches to in order to have it run on OpenVMS.tF >>The page consists of link to where to get it, what to patch and what; >>software it depends on (and links to these dependencies).pF >>I hope this page will help to make it easier to find/install/run the >>software on OpenVMS. >>I >>The present page is an initial version. More packages will be added thecH >>coming weeks, but already contains some nice ports not found elsewhere& >>on OpenVMS pages(i.e. Ted & Pfaedit) >> >>                 Joukt >    ------------------------------  * Date: Fri, 5 Jul 2002 04:53:26 -0700 (PDT). From: Fabio Cardoso <fabiopenvms@yahoo.com.br>& Subject: OpenVMS 7.2-1H1 (Differences)@ Message-ID: <20020705115326.42993.qmail@web20210.mail.yahoo.com>  5 I dont know if it is a bug but I am having problems  d* to use the DIFF command under OVMS 7.2-1H1    % $ diff node1::disk1:[dir]login.com  -a   disk2:[dir]login.com   %DIFF-F-OPENIN, error opening % node1::disk2:[dir]LOGIN.COM; as inputl1 -RMS-F-DEV, error in device name or inappropriate  device type for operationr    4 I am trying to check a local file, but it reports me the remote node (node1)s  
 When I put  $ $ diff node1::disk1:[dir]login.com -!   localnode::disk2:[dir]login.come   It works fine ! !    RegardsU   FC d     =====  ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazil- fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?. Sign up for SBC Yahoo! Dial - First Month Free http://sbc.yahoo.com   ------------------------------  % Date: Fri, 05 Jul 2002 14:09:18 +0200.9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>-* Subject: Re: OpenVMS 7.2-1H1 (Differences)' Message-ID: <3D258C6E.D05067BD@aaa.com>h  = I think that if you don't specify a node for the second file,o= DIFF take whatever was specified (or defaulted) for the first # file as default for the second one.>   Does this work :    > $ diff disk2:[dir]login.com  - >   node1::disk1:[dir]login.com3  8 Your local node should be default for the first file and) node1:: is specifyed for the second file."   Jan-Erik Sderholm.r   Fabio Cardoso wrote: > 5 > I dont know if it is a bug but I am having problemsd, > to use the DIFF command under OVMS 7.2-1H1 > ' > $ diff node1::disk1:[dir]login.com  -r >   disk2:[dir]login.com >  > %DIFF-F-OPENIN, error openinge' > node1::disk2:[dir]LOGIN.COM; as input 3 > -RMS-F-DEV, error in device name or inappropriate  > device type for operation    ------------------------------   Date: 5 Jul 2002 07:43:47 -0600 + From: young_r@encompasserve.org (Rob Young) ( Subject: Re: Quorum discussion/questions3 Message-ID: <ixqQRF+sQGG+@eisner.encompasserve.org>.   In article <kR0V8.6112$Go3.142778@newsfep1-win.server.ntli.net>, "Dane Maslen" <dane.maslen@real-address-withheld-to-avoid-spam> writes: >   > Rob Young wrote in message ...9 >> But here is arguably a better way I learned at a site.e >>- >> 1)  All nodes except satellites get 1 voteeF >> 2)  No quorom disk, not needed!  Helps speed the transitions up for
 >> one thing!a5 >> 3)  Expected Votes = all voting nodes totalled up.l >>B >> Working with a cluster with 6 nodes, we would on the FIRST node >> to boot:l >>D >> >>> boot -fl 1,N  $n$DGA|DUA.a.b.c.d      ! Alpha booting SYSBOOT! >> SYSBOOT>  SET WRITESYSPARAMS 0u! >> SYSBOOT>  SET EXPECTED_VOTES 1l >> SYSBOOT>  CONTINUEg >> >> Where N = that node's root. >># >> Since EV = 1, this node comes upo > * > One minor but very important correction: >   > SYSBOOT>  SET EXPECTED_VOTES 1  > SYSBOOT>  SET WRITESYSPARAMS 0 > SYSBOOT>  CONTINUE > G > Every time a parameter is set, WRITESYSPARAMS is implicitly set to 1.nL > Therefore setting WRITESYSPARAMS to 0 must be the last thing you do beforeJ > the CONTINUE, otherwise the revised SYSGEN parameters will be written to! > disk and used for future boots.  >   C 	Yes.  Thanks.  I've made that mistake a few times and even in the eH 	posting.   That is also why I suggest you examine the parameters after G 	the cluster is up as it only takes one flip-flop and I posted several.d   				Robd   ------------------------------  % Date: Fri, 05 Jul 2002 07:28:01 +0100:( From: Nic Clews <sendspamhere@127.0.0.1>A Subject: Re: System Disk Initialization Parameter recommendationsD) Message-ID: <3D253C71.6DA774CC@127.0.0.1>    JF Mezei wrote:A > O > If you get a empty system disk you have just FORMATTED, how does one go aboutKK > loading *VAX* VMS on it and still provide your own parameters, especiallyu0 > cluster size which can't be changed later on ? > P > When you BACKUP/IMAGE or the B saveset , won't backup do the Init with its own > parameters ?  E Well whatever the B saveset was made with. This is one to be slightly C wary of, if you still have a small disk, if the cluster size on thewD source gets larger, you'll lose space through truncation boundaries.  D I notice that 7.2 VAX/VMS has a cluster size of 9. No good for small disks.  -O > I realise that on alpha, you can boot from CD and then do the INIT from therep& > and then backup/noinit. But on VAX ?   BACKUP/NOINIT on VAX as well...   ? I've also used BACKUP/IMAGE of a disk, changed the cluster size</ (reinitialized), then BACKUP/IMAGE/NOINIT back.u -- e? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciences1 nclews at csc dot com    ------------------------------  % Date: Fri, 05 Jul 2002 07:32:34 +0100i( From: Nic Clews <sendspamhere@127.0.0.1>A Subject: Re: System Disk Initialization Parameter recommendations ) Message-ID: <3D253D82.DF59082C@127.0.0.1>,   Paul Repacholi wrote:b >   > norm.raphael@metso.com writes: > : > > I am about to migrate from a 4.1GB SCSI system disk to> > > an 18GB SAN-based system disk (with OpenVMS Alpha V7.2-2). > ...o >  > > * for   /CLUSTER_SIZE: > ...sF > > [I would guess that keeping 9 is reasonable, but that is a guess.] > G > It would be OK. Beware of small values on big disks, allocation times2  > can take a long lunch-break :(  C This is the first reference I've seen with regard to performance byDG using smaller than the default cluster size. (I've always avoided usingu small cluster sizes).i   Care to elaborate, please?   -- j? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciencess nclews at csc dot com    ------------------------------  % Date: Fri, 05 Jul 2002 20:03:14 +0010n% From: paddy.o'brien@zzz.tg.nsw.gov.auaA Subject: Re: System Disk Initialization Parameter recommendations 5 Message-ID: <01KJQXU82RRM00074X@tgmail.tg.nsw.gov.au>U   Paul Repacholi wrote:U >   > norm.raphael@metso.com writes: > : > > I am about to migrate from a 4.1GB SCSI system disk to> > > an 18GB SAN-based system disk (with OpenVMS Alpha V7.2-2). > ...  >  > > * for   /CLUSTER_SIZE: > ...TF > > [I would guess that keeping 9 is reasonable, but that is a guess.] > G > It would be OK. Beware of small values on big disks, allocation timesI  > can take a long lunch-break :(  + Why?? (A big question, not an interjection)   J I have only 9GB disks at present, the default cluster size is 18.  On the J system disk I have left the default, but changed all my user disks to 9.  E This seemed a reasonable average size for the files created by users./  J I am buying replacement disks of 36Gb, and was intending to use a similar J philosohy ... leave system default at whatever, but initialise user disks  again with 9 cluster size.  E On your comment, is 9 too small and what will be the effect on users?i  N Note also that many of our applications have an expectation of the file size, J so open files with a reasonable initial blocksize and a similar extension  size.n  E Many files created by the users are <9 blocks, including things like K NETSERVER.LOG.   Regards, Paddy   ------------------------------   Date: 5 Jul 2002 05:04:30 -0600I- From: Kilgallen@SpamCop.net (Larry Kilgallen)lA Subject: Re: System Disk Initialization Parameter recommendationsh3 Message-ID: <Vp+NuEKcbQQo@eisner.encompasserve.org>y  T In article <3D253C71.6DA774CC@127.0.0.1>, Nic Clews <sendspamhere@127.0.0.1> writes:  P >> I realise that on alpha, you can boot from CD and then do the INIT from there' >> and then backup/noinit. But on VAX ?O  A VAX/VMS distribution CDROMs have supported booting into a limiteds@ DCL environment since about 1994.  I forget which alternate root provides this, perhaps 1.e   ------------------------------   Date: 5 Jul 2002 07:50:55 GMTa7 From: sy18889@rabmbit.famrp.cosm (Bradford J. Hamilton) A Subject: Re: System Disk Initialization Parameter recommendationst! Message-ID: <LvXbYzlOPDB5@rabbit>    Hi Paul,  \ In article <878z4rxnar.fsf@prep.synonet.com>, Paul Repacholi <prep@prep.synonet.com> writes:  > norm.raphael@metso.com writes: <snip> v   >> * for   /INDEX D >> [This used to matter, but does it on a SAN disk?  Would BEGINNINGC >> be just as good as END, especially if /HEADERS is large enough?]  >  > (Lets buy an argument...)! > F > Middle is still the way to go for the same resons it was before. NowG > before anyone jumps up and metions faster xfers on outer tracks, workdH > out where the `middle' is on a zoned disk. I have seen index files endB > up at the beggining of the disk, but don't know if it is INIT or > BACKUP I should yell at.  M Is there some information in the doc regarding "zoned" disks, and the reasons-K behind using the middle of the disk for /INDEX?  I'm still kind of new, and  would appreciate a pointer...    <snip>   > --  > > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda.yB >                                              West Australia 60760 > Raw, Cooked or Well-done, it's all half baked.H > EPIC, The Architecture of the future, always has been, always will be. -- m Bradford J. Hamilton& braMdhamAilPtoSn@aMtAtPbi.cSom		(home)& sMy1A88P89S@rabMbit.fAmPr.coSm		(work)  ; "All opinions that I express are my own, not my employer's"I "Lose the MAPS"t   ------------------------------   Date: 5 Jul 2002 10:54:44 -0700W. From: SPAMSINK2001@YAHOO.COM (Alan E. Feldman)A Subject: Re: System Disk Initialization Parameter recommendations1= Message-ID: <343f30ae.0207050954.237e55c0@posting.google.com>a  c norm.raphael@metso.com wrote in message news:<OF6F3571AD.7E9844FB-ON85256BEA.0055841F@metso.com>...-8 > I am about to migrate from a 4.1GB SCSI system disk to< > an 18GB SAN-based system disk (with OpenVMS Alpha V7.2-2). > 9 > If I just use BACKUP/IMAGE the cluster factor increases  > from 9 to 35.r >  > I think 35 is too large. > ? > I can INITIALIZE the target disk and user BACKUP/IMAGE/NOINIT>D > to change things, but I have no information on what is reasonable. > F > A search of the DSNlink Database produces no article that referencesF > the SYSTEM DISK.  The original disk was FIS, so the values on it are > a mystery. > 1 > What does VMS development use? recommend? ship?oE > Is there an article on this specific subject ( Shouldn't there be)?t > 
 > Especially:>    F For cluster size, it's a case of space savings versus block allocation@ performance. I don't know what's best here. See the other posts.A Actually, I think this will also affect how quickly the disk gets - fragmented. Can someone else comment on that?-     [SNIP] >  > * for /DIRECTORIES:4 >  > ( >         /DIRECTORIES=number-of-entries >  [SNIP]E > [I usually use 64, but I have no idea of the cost or reasonableness " > of that value on a system disk.]    A It seems to me that if you will never have more than 64 top levelr0 directories, 64 is fine. It's only a few blocks.     >  > * for   /MAXIMUM_FILES:O >  >         /MAXIMUM_FILES=n > @ >      Restricts the maximum number of files that the volume canB >      contain. The /MAXIMUM_FILES qualifier overrides the default- >      value, which is calculated as follows:a [SNIP]    E This qualifier determines the size of the index file bit map. You canoE never have more file headers on a disk than this number. I'd leave itsB at the default unless there is a chance you'll have more than thatD many file headers on the disk at one time. Also, the size of this isD changeable only by re-initializing the disk, I believe. So make sureE it's big enough. It probably is, unless you're going to fill the diskl- with a very large number of very small files.t    : > [The existing volume has a value of about 420,000 files] >  > *for   /HEADERS: > $ >         /HEADERS=number-of-headers > D >      Specifies, for disk volumes, the number of file headers to beB >      allocated for the index file. The minimum and default valueB >      is 16. The maximum is the value set with the /MAXIMUM_FILES >      qualifier.  [SNIP] > E > [What is a guideline for the system disk in terms of maximum files?d  > All, half, some other number?]    F The algorithm for extending INDEXF.SYS has been greatly improved sinceE somewhere around V5.x. Unless you have one of the buggy versions, yout> shouldn't worry too much about this. It used to be extended inF 1000-block increments whenever new file headers were needed. That usedE to result in a many-piece index file and in my experience the dreadedoE HEADERFULL problem wouldn't happen until you had about 50000 files. IeC had a buggy version of the improved extension method with 5.5-2 and ? got a patch from DEC to fix it. I've never had a problem since.   F I wouldn't bother with this if I were you, but if you can estimate theD maximum number of file headers you'll ever have at one time, you canC set it to that. Maybe 500K or 600K in your case. But again, I wouldo@ not lose any sleep over this one. The new extension algorithm is> pretty good in my experience (VMS V5.5-2 patched, v6.1, v6.2).     Summary for number of "files":  C /MAXIMUM_FILES is the size of the index file bit map. There must belC one bit in this map for every file header. If the file header is inuB use, the bit is 1. If not, it is 0. This is unchangeable without aD re-INIT, I believe. You can never have more file headers on the disk than this number.O  F /HEADERS is the initial number of file headers to create in INDEXF.SYSD during the INIT. The file INDEXF.SYS is extended as needed to createD new file headers with a greatly improved algorithm, so the value youD use for this is much less important than it used to be. I can try to! dig up the algorithm if you like.w   [SNIP]  % Can't help you with the rest of them.d     Disclaimer: JMHO Alan E. Feldman  afeldman gfigroup com    ------------------------------  % Date: Fri, 05 Jul 2002 06:28:24 -040012 From: rdeininger@mindspring.com (Robert Deininger)A Subject: Re: System Disk Initialization Parameter recommendationssK Message-ID: <rdeininger-0507020628240001@1cust149.tnt1.nashua.nh.da.uu.net>   3 In article <3D253D82.DF59082C@127.0.0.1>, Nic Clewss <sendspamhere@127.0.0.1> wrote:m   >Paul Repacholi wrote: >> S! >> norm.raphael@metso.com writes:  >> i; >> > I am about to migrate from a 4.1GB SCSI system disk tou? >> > an 18GB SAN-based system disk (with OpenVMS Alpha V7.2-2).  >> ... >> a >> > * for   /CLUSTER_SIZE:  >> ...G >> > [I would guess that keeping 9 is reasonable, but that is a guess.]  >> nH >> It would be OK. Beware of small values on big disks, allocation times! >> can take a long lunch-break :(  >oD >This is the first reference I've seen with regard to performance byH >using smaller than the default cluster size. (I've always avoided using >small cluster sizes). >  >Care to elaborate, please?t   A few cent's worth...f  J Larger cluster sizes can waste space.  On a system disk there tend to be aI lot of smallish files.  Disks with users' home directories often have the H same problem.  The default cluster size on modern disks is often 18, 37,I or even larger.  Fill that up with files in the 3-10 block range, and you0G may be wasting more than half the disk.  This much is pretty obvious, Ih guess.  J Of course, if you know you'll be putting mainly large files on the disk, a1 cluster size of 500, 1000, or larger may be fine.   C If you keep your disks relatively defragmented, small cluster sizes D shouldn't be a problem.  Most of the overhead is per extent, not perC cluster.  BUT, if you let a disk get very fragmented (a significant G fraction of free extents are only 1 cluster in size), large allocationscF will be slow.  The resulting files will be very fragmented as well.  IA view this as a fragmentation problem, not a cluster-size problem.e  % If I had to summarize in a few lines:a1 1. large cluster size + small files wastes space.nB 2. small cluster size + large files wastes some time, and with bad& fragmentation may waste a lot of time.  E To help reduce fragmentation, I like to set the /EXTENSION to several B times the /CLUSTER_SIZE -- usually 5x or 10x.  This makes "stupid"J applications that do many extend operation less damaging.  RMS system-wideH defaults are also helpful here, but I find per-disk customization better than system-wide.@   ------------------------------   Date: 5 Jul 2002 08:27:33 GMTi( From: nmm1@cus.cam.ac.uk (Nick Maclaren)4 Subject: Re: Three HP Press releases (via Bloomberg)0 Message-ID: <ag3l9l$sth$1@pegasus.csx.cam.ac.uk>  - In article <87it3vw6kc.fsf@prep.synonet.com>,n. Paul Repacholi  <prep@prep.synonet.com> wrote:+ >nmm1@cus.cam.ac.uk (Nick Maclaren) writes:m >yC >> Yes.  And considering that the POWER4 is close behind, on a 0.18sF >> micron process, and has been out for a while, it isn't only the EV7G >> that might knock it off its pedestal.  Of course, Intel have already G >> brought the Madison forward (on paper), and could probably produce ag' >> 0.13 micron McKinley if they wanted.@ >o; >I think there is a re-spin of the POWER4 due in Oct or so._  A Several articles have said this recently, including a bizarre oneD? posted yesterday on The Inquirer.  My belief is that the author-@ of that one needs to foster a bit more cynicism and disinterest.   >On the intel side.6 >hB >there was a report on TheRegister where an intel exec was bragingG >how McKinley was all intels own work. Odd, wasn't McK the HP design???f >Or did the truth slip out...  >FK >I suspect that Merced was euthenaised in 99 or so, and the names filed offoM >with no notice. The McK has already shipped labeled Merced, and the Maddison M >is the one about to arrive. This would fit in with the last slip on `Merced' F >as it is about the time needed to add the x86-box back in to the McK.  F I am 95% sure that was not the case, but that the situation is almost D equally contorted.  What I believe, based on public information, is:  A The original IA-64 (sic) architecture was changed then, mainly to A embed x86 support but perhaps in other ways, too.  The Merced wasn< eviscerated at least once and rebuilt under the direction of Dr Victor Frankenstein.   D Rumours are that the HP engineers fought that change tooth and nail,> lost out politically, and that the McKinley was transferred to? Intel for final validation and conversion to silicon.  The samesB rumours said that the joint development deal (which originally wasC indefinite) was terminated as part of the renogiation, and that alls* future IA-64 chips would be Intel's alone.  B I have also heard contrary reports, so I believe that the politicsC is nothing like as simple as that makes out.  But I can't say more,s& to avoid breaching confidence.  Sorry.  A I don't know exactly when x86 support was added, but it shouldn'ttC have meant a complete redesign, if the original McKinley was clean. B I cannot say why the extra pipeline stage was added, but I am veryC surprised (a) it was done at such a late stage, (b) it seemed to gou? so smoothly and (c) so few commentators picked it up as a major  issue.  D >> But, despite all the shouting, I doubt that SpecFP performance isG >> going to be the reason that the McKinley sinks or fails.  It will beTG >> whether it works in the field, perhaps I/O performance and certainly  >> Microsoft's machinations. >PF >Well, if anyone can get a machine that they are allowed to benchmark.  E If Intel don't launch the McKinley this year, the IA-64 line is dead.eB I am 90% certain it will be launched this month, but am waiting to@ see which OEMs announce products with delivery dates and prices.8 Trumpet voluntaries are to be expected, but mean little.  C When I told two OEMs that there was not a hope in hell (probably innD those words) of us procuring a system if the users had to sign IntelE NDAs, they confirmed that most real customers had told them the same.s     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679A   ------------------------------  % Date: Fri, 05 Jul 2002 08:56:59 +0100o% From: Alan Greig <a.greig@virgin.net>e Subject: Re: UAF questions8 Message-ID: <f3kaiusq7f2nv5a3vm66pt76mfk0eaa97k@4ax.com>  E On Thu, 04 Jul 2002 21:09:55 -0400, Howard S Shubs <howard@shubs.net>  wrote:  . >In article <874rffxmyy.fsf@prep.synonet.com>,/ > Paul Repacholi <prep@prep.synonet.com> wrote:l >eE >> Do not EVER use SYS$SYSTEM... Use SYS$COMMON:[SYSEXE] or directorynB >> etc to suit. Unless you like having to seach and destroy orphanA >> SYSUAFs left in SYS$SPECIFIC:[SYSEXE]s and the odd ball errorso >> that will bite you. >NJ >That's only a problem if you let the system create such files.  The only H >SYSUAF.DAT file on your system should be in SYS$COMMON:[SYSEXE], which   ? And if you have multiple boot disks and/or a mixed architecturee cluster for example?  6 >is pointed to by SYS$SYSTEM (SYS$SYSROOT:[SYSEXE] -> E >SYS$SPECIFIC:[SYSEXE],SYS$COMMON:[SYSEXE]).  It will only work on a 2F >version in SYS$SPECIFIC if there's one there already.  Otherwise, it E >will -always- open the one it finds -before- it tries to create one.. >9D >The only way a copy could exist in SYS$SPECIFIC:[SYSEXE] is if the C >system manager copied one to that location or SET DEFAULT to that sG >specific location before doing MCR AUTHORIZE.  The system manager who .D >might do so is a bit of a fool, as screwing around with SYSUAF.DAT @ >should always be done most carefully, keeping in mind the very  >possabilities you mention.i   -- Alan   ------------------------------   Date: 5 Jul 2002 05:06:39 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen)b Subject: Re: UAF questions3 Message-ID: <cHclX0kw+wAp@eisner.encompasserve.org>l  \ In article <874rffxmyy.fsf@prep.synonet.com>, Paul Repacholi <prep@prep.synonet.com> writes:2 > spamsink2001@yahoo.com (Alan E. Feldman) writes: >  >> Oops! I forgot... > O >> $ SET DEFAULT SYS$SYSTEM  ! unless SYSUAF is defined to be your sysuaf file.a > 
 > </flame on>s > D > Do not EVER use SYS$SYSTEM... Use SYS$COMMON:[SYSEXE] or directoryA > etc to suit. Unless you like having to seach and destroy orphano@ > SYSUAFs left in SYS$SPECIFIC:[SYSEXE]s and the odd ball errors > that will bite you.   E That is only an issue with certain broken programs.  AUTHORIZE is not  one of them, nor is TECO.S   ------------------------------  $ Date: Fri, 5 Jul 2002 09:33:12 +02002 From: "Frits A.M. Storms" <frits@storms.tmfweb.nl>8 Subject: Re: Using GNU C on OpenVMS FAQ (Looking for it)? Message-ID: <3d255087$0$12290$e4fe514c@dreader4.news.xs4all.nl>   A "Martin Vorlaender" <martin@radiogaga.harz.de> schreef in berichth5 news:3d231f4b.524144494f47414741@radiogaga.harz.de...l3 > Frits A.M. Storms (frits@storms.tmfweb.nl) wrote:a: > > When linking the object of the small c-program below : > > #include <stdio.h> > > main( )i > > {s" > >     printf("Hello World !\n"); > > }f >t > > When > > $ LINK/NOTRACE HELLO > > I get :i@ > > %LINK-W-USRTFR, image HELLO.EXE has no user transfer address >eK > IIRC, you have to link against the GNU C startup code (filename CRT... ?)e& > which provides the main entry point.   Yep !r That did the trick.c I used CRT0.OBJ but that gave :i
 Hello World !l) %SYSTEM-E-EXQUOTA, process quota exceededu% Using CRT1.OBJ worked with no errors:e= LINK/NOTRACE hello,gnu_cc_library:crt1,sys$share:decc$shr/opt0   Thanks a lot !   ------------------------------  # Date: Fri, 05 Jul 2002 14:51:48 GMTk- From: "John E. Malmberg" <wb8tyw@qsl.network>s8 Subject: Re: Using GNU C on OpenVMS FAQ (Looking for it)* Message-ID: <3D25AE4B.9000500@qsl.network>   Frits A.M. Storms wrote: >  > Yep !I > That did the trick. ! > I used CRT0.OBJ but that gave :e > Hello World ! + > %SYSTEM-E-EXQUOTA, process quota exceededA' > Using CRT1.OBJ worked with no errors:b? > LINK/NOTRACE hello,gnu_cc_library:crt1,sys$share:decc$shr/opte  I I am not sure how your link statement succeeded.  I suspect that you are -< using a LNK$LIBRARY* logical name to bring in the VAX C RTL.  I The SYS$SHARE:DECC$SHR.EXE image has prefixes on all of the modules, and  I the output from GCC, (unless there has been a major change) does not put  ! prefixes on the external symbols.y  H Because the VAXCRTL is old, it is much better to be using the DEC C RTL  instead.  A There is a special DECC shared image that does not have prefixes.d  C For an example of how to use the DECC RTL with GCC, please look at:c  " ftp://ftp.qsl.net/pub/wb8tyw/gzip/   -Johnt wb8tyw@qsl.network Personal Opinion Onlyl   ------------------------------   Date: 5 Jul 2002 09:13:01 -0700n) From: P.Young@unsw.EDU.AU (Patrick Young)r) Subject: Re: VMS IO up to the level of PC = Message-ID: <55f85d77.0207050813.46e393bc@posting.google.com>n  ^ Osmo Kujala <kujala@tukki.cc.jyu.fi> wrote in message news:<afp3l1$djd$1@mordred.cc.jyu.fi>...  & > Does the card do 160MB/sec with VMS?  B Yes. IIRC at the time of my posting I had taken my home PC164 into% work to try out some new 180GB disks.c  C U160 SCSI is fast ... but it is not a matter of "size that counts", & but more a matter of "how you use it".  : Put simply, U160 is for RAID 0/5 and/or using fast drives.  E As a simple example, on my home Alpha (the PC164 mentioned) that I amrJ now using to post this reply, the U160 controller has two drives attached:   $ sh mag pkbG Device PKB0:, device type Adaptec AIC-7899, is online, error logging ism  B (AIC-7899 display from the driver is misleading, and needs fixed).   $ sh mag dkbN Disk $1$DKB0: (HILUX), device type QUANTUM ATLAS_V_18_WLS, is online, mounted, ...e  K Disk $1$DKB100: (HILUX), device type FUJITSU MAN3184MP, is online, mounted,u ...:  ! As a crude, but layman type test:e   $ set def dkb:[0,0]e& $ mc sysgen cre x.x/contig/size=400000 $ create test.coma  $ show time  $ copy x.x nl:e  $ show time  ExitF $ @test9    6-JUL-2002 01:30:45    6-JUL-2002 01:30:55 $ set def dkb100:[0,0]& $ mc sysgen cre x.x/contig/size=400000 $ copy dkb:[0,0]test.com * $ @testh    6-JUL-2002 01:31:41    6-JUL-2002 01:31:46  2 And both at the same time (in different DECterms):   $ set def dkb:[0,0]  $ @testa    6-JUL-2002 01:32:48    6-JUL-2002 01:32:59   $ set def dkb100:[0,0] $ @test     6-JUL-2002 01:32:49    6-JUL-2002 01:32:57  G It is clear that the Fujitsu drive has a faster internal transfer rate.   J > There is also Adaptec 39160 with 7899G and I've seen mentioned that SRM   J From what I am *told* the 39160 is "Window(tm)_stripped_down_ware". I haveB not tried one and don't know - but would be surprised if it works.  G I have found in the past that Adaptec is not forthcoming on data sheetso) where U160 is involved. Has this changed?s   ------------------------------  % Date: Tue, 25 Jun 2002 10:43:43 -0500u, From: "Tony Scandora" <Scandora@cmt.anl.gov> Subject: Re: VMS port delayed!+ Message-ID: <ag4dek$2oj$1@milo.mcs.anl.gov>   5 "Bob Ceculski" <bob@instantwhip.com> wrote in messagel7 news:d7791aa1.0206241510.112a7f0b@posting.google.com...s > ...fH > SOURCES CLOSE to the action in the USA tell the INQUIRER that a little@ > snagette is likely to hold up the much vaunted port of the VMS3 > operating system to the Itanium processor family. H > VMS will need a hardware abstraction layer (HAL) to run on the ItaniumC > and that could take a little longer to prepare than the engineerso
 > thought.  G Has anyone tallied up the number of stories the Inquirer published thatqL turned out to be not close to reality?  The VMS developers have given a fairL amount of their time to posting milestones and status on this newsgroup, andG the HPQ VMS web site is running a pool to guess the first boot time.  IeL doubt that a corporate webmaster would allow customers to play that guessingH game too far in advance, and I hold the named VMS developers who post toI this newsgroup in higher esteem than unnamed sources of a news service off questionable reliabliity.   1 Tony Scandora, Argonne National Lab, 630-252-7541r scandora@cmt.anl.gov   ------------------------------  # Date: Fri, 05 Jul 2002 06:24:36 GMT * From: "Bill Todd" <billtodd@metrocast.net>Y Subject: Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar vB Message-ID: <EYaV8.559231$%y.36742574@bin4.nnrp.aus1.giganews.com>  ; "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in messagel7 news:b096a4ee.0207041956.7e875e46@posting.google.com... 7 > "Bill Todd" <billtodd@metrocast.net> wrote in message = news:<4_8U8.81586$Ca2.4755216@bin2.nnrp.aus1.giganews.com>...s? > > "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in messagee; > > news:b096a4ee.0207011838.3539685d@posting.google.com...e   ...a   > > > Am I missing something > > > here?a > > F > > Yes.  And since it's been called to your attention already, either you're a> > > bit slow or not really making the effort to understand it. >h >w! > Let's refrain from the insults.C  I Feel free to act as you see fit; as for me, I'll decide on a case-by-casefK basis.  When you ask a question, be prepared for a straight answer:  if youlI hadn't attempted to be cute in asking it, I might have been less blunt inr responding to it.    >o >oJ > > He advocates a change that (he believes) would eliminate the perceived needL > > to defend one's home with a gun.  *In the absence of that change* (i.e.,H > > under current conditions), he clearly himself does feel such a need.H > > There's no contradiction (because they're two different situations), henceb > > no hypocrisy.  >0 >hC > Uh, perceived need is not equivalent to real need. People who aremH > paranoid may have perceived needs for owning a gun in situations where > one doesn't really need one. >iE > And your missing situation number three: the interim period betweendF > such a law taking effect and a large reduction in crime. During thatG > period, it would still be prudent to own a gun. And Mr. Rowan clearlyrG > says that when faced with such dangers he would own a gun. Therefore, = > he says that he would own a gun during that interim period.l  G You're really full of shit, you know.  Or completely ignorant of simple0B logic.  What Mr. Rowan said is what he said, no more, no less, andK attempting to read into it support for your own viewpoint rather than stickgA to it literally is the mark of demagoguery rather than intellect.t  H The way to find out Mr. Rowan's view on whether he feels such an interimF situation would be necessary for him to address and if so how he wouldB address it is to ask him.  Lacking that, calling your suppositionsD 'evidence' of the hypocrisy you claim exists (as opposed to multipleC alternate interpretations ranging from simply disagreeing with your F assessment of the 'interim' conditions to the possible naivete alreadyK mentioned) demonstrates considerably less about him than it does about you.c   - bill   ------------------------------  % Date: Fri, 05 Jul 2002 19:21:13 +0010d% From: paddy.o'brien@zzz.tg.nsw.gov.au Y Subject: Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar e5 Message-ID: <01KJQWE4CGG200072R@tgmail.tg.nsw.gov.au>c   Bill Todd said:s  ; (I hope I've got the dead sergeants, i.e. attributes right)   < >"Alan E. Feldman" <spamsink2001@yahoo.com> wrote in message7 news:b096a4ee.0207041956.7e875e46@posting.google.com...t8 >> "Bill Todd" <billtodd@metrocast.net> wrote in message= news:<4_8U8.81586$Ca2.4755216@bin2.nnrp.aus1.giganews.com>... @ >> > "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in message< >> > news:b096a4ee.0207011838.3539685d@posting.google.com... >> >...   >> > > Am I missing somethingi >> > > here? >> >G >> > Yes.  And since it's been called to your attention already, eitherp you're a? >> > bit slow or not really making the effort to understand it.  >> >>" >> Let's refrain from the insults. > J >Feel free to act as you see fit; as for me, I'll decide on a case-by-caseL >basis.  When you ask a question, be prepared for a straight answer:  if youJ >hadn't attempted to be cute in asking it, I might have been less blunt in >responding to it.  J And why do we need so much OT stuff here?  A simple translation of Bill's L last paragraph is: I disagree with you, but won't give up arguing until you  see my viewpoint.    >> >>K >> > He advocates a change that (he believes) would eliminate the perceived  needM >> > to defend one's home with a gun.  *In the absence of that change* (i.e.,nI >> > under current conditions), he clearly himself does feel such a need.dI >> > There's no contradiction (because they're two different situations),- hence- >> > no hypocrisy. >> >>D >> Uh, perceived need is not equivalent to real need. People who areI >> paranoid may have perceived needs for owning a gun in situations where. >> one doesn't really need one.f >>F >> And your missing situation number three: the interim period betweenG >> such a law taking effect and a large reduction in crime. During thataH >> period, it would still be prudent to own a gun. And Mr. Rowan clearlyH >> says that when faced with such dangers he would own a gun. Therefore,> >> he says that he would own a gun during that interim period. >aH >You're really full of shit, you know.  Or completely ignorant of simpleC >logic.  What Mr. Rowan said is what he said, no more, no less, andtL >attempting to read into it support for your own viewpoint rather than stickB >to it literally is the mark of demagoguery rather than intellect. >tI >The way to find out Mr. Rowan's view on whether he feels such an interimeG >situation would be necessary for him to address and if so how he wouldbC >address it is to ask him.  Lacking that, calling your suppositions E >'evidence' of the hypocrisy you claim exists (as opposed to multiplewD >alternate interpretations ranging from simply disagreeing with yourG >assessment of the 'interim' conditions to the possible naivete alreadynL >mentioned) demonstrates considerably less about him than it does about you. >b >- billi  L A simple translation of Bill's last few paragraphs is: I disagree with you, 5 but won't give up arguing until you see my viewpoint.t  K Bill, I respect and enjoy your posts.  But, please, don't do a CJL and use (> inflammatory language just because someone disagrees with you.  L Discuss/debate, but here no US political statements.  We all have to put up L with our politicians/news commentators, etc.  US folk are not the only ones K that read c.o.v./Info-VAX.  Did I ever tell you my feelings about Margaret n
 Thatcher ....n   Regards, Paddy   ------------------------------  # Date: Fri, 05 Jul 2002 11:24:46 GMTu* From: "Bill Todd" <billtodd@metrocast.net>Y Subject: Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar  C Message-ID: <2mfV8.281572$_j6.13873218@bin3.nnrp.aus1.giganews.com>a  2 <paddy.o'brien@zzz.tg.nsw.gov.au> wrote in message/ news:01KJQWE4CGG200072R@tgmail.tg.nsw.gov.au...a > Bill Todd said:b >c= > (I hope I've got the dead sergeants, i.e. attributes right)r >a> > >"Alan E. Feldman" <spamsink2001@yahoo.com> wrote in message9 > news:b096a4ee.0207041956.7e875e46@posting.google.com...a: > >> "Bill Todd" <billtodd@metrocast.net> wrote in message? > news:<4_8U8.81586$Ca2.4755216@bin2.nnrp.aus1.giganews.com>...nB > >> > "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in message> > >> > news:b096a4ee.0207011838.3539685d@posting.google.com... > >> > >... >e > >> > > Am I missing somethingr > >> > > here? > >> >I > >> > Yes.  And since it's been called to your attention already, eitherM
 > you're aA > >> > bit slow or not really making the effort to understand it.a > >> > >>$ > >> Let's refrain from the insults. > >lL > >Feel free to act as you see fit; as for me, I'll decide on a case-by-caseJ > >basis.  When you ask a question, be prepared for a straight answer:  if yousL > >hadn't attempted to be cute in asking it, I might have been less blunt in > >responding to it. >r+ > And why do we need so much OT stuff here?   K It's clearly a matter of 'want', not of 'need'.  If no one raised off-topiciL issues, there'd be zero incidence of them.  If no one responded to those whoA raise off-topic issues, there'd be very little incidence of them.e  G C.o.v. is (at least usually) a fairly closed community.  When off-topic.J discussion occurs, it's usually because at least several of the members ofB that community choose to support it (as you now seem to be doing).      A simple translation of Bill'sI > last paragraph is: I disagree with you, but won't give up arguing untilu youJ > see my viewpoint.0  F You really need to take a Berlitz course if you think that's *anythingI close* to what I said:  perhaps Alan's penchant for reading what he wants 3 into material rather than what it says is catching.X   >  > >> > >>C > >> > He advocates a change that (he believes) would eliminate the>	 perceivedf > needH > >> > to defend one's home with a gun.  *In the absence of that change* (i.e.,K > >> > under current conditions), he clearly himself does feel such a need.AK > >> > There's no contradiction (because they're two different situations),h > hencep > >> > no hypocrisy. > >> > >>F > >> Uh, perceived need is not equivalent to real need. People who areK > >> paranoid may have perceived needs for owning a gun in situations wherew! > >> one doesn't really need one.C > >>H > >> And your missing situation number three: the interim period betweenI > >> such a law taking effect and a large reduction in crime. During thattJ > >> period, it would still be prudent to own a gun. And Mr. Rowan clearlyJ > >> says that when faced with such dangers he would own a gun. Therefore,@ > >> he says that he would own a gun during that interim period. > >iJ > >You're really full of shit, you know.  Or completely ignorant of simpleE > >logic.  What Mr. Rowan said is what he said, no more, no less, and-H > >attempting to read into it support for your own viewpoint rather than stick.D > >to it literally is the mark of demagoguery rather than intellect. > >AK > >The way to find out Mr. Rowan's view on whether he feels such an interim I > >situation would be necessary for him to address and if so how he wouldlE > >address it is to ask him.  Lacking that, calling your suppositionsSG > >'evidence' of the hypocrisy you claim exists (as opposed to multiplelF > >alternate interpretations ranging from simply disagreeing with yourI > >assessment of the 'interim' conditions to the possible naivete alreadyaI > >mentioned) demonstrates considerably less about him than it does aboutt you. > >y	 > >- billt > H > A simple translation of Bill's last few paragraphs is: I disagree with you,7 > but won't give up arguing until you see my viewpoint.,  A By George, you've done it again!  This disease may be dangerouslym
 communicable.c  L When someone posts something I disagree with, I may choose to respond to it.H But I never *demand* that they convert to my viewpoint:  if they insteadI just choose to shut up, so will I (and sometimes I just shut up myself ifaE they're so moronic that further replies are clearly a waste of time).    >lL > Bill, I respect and enjoy your posts.  But, please, don't do a CJL and use@ > inflammatory language just because someone disagrees with you.  L I seldom use inflammatory language *just* because someone disagrees with me:D it's usually because of the *manner* in which they choose to expressH themselves.  Building vaporous edifices out of the imagined position Mr.J Rowan was asserted to hold in an area he had not even touched upon was theK impetus above:  when one takes such extensive liberties in representing theJI opinions of someone else who is not present to defend himself, one shouldd$ expect to be brought up a bit short.  L Of course, after a full year of combatting the garbage put out by Compaq andL its toadies my tolerance level in general is considerably lower than normal.D So when I bring someone up short, it may be by more than just a bit.   >o6 > Discuss/debate, but here no US political statements.  G Really?  Did I miss your appointment as official c.o.v. censor?  If yousE meant that as a request rather than as a proclamation, you phrased itA poorly.g   - bill   ------------------------------  # Date: Fri, 05 Jul 2002 12:05:26 GMT 1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie)IY Subject: Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar i; Message-ID: <aYfV8.43589$q53.1208506@twister.austin.rr.com>s  ) Bill Todd (billtodd@metrocast.net) wrote:  : J : The way to find out Mr. Rowan's view on whether he feels such an interimH : situation would be necessary for him to address and if so how he would : address it is to ask him.r :x  2    http://www.cnn.com/2000/US/09/23/obit.rowan.ap/F    CNN.com -  Columnist Carl Rowan dies at age 75 - September 23, 2000  E   "WASHINGTON (AP) -- Carl Thomas Rowan, a well-known commentator and-@    nationally syndicated newspaper columnist who was once calledG    America's "most visible black journalist," died Saturday. He was 75.   H    LeRoy Tillman, a spokesman for Washington Hospital Center, said RowanH    died of natural causes in the hospital's Intensive Care Unit at about    3 a.m..."    2 --Jerry Leslie   (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for emailo   ------------------------------   Date: 5 Jul 2002 06:32:26 -0700o. From: SPAMSINK2001@YAHOO.COM (Alan E. Feldman)Y Subject: Re: [OT]  Carl Rowan and guns (was Re: A possible shift in the status of VMS ar  = Message-ID: <343f30ae.0207050532.3f1253bd@posting.google.com>   t "Bill Todd" <billtodd@metrocast.net> wrote in message news:<EYaV8.559231$%y.36742574@bin4.nnrp.aus1.giganews.com>...= > "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in messager9 > news:b096a4ee.0207041956.7e875e46@posting.google.com...d9 > > "Bill Todd" <billtodd@metrocast.net> wrote in messageL@ >  news:<4_8U8.81586$Ca2.4755216@bin2.nnrp.aus1.giganews.com>...A > > > "Alan E. Feldman" <spamsink2001@yahoo.com> wrote in message = > > > news:b096a4ee.0207011838.3539685d@posting.google.com.../ >  > ...= >  > > > > Am I missing something
 > > > > here?. > > > H > > > Yes.  And since it's been called to your attention already, either >  you're ae@ > > > bit slow or not really making the effort to understand it. > >] > >c# > > Let's refrain from the insults.1 > K > Feel free to act as you see fit; as for me, I'll decide on a case-by-case M > basis.  When you ask a question, be prepared for a straight answer:  if you K > hadn't attempted to be cute in asking it, I might have been less blunt in3 > responding to it.h    , I don't recall being "cute". Sorry if I was.    L > > > He advocates a change that (he believes) would eliminate the perceived >  needrN > > > to defend one's home with a gun.  *In the absence of that change* (i.e.,J > > > under current conditions), he clearly himself does feel such a need.J > > > There's no contradiction (because they're two different situations), >  hence > > > no hypocrisy.  > >- > >-E > > Uh, perceived need is not equivalent to real need. People who arenJ > > paranoid may have perceived needs for owning a gun in situations where  > > one doesn't really need one. > > G > > And your missing situation number three: the interim period between6H > > such a law taking effect and a large reduction in crime. During thatI > > period, it would still be prudent to own a gun. And Mr. Rowan clearlydI > > says that when faced with such dangers he would own a gun. Therefore,c? > > he says that he would own a gun during that interim period.) > I > You're really full of shit, you know.  Or completely ignorant of simplehD > logic.  What Mr. Rowan said is what he said, no more, no less, andM > attempting to read into it support for your own viewpoint rather than stick C > to it literally is the mark of demagoguery rather than intellect.n    E This argument is bogus. You rule out simple conclusions based on whatlA he said simply because he didn't say them. I find that illogical.r   Demogoguery? From Webster.com: E 1 : a leader who makes use of popular prejudices and false claims andw promises in order to gain powerhB 2 : a leader championing the cause of the common people in ancient timesi  E I am not trying to gain power and I am not in ancient times. And justaC to be clear, I am not advocating gun control nor am I trying to say:@ anything directly against it. I am only referring to Mr. Rowan's> remarks and am saying that at the very list there is a hint of0 hypocrisy and that they are not self-consistent.    J > The way to find out Mr. Rowan's view on whether he feels such an interimH > situation would be necessary for him to address and if so how he would > address it is to ask him.     C The interim period is not optional. There *will* be a period duringuD which gun ownership by all private citizens is illegal and the crimeF rate will still be unacceptably high. And since there *will* be such aE period, and Mr. Rowan clearly stated he would protect his family when D the crime rate is "too high", he'll be owning a gun illegally duringA that period. If you think that there might not be such an interimi period, then you are an idiot.    ) > Lacking that, calling your suppositionsiF > 'evidence' of the hypocrisy you claim exists (as opposed to multipleE > alternate interpretations ranging from simply disagreeing with your H > assessment of the 'interim' conditions to the possible naivete alreadyM > mentioned) demonstrates considerably less about him than it does about you.r >  > - bill     That's your opinion. t  A And since you insist on uncalled for insults Mr. Know-it-all, whyo/ should everyone, or even anyone, listen to you?e     Disclaimer: JMHO Alan E. Feldman  afeldman gfigroup com    ------------------------------   End of INFO-VAX 2002.367 ************************