1 INFO-VAX	Tue, 16 Apr 2002	Volume 2002 : Issue 209       Contents:6 Re: A simple way to manipulate the date/time required.  Re: Alphaserver EV7 announcement  ANN: FTP_MIRROR update available0 Re: another hint of post-merger direction, maybe0 Re: another hint of post-merger direction, maybe0 Re: another hint of post-merger direction, maybe' Re: BACKUP causes crash on VMS V5.5-2H4  Re: Best DIGITAL product ever? Re: Best DIGITAL product ever? Re: Best DIGITAL product ever? Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: Blade architectures  Re: CLD problem (flight) Re: cluster  Re: CMS library repair; Re: Could increasing kernel stack size improve performance?  RE: Daylight savings Re: DCL command procedure  RE: Disaster Recovery  Re: Disk Monitoring  Re: Evolution of the DEC mouse RE: Evolution of the DEC mouse Re: Evolution of the DEC mouse+ Re: Excursion PCs as X-Terminals on OpenVMS + Re: Excursion PCs as X-Terminals on OpenVMS  Re: Exporting SYSUAF.LIS Re: Exporting SYSUAF.LIS  Re: Fearless VMS Prognostication full file backup to disk Re: full file backup to disk FW: cluster  FW: cluster  FW: cluster  Re: GNU cc vs VMS CC listings  Re: GNU cc vs VMS CC listings  Hobbist VMS Source ListingF Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxF Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxF Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinux- Re: I was right!  Alpha will live in Itanium! - Re: I was right!  Alpha will live in Itanium!  Re: IA64 is not the VAX  Re: IA64 is not the VAX  Initial Sequence Numbers irc chat channel #vms on efnet Itanium Oddities Re: Itanium troubles Re: Itanium troubles! Re: Java with the WASD webserver. > Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles)> Re: learning how to use a computer (was: Re: Itanium troubles) Re: LVD SCSI on PWS 500au  Re: LVD SCSI on PWS 500au # Mixing Alphas and PCs on KVM switch  MMOV$ALPHAVCR.EXE: bad colors?# OS400 in similar non-marketing bind ' Re: OS400 in similar non-marketing bind ( Re: OT Best Practises in the IT Industry( Re: OT Best Practises in the IT Industry( Re: OT Best Practises in the IT Industry0 Re: Prediction:  VMS lives, merger or no merger!0 Re: Prediction:  VMS lives, merger or no merger! Re: Printserver 32 Error Code  Re: Processes in state RWSCSA Question about system messages, $SET MESSAGE, MESSAGE utility ... E Re: Question about system messages, $SET MESSAGE, MESSAGE utility ...  Re: RAM for MicroVAX 3300  Re: remote mailbox0 Re: Sue, a great promotion for "unhackable" vms!0 RE: Sue, a great promotion for "unhackable" vms!0 Re: Sue, a great promotion for "unhackable" vms!0 Re: Sue, a great promotion for "unhackable" vms!0 Re: Sue, a great promotion for "unhackable" vms!! Re: Survey on Your Alpha Hardware ! Re: Survey on Your Alpha Hardware " Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs" Re: Throughput and additional CPUs. Re: VMS "unhackable" example just for Andy .... RE: VMS "unhackable" example just for Andy ...! Re: VMS 7.2-1 to 7.2.2- Follow Up ! Re: VMS 7.2-1 to 7.2.2- Follow Up  we GUARANTEE you a check! ( Re: Why VMS is "unhackable" lesson 1 ...  F ----------------------------------------------------------------------  % Date: Tue, 16 Apr 2002 12:45:31 +1000 = From: "Mark\(unMASK\)Forsyth" <forsytMhm@optAushoSme.com.aKu> ? Subject: Re: A simple way to manipulate the date/time required. " Message-ID: <d83g9a.2rc.ln@really>  D "nortools" <norma.toolsie@bartsandthelondon.nhs.uk> wrote in message7 news:27d25fc7.0204140812.397340d8@posting.google.com...  > Hi, G > Does anyone know a "simple" way to extract the date and time and then A > remove all the superfluous hyphens, colons, full-stops, spaces, G > leaving just the numbers and letters, e.g: 14APR1711. This is without B > having to resort to lots of symbol creations and using f$locate,H > f$extract to manipulate the created symbols. My latest procedure worksG > but runs to about 14 lines. I am sure the same result can be achieved  > in less lines than this.  J Assuming you're after something in the format DDMMMHHMM why not just do :- $ aaa = f$edit(f$cvt(,,"day")+-  _$ f$cvt(,"absolute","month")+-  _$ f$cvt(,,"hour")+- _$ f$cvt(,,"minute"),"trim") $ sh sym aaa   AAA = "16APR1237"  $    Ooroo 	 Mark F...        > Thanks > N    ------------------------------  # Date: Mon, 15 Apr 2002 20:42:03 GMT # From: "John Smith" <a@nonymous.com> ) Subject: Re: Alphaserver EV7 announcement E Message-ID: <vWGu8.1558$Gq1.874@news02.bloor.is.net.cable.rogers.com>   + Not unheard of.  My condolences, of course.     @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message& news:a9f1c4$fl21$1@lead.zk3.dec.com...K > Apparently Windows crap I was using to read/write the group had some type  of	 > attack.  >  > John Smith wrote in message > > <8PBt8.1549$OI1.767@news02.bloor.is.net.cable.rogers.com>...! > >Are you sure about that?   ;-)  > >  > > C > >"Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message ) > >news:a94vdf$evvq$1@lead.zk3.dec.com...  > >># > >> Bill Todd wrote in message ...  > >> > > >> >L > >> >That said, I have no idea how much it would cost to shrink EV6 (though > >theL > >> >fact that the core is almost identical to EV7's might allow it to ride > onI > >> >the coattails of the EV7 shrink), and hence whether that would make 	 > >sense.  > >> > > >>4 > >> Remember that EV7 uis built around a EV68 core. > >> > >> > >> > >> > >> > >  > >  >  >    ------------------------------  # Date: Tue, 16 Apr 2002 02:31:59 GMT - From: goathunter@goatley.com (Hunter Goatley) ) Subject: ANN: FTP_MIRROR update available 1 Message-ID: <3cbb8c1c.214816660@news.process.com>   0 FTP_MIRROR V1.0-4 is now available for download.  D FTP_MIRROR is a DCL command procedure that can be used to keep localE directories in sync with directories on a remote FTP server (in other < words, to maintain a mirror for the remote FTP directories).  F This release adds support for HGFTP, a new PURGE configuration option,C and the generation of a mail message if the local directories don't F map with the remote directories (i.e, if there's a directory structure@ change on the remote end, a message will warn you of that fact).  < You can download FTP_MIRROR V1.0-4 using the following URLs:   http://www.process.com/openvms/   : ftp://ftp.process.com/vms-freeware/fileserv/ftp_mirror.zip? http://ftp.process.com/ftp/vms-freeware/fileserv/ftp_mirror.zip 6 ftp://ftp.tmk.com/vms-freeware/fileserv/ftp_mirror.zip7 http://www.tmk.com/vms-freeware/fileserv/ftp_mirror.zip   2 And on the other mirrors within the next 24 hours.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ 8 goathunter@goatley.com    http://www.goatley.com/hunter/< New Robert R. McCammon site: http://www.RobertRMcCammon.com/   ------------------------------    Date: 15 Apr 2002 20:02:49 -0000= From: Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> 9 Subject: Re: another hint of post-merger direction, maybe 5 Message-ID: <20020415200249.6544.qmail@gacracker.org>   K On Mon, 15 Apr 2002, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote:    <snip>  K >Get a life.  There are many, many happy GS users.  And they have one thing  >you'll never have - VMS.   2 Did you have to make me choke on my beer Fred? :-)     Doc. --  6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                              http://vmsbox.cjb.net    ------------------------------  # Date: Mon, 15 Apr 2002 20:43:52 GMT * From: "Bill Todd" <billtodd@metrocast.net>9 Subject: Re: another hint of post-merger direction, maybe A Message-ID: <cYGu8.47885$GS6.5118286@bin3.nnrp.aus1.giganews.com>   @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message& news:a9f1o9$280r$1@lead.zk3.dec.com...: > Aw, and I thought you said you *weren't* a market-droid. > D > Come on, admit it.  You're in sales right?  Never really have done anythingE > technical.  Right?  You getting a lot of business from these posts?   L Andrew has already replied to that at least once by citing work he has done.G And while some of his statements do tend to shed doubt on his technical H competence, they do so no more than a lot of your recent statements shed doubt on your own.  B Sun probably doesn't get too much business as a result of Andrew'sI participation here (which would be less offensive if he stuck more to the L real grains of truth many of his observations contain):  they likely get farG more business as a result of Compaq's actions than as a result of *any*  activity in c.o.v.   - bill   ------------------------------    Date: 15 Apr 2002 17:41:22 -0700( From: bob@instantwhip.com (Bob Ceculski)9 Subject: Re: another hint of post-merger direction, maybe = Message-ID: <d7791aa1.0204151641.7faa4254@posting.google.com>   i "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message news:<a9f1o9$280r$1@lead.zk3.dec.com>... : > Aw, and I thought you said you *weren't* a market-droid. > M > Come on, admit it.  You're in sales right?  Never really have done anything E > technical.  Right?  You getting a lot of business from these posts?  >   B after reading Andrews posts, I would think they would cause him to lose business!   ------------------------------  # Date: Tue, 16 Apr 2002 02:49:09 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)0 Subject: Re: BACKUP causes crash on VMS V5.5-2H49 Message-ID: <FiMu8.42$Nc.1330267@cacnews.cac.cpqcorp.net>   b In article <3cb2f7bc$1@news.si.com>, "Brian Tillman" <tillman_brian@notnoone.notnohow.com> writes:? :I've an older VAX 4000-500A running VMS V5.5-2H4..  It has the M :VAXY2K01_U2055 and VAXF11X06_U2055 ECOs applied.  Yet, when I try to make an K :image BACKUP of certain disks, the system crashes.  I've looked at all the K :ECOs I can find looking for another to address this problem, but the above M :two seem the most likely, and they're already applied, as far as I can tell.   K   What does the system crash with and what is the symbolic for the failing  L   program counter (PC) address, and what details can you provide concerning H   the particular (failing and successful) disks, which controller(s) areK   involved, and does (for instance) ANALYZE/DISK have anything interesting  G   to report?  Anything written to the error logs around the crash time?       N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 15 Apr 2002 22:19:09 +0100 + From: "antonio.carlini" <arcarlini@iee.org> ' Subject: Re: Best DIGITAL product ever? & Message-ID: <3CBB43CD.8020003@iee.org>   Jan-Erik Sderholm wrote:   E > Someone sad that DTR-11 was the most overlayed image to ever run on B > the PDP/RSX. I used it on a 11/84 to produce a number of reports5 > from mostly sequentional files. Realy nice product.     * Time for "A Day In the Life ..." methinks:  = http://www.arnes.si/news/archive/si.comp.os.vms/msg00126.html   - Must have been quite frightening on an 11 :-)      Antonio    --     --------------- - Antonio Carlini             arcarlini@iee.org    ------------------------------  % Date: Mon, 15 Apr 2002 21:40:16 -0400 ( From: Bill Gunshannon <bill@cs.uofs.edu>' Subject: Re: Best DIGITAL product ever? B Message-ID: <20020415213840.L14125-100000@server2.cs.scranton.edu>  * On Sat, 13 Apr 2002, Didier Morandi wrote:   > Let's see: > 9 > PDP, DECsystems, VAX and Alpha-based computers are dead   - Mentec is still developing and selling PDP's.   + > RT11, RSX, RSTS, TOPS and VAXeln are dead   9 RT11, RSTS/E and RSX are all current and available today.    Just not from Compaq.    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |> Scranton, Pennsylvania   |         #include <std.disclaimer.h>   ------------------------------   Date: 16 Apr 2002 04:34:47 GMT2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>' Subject: Re: Best DIGITAL product ever? , Message-ID: <a9g9l701ljr@enews4.newsguy.com>  ) Bill Gunshannon <bill@cs.uofs.edu> wrote: , > On Sat, 13 Apr 2002, Didier Morandi wrote:  
 >> Let's see:  >>: >> PDP, DECsystems, VAX and Alpha-based computers are dead  / > Mentec is still developing and selling PDP's.   , >> RT11, RSX, RSTS, TOPS and VAXeln are dead  ; > RT11, RSTS/E and RSX are all current and available today.    > Just not from Compaq.   J Also, you can still buy a system running TOPS-20 from XKL, in fact I thinkF they're working on a new model from something I read a few months ago.   			Zane    ------------------------------  % Date: Mon, 15 Apr 2002 18:28:18 +0100 T From: Andrew Harrison SUNUK Consultancy <andrew_nospam.remove_this.harrison@sun.com>  Subject: Re: Blade architectures& Message-ID: <3CBB0DB2.9040303@sun.com>   Peter da Silva wrote:   4 > In article <8gk7rau8ld.fsf@galapagos.cse.psu.edu>,> > Scott Schwartz  <"schwartz+@usenet "@bio.cse.psu.edu> wrote: > * >>peter@abbnm.com (Peter da Silva) writes: >>J >>>>In the Linux world there seem to be people who insist that it's reallyE >>>>important to have a swap that's at least 3x the size of physical.  >>>> > 4 >>>Hmm. 2x seems to be the rule of thumb in FreeBSD. >>>  > D >>On solaris, which has a very nice VM-based /tmp filesystem, havingA >>lots of extra swap space is useful even if no program uses much 	 >>memory.  >>     > I > There's a lot of systems that have a tmpfs like that. I'm not sure what M > advantage it has over a regular UFS partition mounted "async" and recreated ; > on boot, so long as you have a good unified VM subsystem.  >     > Well tmpfs never actually does any disk I/O unless you run out7 of memory so it can be faster than UFS or some other FS : mounted async. You can limit the size of tmpfs if you want; to ensure that you don't eat too much memory or swap space.    Regards  Andrew Harrison    ------------------------------    Date: 16 Apr 2002 01:49:55 +0800, From: Paul Repacholi <prep@prep.synonet.com>  Subject: Re: Blade architectures- Message-ID: <87ads43l8c.fsf@prep.synonet.com>   - Tim Shoppa <shoppa@trailing-edge.com> writes:   B > If the margin is related to what you actually need, that's fine.: > But Jesus Christ, they could simultaneously stream everyF > pornographic movie ever made through 8 OC-3's!  (No, I did not blurt* > that out at the meeting, but it's true!)  ? "Oh my, you could have a profitable online business with them."    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------   Date: 15 Apr 2002 19:38:53 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)  Subject: Re: Blade architectures0 Message-ID: <a9fa8d$gpe$1@pegasus.csx.cam.ac.uk>  - In article <a9ed5l$73b@web.eng.baileynm.com>,-' Peter da Silva <peter@abbnm.com> wrote: 0 >In article <3CBA0F55.6030500@brussels.sgi.com>,- >Alexis Cousein  <al@brussels.sgi.com> wrote:rB >> I'm not saying you need them often -- but *when* you need them,? >> you'd better have them at hand - I don't climb rocks without E >> a safety rope either, even though the intent is never to rely uponr	 >> it ;).i >iI >On the other hand, you don't need a safety rope when you're not climbing  >rocks.R >vG >If you're running a real-time control system where any crash has to be E >analysed and replicating it would involve shutting down a productionXL >system that's supposed to be "five 9s", you need a safety rope and a spare.  @ Hence my hobby horse of automatic, enabled (by default), tracing@ and diagnostic facilities.  NOT "rebuild for debugging and rerun under the test harness" :-(W  I >If you're running a routine task on a generic PC with spares around, andoF >a crash is almost always a hardware problem and the easiest fix is toL >replace the machine, a safety rope would be like packing a parachute before >going above the 2nd floor.v  E If you are doing that, WHAT operating system are you running?  I wanti' it now!  Yesterday.  Pretty please ....   G Frankly, half of the hardware crashes that I see ARE software problems.4B If the software wasn't such a pig's ear, it would have handled theD (often isolated and well-defined) hardware problem and allowed me toC at least run the diagnostic tools that the system comes with (well,r$ it does come with them, doesn't it?)     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 334679o   ------------------------------  # Date: Mon, 15 Apr 2002 22:20:06 GMTa7 From: ssthapa@classes.cs.uchicago.edu (Suchandra Thapa)t  Subject: Re: Blade architectures8 Message-ID: <slrnabmjvg.2tv.ssthapa@hepcat.telocity.com>  + Floyd Davidson <floyd@ptialaska.net> wrote::I >>    Actually, linux itself is pretty bad about handling virtual memory.oL >>The 2.4 kernels needed the swap space to be at least twice the size of theM >>physical memory due to the way the VM works.  Swap space isn't reclaimed sosM >>pages can be duplicated in memory and swap.  So having swap space the same aO >>size as your physical memory doesn't do anything if you are swapping heavily.o > A > That was true for a very short time, but is not true of current  > 2.4 kernels.  K     Which kernel got rid of the duplication of the swap pages?  I was underMN the impression that it might be addressed in 2.5 but not before then?  Was it  in the VM change in 2.4.10?    -- BL ----------------------------------------------------------------------------& Suchandra Thapa                       & s-thapa-11@NOSPAMalumni.uchicago.edu  L ----------------------------------------------------------------------------   ------------------------------    Date: 15 Apr 2002 15:52:22 -07003 From: Eric Smith <eric-no-spam-for-me@brouhaha.com>n  Subject: Re: Blade architectures0 Message-ID: <qh3cxwziah.fsf@ruckus.brouhaha.com>  9 ssthapa@classes.cs.uchicago.edu (Suchandra Thapa) writes:kI >     Actually, linux itself is pretty bad about handling virtual memory.eL > The 2.4 kernels needed the swap space to be at least twice the size of the. > physical memory due to the way the VM works.  E False.  I've run a lot of different 2.4 kernels with ZERO swap space,r and never had a problem.  H I think what you meant to say about those kernels is that if you DID useE swap space, you needed more swap than you had physical, or you got nod benefit.   ------------------------------  % Date: Mon, 15 Apr 2002 21:30:51 -0500 $ From: "del cecchi" <dcecchi@msn.com>  Subject: Re: Blade architectures3 Message-ID: <12Mu8.679$jj4.30999@eagle.america.net>d  7 "Dowe Keller" <dowe@krikkit.127.0.0.1> wrote in messageM( news:m3hemd1vp9.fsf@krikkit.127.0.0.1...7 > Eric Smith <eric-no-spam-for-me@brouhaha.com> writes:. > G > > That's not an acceptable excuse for installers that REFUSE to allowuF > > the user to not set up a swap, or that make it really difficult to
 > > avoid it.- >-D > For every development group, there will be *at least* one fasistic > moron. > E > Surely, this has to be a law atributed to someone (It's too obviousA
 > not to be).m >p > --4 > dowe@sierratel.com <http://www.sierratel.com/dowe>G > In less than a century, computers will be making substantial progresss on. > ... the overriding problem of war and peace. > -- James Slagle   H If you are going to crosspost drivel, at least learn how to spell, or at' least learn how to use a spell checker.s  ! Thank you for your consideration.t   ------------------------------  # Date: Tue, 16 Apr 2002 03:05:31 GMTe2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)! Subject: Re: CLD problem (flight)a9 Message-ID: <%xMu8.44$Nc.1330267@cacnews.cac.cpqcorp.net>   _ In article <3cb33884.8082221@news.process.com>, goathunter@goatley.com (Hunter Goatley) writes: - :Not a bug.  You see that with DELETE/SYMBOL:  :t :  $ del/sym/glob yyzfE :  %DCL-I-IGNQUAL, qualifiers appearing before this item were ignored-	 :   \SYM\, :iM :I'm not sure how that gets set in the CLD file, though.  I've never wondered  :enough to go look for it.    I   This is established by the SYNTAX keyword within the CLD file -- when amH   qualifier switches the parsing over to a wholely new verb SYNTAX, thisI   message warns that qualifiers have been ignored during the CLI command e   parsing operation.    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 16 Apr 2002 01:01:14 GMTs) From: rob.buxton@wcc.govt.nz (Rob Buxton)  Subject: Re: cluster2 Message-ID: <3cbb76f5.2234016936@news.wcc.govt.nz>  F Think you're spot on there Vinit, I think the disk issue is definately/ secondary to this not actually being a cluster.t  / Sounds like they have two separate VMS Systems.l  @ Once the Cluster is formed there's also a couple of other SYSGEN  Parameters that may be required.
 MSCP_LOAD = 1  MSCP_SERVE_ALL = 1  D I think these may be needed to Serve the Disks to other nodes in the Cluster.  = On 15 Apr 2002 08:46:56 -0700, vadya@ibjus.com (Vinit) wrote:2  v >A Bonaveidogo <Asena@fsc.com.fj> wrote in message news:<BFBEDDF2CFEDD411917400508BF3A6FFCBCB01@exchsvr.FSC.COM.FJ>...M >> I have two node (A & B)  forming a cluster and  both connected on shadoweddL >> environment.  On node A I have $1$dka0, on node B its $2$dka0.  these twoO >> disks are shadowed. When I boot each node, it couldn't recognised the member,9 >> of the shadow set.  Its says no such device available.hN >> On Node A... this error message appeared "$2$dka0 failed as a member of the >> shadow set"? >> On Node B ... "$1$dka0 failed as a member of the shadow set"i >> rL >> When I do a show cluster on each node....it shows only the  current node. >> lB >> Can someone advise how to configure these two nodes properly.   >> f >> Asena >: >Hello,/ >wG >If "show cluster" shows only one system, you do not have a VMS cluster_ >configured on the nodes.e >You will have to run 6 >"@sys$manager:CLUSTER_CONFIG.COM" to set up clusters.F >You will ned to know What kind of cluster interconnect are you using. >o >cheers  >Vinit   ------------------------------  # Date: Tue, 16 Apr 2002 02:17:56 GMTl2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: CMS library repairg9 Message-ID: <oRLu8.39$Nc.1330267@cacnews.cac.cpqcorp.net>.  Z In article <3CAE4F13.29039.540658@localhost>, "Stanley F. Quayle" <stan@stanq.com> writes:@ :When I do a CMS VERIFY/REPAIR on some CMS libraries, I get the  :following:h :h? :%CMS-E-ERRVER2, internal contiguous space verified with errors.B :-CMS-E-MSSBLKSTR, 4 35 type blocks found on pass 1, and 1 blocks  :found on pass 24 :%CMS-E-ERRVERCON, control file verified with errors : 5 :However, the library still seems to "work" normally.y    5   That looks to be a (relatively benign?) corruption.n     F :These CMS libraries are accessed by different versions of CMS across 5 :my cluster.  Could there be some problems with that?     ?   I haven't tried this combination.  Use of the same version is @   prefered, but I have seen multiple versions used -- that said,?   there was an incompatible upgrade a couple of revisions back,i   when 00CMS turned into 01CMS.8     :The versions are: :s :   VAX VMS 7.3		CMS 4.1 :   Alpha VMS 7.3	CMS 4.0r :   VAX VMS 7.2		CMS 3.7-2 : F :Should I upgrade my 3.7-2 to 4.0 or 4.1?  Will that make my problems 	 :go away?a       I'd tend to assume not.a  )   You appear to have a corrupted library.y  E   I'd get in contact with the support center, and see about acquiring $   some tools to rebuild the library.    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 16 Apr 2002 01:39:28 GMTh2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)D Subject: Re: Could increasing kernel stack size improve performance?9 Message-ID: <khLu8.35$Nc.1330267@cacnews.cac.cpqcorp.net>-  f In article <dd3f0cb7.0204021627.52a20830@posting.google.com>, chessmaster1010@hotmail.com (JG) writes: ..A :My real question is if this is true then would I see performance4C :improvements on systems running a webserver just by increasing the@ :size of the kernel stack? s  J   I'd encourage following the performance guide's guidelines, rather than J   making assumptions -- the key to performance management is a methodical K   approach, tweak one control at a time, and use and maintain a collection  M   of performance data for the purpose of relative comparisions.  The OpenVMS kJ   performance guide provides an overview on how to gather data and how to    tune the system.  G :                            We have AlphaServer 1000s that seem to hitjG :a bottleneck serving cached graphics when there is plenty of free CPU.3  ?   "Serving cached graphics"?   Via X Windows?  Webserver?  SCS?m  G   The AlphaServer 1000 graphics via the integrated graphics controller oI   are resource-intensive -- if you are using graphics on this system and  H   don't already have a PCI graphics controller installed, you will want J   to consider it.  (Also please realize this system itself is now old and L   slow and often memory-constrained.)  That said, you will want to identify A   the bottleneck before you can effectively attempt to remove it.   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 16 Apr 2002 02:26:32 GMT-2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: RE: Daylight savingso9 Message-ID: <sZLu8.40$Nc.1330267@cacnews.cac.cpqcorp.net>p  _ In article <CIEJLCMNHNNDLLOOGNJIAEFNELAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes: < :Well I think it was a 6.1 system, it was on a spare drive.   @   Spare drive?  As in, "not booted from it" when you tried this?  O :I looked through [000000]directory.lis but didn't see any obvious indicator ofIH :version.  welcome.txt says 6.1, but it could be old.  I copy also from  :a 7.3 system   G   Please.  Don't.  Please.  Please don't copy files around like that.   D   Please.  Please don't get "creative".  Please?  (If you have been,D   um, "creative", with this system disk, I would strongly encourage D   OpenVMS be reinstalled.  Why?  Because you'll risk chasing various@   weirdness for years.  I'd also get off V6.1, but I digress...)    / :NORNS> @SYS$SYSROOT:[SYSMGR]UTC$TIME_SETUP.COMn :i" :  Configuring the Local Time Zone :  : 1 :  TIME ZONE SPECIFICATION -- MAIN Time Zone Menu  :2> :%DCL-W-UNDSYM, undefined symbol - check validity and spelling : \ROW1\> :%DCL-W-UNDSYM, undefined symbol - check validity and spelling :0
 :      0) GMTr :oD :Select the number above that best represents the desired time zone: : , :So I guess I am still missing a few things.    E   Among these, the trivia that V6.1 did not have a TDF nor particularS   support for timezones.  E   Please post specific details of the OpenVMS release and the OpenVMS ;   platform, and please post details of the problem(s) seen.       N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 16 Apr 2002 02:14:23 GMT02 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)" Subject: Re: DCL command procedure9 Message-ID: <3OLu8.38$Nc.1330267@cacnews.cac.cpqcorp.net>   X In article <a8lmkd$nth2@rain.i-cable.com>, "Kenneth" <yeung_kenneth@hotmail.com> writes:0 :Is the $ sign a must in the command procedure?      Yes.  K   Compaq OpenVMS documentation indicates all DCL commands in a DCL command oH   procedure file must include the $ sign; that the $ is required syntax.  G :It's working even I don't give the $ sign on each command line in the m
 :procedures. 	  I   Yours is an interesting approach and an interesting conclusion -- your 0L   DCL command procedure apparently works without the inclusion of the $, so K   you conclude that it is clearly not needed.  That's certainly apparently @L   quite reasonable for your case, but I would not assume that your DCL code M   would work in other cases, nor that your DCL code would survive an OpenVMS  I   upgrade, nor that you could safely generalize the requirement (or lack eK   thereof) for the $ delimiter.  Why?  Because you are doing something thatpI   the documentation does not support -- and because these interfaces tendeK   to be tightened over time, and there are traditionally improvements made  L   in syntax checking and in better diagnostics over time.  (You have clearly!   been lucky here, of course. :-)   ' :Then what is the use of this "$" sign?   J   The $ is used to indicate a DCL command, and thus to delimit the end of I   the application data for any application that is reading data from the eH   input stream.  This allows DCL to terminate input into the applicationF   (the $ is an implicit EOF when read by an application, much like theJ   ^Z) and the $ is also used resynchronize on error; to find the next DCL    command in the input stream.   	--e     No offense is intended.d    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 15 Apr 2002 16:47:50 -0400y+ From: "Main, Kerry" <Kerry.Main@Compaq.com>- Subject: RE: Disaster RecoveryT Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4023D8E79@kaoexc01.americas.cpqcorp.net>   John,i  F As a fyi - interesting article on how many Customers focus on disaster/ recovery has changed in light of recent events.   
 Reference:5 http://www.informationweek.com/story/IWK20011121S0005t   Regards,  
 Kerry Main Senior Consultantl Compaq Canada Corp.h Professional Servicesu Voice: 613-592-4660, Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----+ From: John Smith [mailto:a@nonymous.com]=20u Sent: April 15, 2002 10:30 AM  To: Info-VAX@Mvb.Saic.Come Subject: OT: Disaster Recovery      6 "Main, Kerry" <Kerry.Main@Compaq.com> wrote in messageH news:BE56C50EA024184DAF48F0B9A47F5CF401AB1E6B@kaoexc01.americas.cpqcorp. net. .i Paul,b  = Re: Disaster backup sites overflowing with other Customers ..-  @ You are correct and there is some additional items to consider -  H 1. Disaster tolerant "rental" sites typically cost in the range of $200kC USD per year (every year) - just to have the equipment and space oneD standby. Now, if implemented, the daily rates are often in excess ofE $10K USD per day. Think about what the final costs would be to go out B and build / find a new site - optimistically 3 months. Now add theH complexity associated with many other Customers also looking for similar spare datacenter space.u  B 2. Based on issues discovered in point 1.), some Customers are nowB questioning whether 2 sites is enough i.e. perhaps 3 site clustersC (active-active-passive?) are the right approach for the future? Big B issue is that if the primary site goes away permanently, then thatF business must run  everything on a single site and that is a huge risk to their business.  H Question for future discussion - does the disaster recovery site you are5 renting from have a backup site? Some do, some don't.a  B 3. While obviously not for everyone, keep in mind that the cost ofF multi-site solutions for those that are looking at them typically haveH downtimes that mean even a few hours would pay for all the multi-site HWH and software. That does not even consider that Vendors credibility as anF online financial provider. Picture becomes even more critical when one@ considers the ever extending times for online financial trading.  A 4. While bandwidth and latency are always going to be issues with.D multi-site clusters, keep in mind that for large centers, Cisco (andF others in that space) already offer 10Gb WAN solutions today with 40Gb
 demonstrated.     F Many businesses carry 'business interruption' insurance policies. LikeG all insurance policies, there is a premium paid for this. Some insurers>? will lower the premiums on policies such as these if there is anD hot-backup site for critical business function, much in the same wayD that you can get lower premiums on your home insurance if you have aG monitored burglar/fire alarm system. So, while hot-backup is not cheap, + there can be some reduction in other costs.>    D Frankly I cannot understand why anyone would put a disaster recoveryC site so close to a primary site (NYC and Brooklyn/Jersey City as anuC example). I do know all the arguments about proximity of staff, and 5 other infrastructure. But this ignores other factors.t  D In 1979 there were two instances in which mass evacuations of entireG areas took place - downwind of the Three Mile Island nuclear plant, and-B another involving 250,000 residents of a suburb of Toronto, Canada> following a train derailment and a risk of widespread chemicalG contamination. Admittedly no computing sites in the areas were directly-H affected by explosions, etc... and that remote system administration wasG light-years behind capabilities available today. However, systems do goDD down, disks fail, trees fall on comms lines, etc....and if nobody isH allowed into an evacuated area and both your primary and hot-backup site$ are within that zone - you're toast.  @ The 1986 Chernobyl reactor explosion in the Ukraine left an areaG approximately the size of Iowa and Illinois contaminated with radiationaG at levels high enough to warrant concern. Depending on wind patterns atbH the time of the accident, some towns as much as 200 kilometers away wereC hit with enough fallout to warrant their permanent evacuation. It'suG difficult to find exact figures of the area immediately surrounding the ; reactor complex that is off-limits - most figures I've seenuC (http://www.bullatomsci.org/issues/1990/s90/s90gale.html  and othere> sites) indicate that an area with a radius of approximately 30H kilometers centered on the reactor is permanently evacuated - about 2900G square kilometers. Pity the business in any industrialized country thatcF has its primary/hot-site within a similar zone in a similar situation.  G Going a bit further along this thread, I know of many primary/hot-siteseG that are within the impact zone of the same geological fault, and whoseu@ structures are not earthquake reinforced. Not all building codes" contemplate earthquake resistance.  > If you're in 'Tornado Alley', maybe your backup site should be> underground, and also not in a flood plain of any major river.  E Not all of these criteria can be met by simply locating your hot-sitee 'across the river'.r  @ If you have not already done so, investigate your local disasterE recovery groups and the planning/learning resources available to you.-1 One suggestion of a place to begin is www.drj.comM   ------------------------------  # Date: Tue, 16 Apr 2002 01:47:39 GMTp2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: Disk Monitoring9 Message-ID: <%oLu8.36$Nc.1330267@cacnews.cac.cpqcorp.net>a  [ In article <pkemau865sahie27e9jfdbm50a5ugolo9b@4ax.com>, Jon <jsmyth69@hotmail.com> writes:l  C   When posting, please remember to provide the OpenVMS platform andm   operating system release.i  F : I was asked to create a disk i/o monitoring and I was REALLY wantingE :to avoid doing a monitor disk to an output file and read it in.  AreN) :there lexicals that can return disk io?    A   DCL lexicals usually don't gather this sort of information, itsh&   usually only available to programs.   + :I'm specifically looking for queue length.n  >   The supported approach involves the OpenVMS command, and its   recording capabilities.-  "     MONITOR DISK/ITEM=QUEUE_LENGTH  B   If you need something else, then you can get the same data as isE   displayed by MONITOR using the (undocumented) exe$getspi interface.n    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 16 Apr 2002 03:19:08 GMTy2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)' Subject: Re: Evolution of the DEC mouse 9 Message-ID: <MKMu8.45$Nc.1330267@cacnews.cac.cpqcorp.net>S  _ In article <CIEJLCMNHNNDLLOOGNJIKENOELAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes: 7 :the older mice had a circular 7 pin connector; whereasw5 :the newer ones have a 6 pin pc compatible connector.s  =   Differences in wiring and connectors are (usually) easy to .;   resolve.   Signals and protocols differ, which makes thist   effort rather more difficult.r  : :Does anyone know if it is possible to associate the wires7 :from one with the other?  The objective is to put somec8 :VAXs, DEC 3000/300 and DECstation 5000 on a KVM switch.  @   Theoretically yes, but I've never seen the electronics needed.B   Please see the FAQ for some background and pointers to low-levelA   signaling details.  Best to banish these older boxes from your >B   desktop, and use more recent systems and SET HOST or telnet (and@   remote X Windows displays, if needed) to connect to the older    boxes as required.  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 15 Apr 2002 21:04:26 -0700l# From: "Tom Linden" <tom@kednos.com>j' Subject: RE: Evolution of the DEC mouse 9 Message-ID: <CIEJLCMNHNNDLLOOGNJICEFMEMAA.tom@kednos.com>e  E Unfortunately I still need to hang on to some of these older systems.wB Aside from the support issue, I actually still do editing on a 6.2E system because (1) can't find a working copy of emacs for 7.3 and (2)f; it is easier to change programming language than editor :->n  F Since writing the original post, I have learned that most KVM switchedF make  assumptions about the signals, i.e. they just don't switch them C en masse.  For example, some don't get switched because they aren't E used by PS/2 so you can't use thos pins.  I think the least obtrusiverD and time consuming approach is to use a KVM switch for the video andH a manual db15 straight through switch for the KB and Mouse, that way you! worry about the specific signals.s   > -----Original Message-----; > From: Hoff Hoffman [mailto:hoffman@xdelta.zko.dec.nospam]C& > Sent: Monday, April 15, 2002 8:19 PM > To: Info-VAX@Mvb.Saic.ComC) > Subject: Re: Evolution of the DEC mouse  >  > A > In article <CIEJLCMNHNNDLLOOGNJIKENOELAA.tom@kednos.com>, "Tom t" > Linden" <tom@kednos.com> writes:9 > :the older mice had a circular 7 pin connector; whereast7 > :the newer ones have a 6 pin pc compatible connector.t > ? >   Differences in wiring and connectors are (usually) easy to  = >   resolve.   Signals and protocols differ, which makes this-! >   effort rather more difficult.n > < > :Does anyone know if it is possible to associate the wires9 > :from one with the other?  The objective is to put some-: > :VAXs, DEC 3000/300 and DECstation 5000 on a KVM switch. > B >   Theoretically yes, but I've never seen the electronics needed.D >   Please see the FAQ for some background and pointers to low-levelC >   signaling details.  Best to banish these older boxes from your rD >   desktop, and use more recent systems and SET HOST or telnet (andB >   remote X Windows displays, if needed) to connect to the older  >   boxes as required. > 3 >  ---------------------------- #include <rtfaq.h>   > -----------------------------46 >       For additional, please see the OpenVMS FAQ --  > www.openvms.compaq.com    5 >  --------------------------- pure personal opinion a > ---------------------------t4 >    Hoff (Stephen) Hoffman   OpenVMS Engineering    > hoffman#xdelta.zko.dec.com >    ------------------------------  % Date: Tue, 16 Apr 2002 01:46:31 -0400f- From: JF Mezei <jfmezei.spamnot@videotron.ca>o' Subject: Re: Evolution of the DEC mouse , Message-ID: <3CBBBAB6.69F52569@videotron.ca>   Hoff Hoffman wrote:tB >   signaling details.  Best to banish these older boxes from yourD >   desktop, and use more recent systems and SET HOST or telnet (andA >   remote X Windows displays, if needed) to connect to the oldero >   boxes as required.  K That is HERESY coming from a VMS engineer. How dare you suggest that a goodr! old VMS system be banished ??????u  M Some of us have feelings you know, and the thought of banishing such a systemrL is tantamount to sending your puppy to the SPCA to be destroyed (or whatever# politically correct term they use).o   ------------------------------    Date: 15 Apr 2002 13:21:03 -0500- From: koehler@encompasserve.org (Bob Koehler) 4 Subject: Re: Excursion PCs as X-Terminals on OpenVMS3 Message-ID: <hgvmQjb6RJyB@eisner.encompasserve.org>   o In article <mTCu8.5158$Du7.1911080186@hestia.telenet-ops.be>, "Johan Schoofs" <jschoofs@compuserve.com> writes:i	 > Hi all,- > M > Can anyone explain how to get the standard CDE based login screen displayedlM > on a PC running Excursion (or any other X-Windows server software)? This is6K > really a nobrainer for Tru64 hosts: just enable XDCMP (sic?) in ExcursionyL > and you get a nice list of the Tru64 systems running X. You select one outL > of the list and you get the CDE login screen displayed on the PC. We wouldN > really like to see our OpenVMS systems doing something similar. I've checked@ > the Excursion doc but it is of no help concerning our problem. >   F    Depends on your IP stack.  Multinet supports XDCMP broadcast, which%    is what Excursion is watching for..  9    Alternatively you can connect, psecifying the command            mcr decw$session .    as the command to initiate on the VMS host.    e   ------------------------------  % Date: Mon, 15 Apr 2002 22:02:25 +0200v) From: Bart Zorn <B.Zorn@xs4all.nospam.nl> 4 Subject: Re: Excursion PCs as X-Terminals on OpenVMS/ Message-ID: <3CBB31D1.4070501@xs4all.nospam.nl>.   Johan Schoofs wrote:	 > Hi all,e > M > Can anyone explain how to get the standard CDE based login screen displayedeM > on a PC running Excursion (or any other X-Windows server software)? This is2K > really a nobrainer for Tru64 hosts: just enable XDCMP (sic?) in Excursion%L > and you get a nice list of the Tru64 systems running X. You select one outL > of the list and you get the CDE login screen displayed on the PC. We wouldN > really like to see our OpenVMS systems doing something similar. I've checked@ > the Excursion doc but it is of no help concerning our problem.  I TCP/IP services for OpenVMS V5.1 do support XDM. You just have to enable  @ the service. However, when I enable XDM in eXcursion, I get the D selection box and after I select my host, I get a login box. When I E enter a username/password (of which I am pretty sure they are OK), I p& just get the host selection box again.  L Maybe I find the time to explore this a little bit further when I have time.  	 Bart Zornn   ------------------------------  % Date: Mon, 15 Apr 2002 22:52:01 -0400-- From: JF Mezei <jfmezei.spamnot@videotron.ca>m! Subject: Re: Exporting SYSUAF.LISf, Message-ID: <3CBB91D0.62D057F2@videotron.ca>   Hoff Hoffman wrote:$1 >   Why not audit the system directly on OpenVMS?<  J When I worked at a bank, I was the only VMS resource inside the bank. WhenL they wanted to audit my systems, I offered to walk them through it, but theyJ prefered I generate printed listings they could read so that they woudln't have to learn VMS security.h  E Told them that any user with more then netmbx,tmpmbx was considered anL privileged user and they were happy with that. They had no way to relate theS VMS concepts with RACF security on their IBM mainframes on which they were trained.r   ------------------------------  # Date: Tue, 16 Apr 2002 02:39:38 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)! Subject: Re: Exporting SYSUAF.LISl9 Message-ID: <K9Mu8.41$Nc.1330267@cacnews.cac.cpqcorp.net>e  f In article <9bbdf680.0204090326.11a2aac9@posting.google.com>, andysands@yahoo.com (Andy Sands) writes:B :I asked a client to use the following script to download the userD :profiles from their VMS box to allow me to perform some audit work.    /   Why not audit the system directly on OpenVMS?e     :Prompt		Command :$ 		SET DEFAULT SYS$SYSTEMe :$ 		RUN AUTHORIZE :UAF> 		LIST [*,*]/FULLr :UAF> 		EXIT : E :The SYSUAF.LIS file that I have been given has no structure in it athG :all (no carriage returns specified etc) so cannot be read into a fixedmD :width interpreter like Monarch or a delimited file interpreter like :ACL/Excel.a    B   You seem to believe folks here are familiar with these packages,C   or with whatever operating system platform you are using to read  D   the sequential file.  (This is not necessarily a safe assumption.)B   I'd encourage dumping the first block of the file, and comparingF   the version you have with a dump of the version on the OpenVMS host.    F :I have received VMS downloads of these files in the past with clearlyF :structured records in them.  Does anyone know what has gone wrong andF :if so what is the correct command syntax for obtaining the sysuaf.lis :files in a structured form?  H   On OpenVMS Alpha V7.3, SYSUAF.LIS is a sequential variable file, with F   carriage return carriage control.  I have no idea if your particularH   foreign platform (Windows?) can read this or other OpenVMS sequential A   file formats, nor do I know if the file transfer was performed iC   appropriately.  I do know that transfering sequential files isn'taD   as easy as it might first be assumed, and that the native C streamC   linefeed file format tends to be the easiest to transfer.  I also C   know that some of the Windows tools tend to have different ideas r=   about sequential record formats and line wrapings and such.y    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Mon, 15 Apr 2002 19:00:46 GMTs* From: "Bill Todd" <billtodd@metrocast.net>) Subject: Re: Fearless VMS Prognostications@ Message-ID: <yrFu8.62691$%8.5269280@bin2.nnrp.aus1.giganews.com>  @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message& news:a9ere3$flfv$2@lead.zk3.dec.com...  > Bill Todd wrote in message ... > >U > >Two glaring examples aret > >nI > >1.  The numbers that refuted Compaq's lie that Alphas cost too much to3B > >develop:  Winkler's and Marcello's cost figures, the associated informationgG > >in the 1999 and 2000 annual reports, VMS's profit figures, VMS's andn	 > Tru64's J > >revenue figures.  It's not as if you provided any numbers yourself thatI > >called those numbers into question:  you just chose to dismiss them asu > >'nay-saying' and 'whining'. > >y >tH > Not to pick a nit.  Having the VMS profit and revenue numbers, and the UNIXE > revenue numbers tells you almost nothing about the profitability ofi > AlphaServers, or Tru64 UNIX.  J Not to point out the obvious, but when you also have the Alpha developmentF cost figures (which I did include above) there's no nit to pick:  theyG clearly establish at least a lower bound on overall Alpha profitabilityiK (unless you want to claim that AlphaLinux systems and/or Tru64 systems werea" being sold at a significant loss).   >oI > Let me know when you have the profit figures for the AlphaServer group,cG > which includes Tru64, OpenVMS, and Linux along with the associated HW I > platforms and chip development costs, excluding services profit - whicht canuK > be provided seperately.  I know that *I* have never officially seen them.   H Guess you haven't talked with or bothered to read the statements of yourJ upper management as much as I have (the former isn't very much, but enough* to get the figures I've quoted from them).   > F > >2.  The numbers that refuted Compaq's lie that Alphas couldn't have
 > retainedK > >a major performance lead over Itanic over time:  Compaq's 1999 Alpha vs. G > >Itanic comparison paper, the date and performance projections of thei AlphauL > >development team, the date and performance projections Intel (and Compaq)J > >made for Itanic, Paul DeMone's projections.  You did provide one number in= > >return in that statement I happened to find the other day:d > >k >/L > The quote "lie" unquote you want to prove is that over the next 5-10 yearsK > that the costs of developing a unique chip, and unique hardware platforms  toG > leverage the chips performance would have led to a significant markete share, > and high profitabilty.  L Reasonable profitability was already established as noted above (your denialK is your own personal problem), despite monumental lack of marketing effort.oH Significantly higher market share and profitability was certainly likelyI given any significant marketing effort (if you want to argue against thath4 one, you'll have to take on most of this newsgroup).  /   You have past performance to show that it hadmK > captured only a small market share, and no insight into the profitabilityr > prospects.  4 I don't appear to be the party here with no insight.   >9J > Versus the cost/benefit to consolidate development on a single 3rd partyI > (widely available) architecture, use common system platforms across allo O/SoG > domains.  The chip maker has a proven ability to push the process andtL > increase clock frequency, and is "likely" to catch our unique architectureL > at some future time in the decade - simply because they will sell millionsI > of them, which will give them a lower the cost per chip (with no NRE onm ouriI > part) and they control some of the fastest chip processes, and can spin  newtF > chips generations in less than 1 year cycles once they get the basicJ > architecture down (just look at how they churn pentium generations out).  D As has been pointed out by many people over the past 9+ months, yourL description above bears very little resemblance to the processor cHomPaq hasK pinned its hopes on (though it might describe the IA32 platform fairly wellwK in some respects).  That makes it pretty funny that you're the one accusinge. me of having no reasonable basis for my views.   - bill   ------------------------------  % Date: Mon, 15 Apr 2002 15:49:32 -0700c+ From: "Mike Scott" <mscott_NOSPAM@axys.com> ! Subject: full file backup to disk. Message-ID: <3cbb59b9@nubby2.>  J I don't have a tape device, and wish to make a full (file) backup to disk.F I only have the one disk in my system.  I tried the following command:8 $ BACKUP/IGNORE=INTER/IGNORE=LABEL DKA0:[000000...]*.*;* DKA0:[000000]ALL_FILES.BCK2 and got a ton of errors about FILE ALREADY EXISTS.. Am I missing some bit of syntax?  -with thanks  H p.s. my motivation is to have a complete system archive so I can restore/ individual files back to a known state (day-0).q   ------------------------------  % Date: Mon, 15 Apr 2002 19:20:27 -0400<2 From: norm lastovica <norman.lastovica@oracle.com>% Subject: Re: full file backup to diska* Message-ID: <3CBB603B.F3E3EAD4@oracle.com>  & Use /SAVE_SET on the destination file:   < $ BACK/IGN=INTER DKA0:[*...] DKA0:[000000]ALL_FILES.BCK/SAVE   Mike Scott wrote:  > L > I don't have a tape device, and wish to make a full (file) backup to disk.H > I only have the one disk in my system.  I tried the following command:: > $ BACKUP/IGNORE=INTER/IGNORE=LABEL DKA0:[000000...]*.*;* > DKA0:[000000]ALL_FILES.BCK4 > and got a ton of errors about FILE ALREADY EXISTS.0 > Am I missing some bit of syntax?  -with thanks > J > p.s. my motivation is to have a complete system archive so I can restore1 > individual files back to a known state (day-0).r   --  > norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  % Date: Tue, 16 Apr 2002 08:15:58 +1200.& From: A Bonaveidogo <Asena@fsc.com.fj> Subject: FW: clusterG Message-ID: <BFBEDDF2CFEDD411917400508BF3A6FFCBCB05@exchsvr.FSC.COM.FJ>n  0 This is how both disks are mounted on both nodes  = MOUNT/CLUSTER/NOASSIST DSA0:/SHADOW=($1$DKA0:,$2$DKA0) WBDATAy   -----Original Message-----. From:	Mark E. Levy [mailto:mlevy70@attbi.com] & Sent:	Tuesday, April 16, 2002 12:15 AM To:	Info-VAX@Mvb.Saic.Como Subject:	Re: cluster  H I would expect that behaviour. I assume that this is an NI cluster. WhenG Node A boots, Node B's drives are not yet available, so you'll get thateJ error. When Node B boots, Node A's drives will be available, and it shouldH re-form the shadow set. Why it's not is difficult to determine since youE didn't say exactly how and when you're mounting the drives. It's also 9 possible that some sysgen parameters need to be adjusted. L A *MUCH* better way to resolve this is to share the SCSI bus between the twoL CPUs, such that both drives are available to both CPUs regardless of whetherL or not the other CPU is up. Basically, you'll need to set the SCSI ID of theG controller in one machine to 6. This can get complicated, and since youoI didn't say what kind of CPUs these are, so I can't be more specific as to  how to do this.d	 Mark Levy  SMAd  3 "A Bonaveidogo" <Asena@fsc.com.fj> wrote in message-A news:BFBEDDF2CFEDD411917400508BF3A6FFCBCB01@exchsvr.FSC.COM.FJ...- >-L > I have two node (A & B)  forming a cluster and  both connected on shadowedK > environment.  On node A I have $1$dka0, on node B its $2$dka0.  these two-G > disks are shadowed. When I boot each node, it couldn't recognised thez member8 > of the shadow set.  Its says no such device available.I > On Node A... this error message appeared "$2$dka0 failed as a member ofo then
 > shadow set"s> > On Node B ... "$1$dka0 failed as a member of the shadow set" >nK > When I do a show cluster on each node....it shows only the  current node.r >t? > Can someone advise how to configure these two nodes properly.c >e > Asenat >n >  >k   ------------------------------  % Date: Tue, 16 Apr 2002 08:20:46 +1200u& From: A Bonaveidogo <Asena@fsc.com.fj> Subject: FW: clusterG Message-ID: <BFBEDDF2CFEDD411917400508BF3A6FFCBCB06@exchsvr.FSC.COM.FJ>e   Vadya,  L The disk to be shadowed is the data disk not systems disk.  I have two disksL on both node.  Dka100 is the system disk, dka0 is the data disk.  Dka0 disks  on both nodes are to be shadowed   thanks   -----Original Message-----/ From:	vadya@ibjus.com [mailto:vadya@ibjus.com] B% Sent:	Tuesday, April 16, 2002 3:55 AMh To:	Info-VAX@Mvb.Saic.Com- Subject:	Re: cluster  1 A Bonaveidogo <Asena@fsc.com.fj> wrote in messageeC news:<BFBEDDF2CFEDD411917400508BF3A6FFCBCB01@exchsvr.FSC.COM.FJ>...bL > I have two node (A & B)  forming a cluster and  both connected on shadowedK > environment.  On node A I have $1$dka0, on node B its $2$dka0.  these twooG > disks are shadowed. When I boot each node, it couldn't recognised the  member8 > of the shadow set.  Its says no such device available.I > On Node A... this error message appeared "$2$dka0 failed as a member ofS thes
 > shadow set"d> > On Node B ... "$1$dka0 failed as a member of the shadow set" > K > When I do a show cluster on each node....it shows only the  current node.t > A > Can someone advise how to configure these two nodes properly.  s >  > Asenan    A To Enable shadowing on system disk you need to set the parameter:t Shadow_sys_disk to 1$ a breif description is posted below:   SHADOW_SYS_DISKlL A SHADOW_SYS_DISK parameter value of 1 enables shadowing of the system disk.L A value of 0 disables shadowing of the system disk.  The default value is 0.B Also specify a system disk shadow set virtual unit number with theL SHADOW_SYS_UNIT system parameter, unless the desired system disk unit number is DSA0.I To enable minimerge on a system disk, add the value 4096 to your existing H SHADOW_SYS_DISK value. For example, if you have SHADOW_SYS_DISK set to aK value of 1, change it to 4097 to enable minimerge. Also, be sure to set theMH DUMPSTYLE parameter to dump off system disk, as described in the OpenVMS System Manager's Manual.   ------------------------------  % Date: Tue, 16 Apr 2002 08:35:47 +1200 & From: A Bonaveidogo <Asena@fsc.com.fj> Subject: FW: clusterG Message-ID: <BFBEDDF2CFEDD411917400508BF3A6FFCBCB07@exchsvr.FSC.COM.FJ>    Vidya,  > Do I have to execute cluster_config on both nodes?  ie executeI cluster_config on Node A and add node B  likewise on Node B to add node Ad Am I right?t   Sorry, I'm newbie to vms   thanks   -----Original Message-----/ From:	vadya@ibjus.com [mailto:vadya@ibjus.com] r% Sent:	Tuesday, April 16, 2002 3:47 AM  To:	Info-VAX@Mvb.Saic.Comu Subject:	Re: cluster  1 A Bonaveidogo <Asena@fsc.com.fj> wrote in messagegC news:<BFBEDDF2CFEDD411917400508BF3A6FFCBCB01@exchsvr.FSC.COM.FJ>...rL > I have two node (A & B)  forming a cluster and  both connected on shadowedK > environment.  On node A I have $1$dka0, on node B its $2$dka0.  these twocG > disks are shadowed. When I boot each node, it couldn't recognised thec member8 > of the shadow set.  Its says no such device available.I > On Node A... this error message appeared "$2$dka0 failed as a member ofl the 
 > shadow set"e> > On Node B ... "$1$dka0 failed as a member of the shadow set" > K > When I do a show cluster on each node....it shows only the  current node.  > A > Can someone advise how to configure these two nodes properly.  s >  > Asenao   Hello,  F If "show cluster" shows only one system, you do not have a VMS cluster configured on the nodes. You will have to run  5 "@sys$manager:CLUSTER_CONFIG.COM" to set up clusters.dE You will ned to know What kind of cluster interconnect are you using.n cheers Vinitn   ------------------------------   Date: 15 Apr 2002 23:54:05 GMT/ From: Thomas Dickey <dickey@saltmine.radix.net> & Subject: Re: GNU cc vs VMS CC listings* Message-ID: <a9fp6t$qdn$1@news1.Radix.Net>  , Douglas B Rupp <rupp@nospam_gnat.com> wrote:I > It's amazing the amount of misinformation that abounds on this subject.1  K it's typical of newsgroups (most responses aren't addressing the question).:  K > Speaking only wrt Gnu C for Alpha, since I know very little about the VAXa
 > version:  N > Gcc output that contains the expansion of the include files can be obtain byO > either passing the "-E" switch, or with the -save-temps switch.  However thiseH > isn't generally needed, since the traceback correlates to the original" > source, not the expanded source.  H VAX C tracebacks referenced line numbers - the expanded source, countingI include-files - in the listing (I haven't been doing either that or DEC CCJ recently, but seem to recall that DEC C does the same).  While generally aC waste of diskspace, that approach does have the advantage of simpleu coordination with a debugger.o  H otoh, the VAX/DEC C preprocessor is less functional than the normal UnixJ version (it has no feature for retaining comments from the original sourceG files, which can aid the developer in determining which chunk of sourcehK expanded into a particular area).  The listing file can show the equivalent.J information, but that is not as generally useful as the ability to compile5 the expanded source (with comments and #line's, etc).    So there're tradeoffs.  F Apollo also had a "Unix" whose compiler could generate listings - thatI was not uncommon for commercial compilers in the 80's outside of the AT&Tf< environment.  Of course we recall what happened to Apollo...   --  = Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>h http://dickey.his.come ftp://dickey.his.com   ------------------------------   Date: 15 Apr 2002 23:59:57 GMT/ From: Thomas Dickey <dickey@saltmine.radix.net>n& Subject: Re: GNU cc vs VMS CC listings* Message-ID: <a9fpht$qdn$2@news1.Radix.Net>   clark@sander.stsci.edu wrote:d  A > I guess my original post was based on problems I am having withMB > determining where #IFDEF macros are being evaluated. I am having? > problems determining what the <include files> are doing. More A > specifically I am getting a variable definition being redefinedoA > several times. On VMS the /list/show=include marks the lines inc@ > the <include files> that are being skipped over and those thatC > are being used. I don't see how something that basic can be seen i > to be a wave of the past.e  C > Just how does a modern programmer determine what the preprocessorx& > is doing with the #IFDEF statements?  D I would use the -E and -C options of the preprocessor (though the -CL option in very recent gcc's - 2.96 and up - results in uncompilable source).  J The -C option passes-through the comments from the original files.  The -EM option passes-through #line and #include lines and expands the macros, making N it possible to see the overall context of the expanded source.  (That feature,K btw, was borrowed from other compilers which have since omitted some of thec: information which made it useful, e.g., SunOS vs Solaris).   -- t= Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>f http://dickey.his.comn ftp://dickey.his.com   ------------------------------    Date: 15 Apr 2002 14:09:14 -0700* From: TaroSuzuk@excite.co.jp (Taro Suzuki)# Subject: Hobbist VMS Source Listings= Message-ID: <d0bc4ae0.0204151309.3fe32869@posting.google.com>   
 Dear Readers,w  E As a hobbist is it possible to buy the source listing for VMS cheapera than for commerical use?  . Is so where is a good place in the USA to try?   Thanks I   Taro   ------------------------------  % Date: Mon, 15 Apr 2002 18:25:21 +0100 T From: Andrew Harrison SUNUK Consultancy <andrew_nospam.remove_this.harrison@sun.com>O Subject: Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxi$ Message-ID: <3CBB0D01.70007@sun.com>   Doc.Cypher wrote:i     > K > The correction wasn't the important part, the question was. Of course, asiL > per usual, you didn't answer the question. I've now done a little researchM > into the google archive and it would appear that you have *never* backed upcM > an argument with a cite from deja (as it was) or google (as it now is). So,cL > perhaps you'll refrain from telling others to do what you don't seem to beL > able, or willing, to do yourself. If you debate without giving references, > why shouldn't others?- >     . I didn't answer the question because it wasn't! worth answering. Ask a good one !o     Regards- Andrew Harrison-   ------------------------------    Date: 15 Apr 2002 19:17:50 -00004 From: Doc.Cypher <Use-Author-Address-Header@[127.1]>O Subject: Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinux 5 Message-ID: <20020415191750.4914.qmail@gacracker.org>   6 On Mon, 15 Apr 2002, Andrew Harrison SUNUK Consultancy3 <andrew_nospam.remove_this.harrison@sun.com> wrote:p >Doc.Cypher wrote: >T >n >> sL >> The correction wasn't the important part, the question was. Of course, asM >> per usual, you didn't answer the question. I've now done a little researchdN >> into the google archive and it would appear that you have *never* backed upN >> an argument with a cite from deja (as it was) or google (as it now is). So,N >> perhaps you'll refrain from telling others to do what you don't seem to be M >> able, or willing, to do yourself. If you debate without giving references,u >> why shouldn't others? e  / >I didn't answer the question because it wasn'ti" >worth answering. Ask a good one !  G I thought it was a perfectly good question, perhaps if I rephrase it...d  I Why do you keep telling others to use the google archive to justify their E arguments when you won't do so yourself, and instead vaguely refer tobF google as justification for whatever argument you're currently putting across?r     Doc. -- a6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                              http://vmsbox.cjb.netn   ------------------------------    Date: 15 Apr 2002 17:43:02 -0700( From: bob@instantwhip.com (Bob Ceculski)O Subject: Re: HP's viewpoint on Linux, was: Re: Sun eating major helping ofLinuxs= Message-ID: <d7791aa1.0204151643.2d6a95b1@posting.google.com>    Andrew Harrison SUNUK Consultancy <andrew_nospam.remove_this.harrison@sun.com> wrote in message news:<3CBB0D01.70007@sun.com>... > Doc.Cypher wrote:2 >  >  > > M > > The correction wasn't the important part, the question was. Of course, asrN > > per usual, you didn't answer the question. I've now done a little researchO > > into the google archive and it would appear that you have *never* backed uptO > > an argument with a cite from deja (as it was) or google (as it now is). So, N > > perhaps you'll refrain from telling others to do what you don't seem to beN > > able, or willing, to do yourself. If you debate without giving references, > > why shouldn't others?v > >  >  > 0 > I didn't answer the question because it wasn't# > worth answering. Ask a good one !  >  > 	 > Regardsc > Andrew Harrisonu  B I asked you one ... are you going to put up at defcon10 a slowaris* box againset vms to see who is unhackable?   ------------------------------  # Date: Mon, 15 Apr 2002 20:58:04 GMTi* From: "Bill Todd" <billtodd@metrocast.net>6 Subject: Re: I was right!  Alpha will live in Itanium!A Message-ID: <w9Hu8.45024$3L2.3710945@bin7.nnrp.aus1.giganews.com>   : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3CBB15E6.3F7B6799@videotron.ca...* > Andrew Harrison SUNUK Consultancy wrote:9 > > I have repeatedly given examples of how the GS seriesr> > > are not competitive based on Compaqs own public, published$ > > benchmark results for the boxes. >o >nI > Andrew, did Sun ever fix the problem with its memory going bezerk underrE > certain circumstances (aka: all those ebay problems for instance) ?m  E Yes, IIRC they did - by disabling the cache in the affected series ofuI products, and by recently issuing a new series without the cache problem.m  
  If so, is; > it still fair to say that Sun has very bad memory today ?w  I It would be if they were criticizing others for the same type of problem,h$ but AFAICT that's not the case here.   >eE > In the same vein, has Digital fixed some of the initial performanceo problems > in the wildfires ?  H Not to my knowledge:  I haven't seen any indication that Wildfire memory' access latencies have improved, anyway.   7  If so, is it still fair to say that they don't performeI > properly today ? And when you consider the upcoming EV7 based wildfirest6 > (marvel), will the new wildfires really perform ????  C They won't be Wildfires, but Marvels.  And while they well may blowaI everyone's socks off, they aren't yet available nor do they have any real-7 bearing on the comments being made about the GS series.M   >cK > It is very easy to use older statistics to show a competitor is weak. ButD itJ > will come back to hurt you because when the customer checks and gets theI > updated statistics from the other vendor, he will see through to tacticl and ) > that gives you negative brownie points.L  H You appear to be confusing currently-available products with future onesL (Rob has this problem as well).  While I suspect Marvel will indeed turn outH to be everything that Wildfire was not, using Wildfire statistics is notH 'using older statatistics' yet:  they're the most up-to-date information
 available.   > J > You can hammer Digital/Compaq/HP all you want about VMS's costs and lack ofH > marketing and lack of visible future. That is plenty to steer clear ofJ > Digital/Compaq/HP/whatever. But on a performance basis, are you ready to state:L > that a bunch of separate Sun boxes, connected by wathever, will outperform aa+ > wildfire with same number of processors ?b  G Why do you want to compare a bunch of Sun boxes with a single Wildfire?wH IIRC Sun sells single boxes with more than twice the processors a single Wildfire can contain.S  H A single Sun E10K box (with its cache turned off) may well not match theG performance of a Wildfire with the same number of processors (though itoK might if the cache could be used, and some lesser Sun boxes that don't haveaK the same cache problem might too).  My impression is that a single Sun E15K9J box may well outperform a Wildfire with the same number of processors, but. it would be interesting to see actual numbers.   - bill   ------------------------------    Date: 15 Apr 2002 17:49:19 -0700( From: bob@instantwhip.com (Bob Ceculski)6 Subject: Re: I was right!  Alpha will live in Itanium!= Message-ID: <d7791aa1.0204151649.5e5bd58f@posting.google.com>-  a JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message news:<3CB733AD.F7E0935B@videotron.ca>...X > jlsue wrote:H > > What a laugh.  Andy-boy giving advice.  Like anyone in here respectsA > > this doofus and his elevated-opinion-of-himself-ramblings (asv& > > perfectly evidenced in this post). > N > Like anyone here respects those who feel compelled to insult Andrew anythingL > they can. We are smart enough to filter out the fluff from Andrew and takeJ > some of the tidbits that sometimes make sense. There is no real need forE > Digital employees to make a point of insulting Andrew all the time.s >   D we don't have to ... Andrew does a fine job making himself look like an idiot all by himself ...    ------------------------------  # Date: Mon, 15 Apr 2002 17:50:11 GMT / From: "Don Chiasson" <don_chiasson@notmail.com>h  Subject: Re: IA64 is not the VAXE Message-ID: <npEu8.3684$SK1.984@news01.bloor.is.net.cable.rogers.com>i  . > Christopher C. Stacy wrote re LISP machines:
 [snip....] >A5 > > (By the way, there were five completely different1< > > architectures, implemented in a variety of processes for= > > dozens of different machines over the years, by the way.)w > > 	 [snip...].  G I find this amazing. How could a company put the resources into so many H architectures.  The method of implementation is a detail and will changeF with the technology, but IBM is still using, with extensions, the sameH 360 architecture from about 38 years ago. Similarly, your 2.4GHz PentiumD 4 starts up as an 8086 (or something of that ilk). DEC had the PDP-68 (and follow ons) for about 20 years. Symbolics must have< spent - wasted - a huge amount of energy in chasing multiple architectures.   Donp e-mail: it's not not, it's hot.    ------------------------------    Date: 15 Apr 2002 08:55:18 -0700 From: Greg Finn <finn@isi.edu>  Subject: Re: IA64 is not the VAX( Message-ID: <6w1ydht0rd.fsf@cnn.isi.edu>  / cstacy@grant.org (Christopher C. Stacy) writes:i   [snip]  H You have your opinion.  A few years after their introduction researchersI spent their contract money elsewhere.  That is a good deal of independente. decision making coming to the same conclusion.   ------------------------------    Date: 15 Apr 2002 20:14:05 -0700) From: codygould@yahoo.com.au (Cody Gould)V! Subject: Initial Sequence Numberst= Message-ID: <e922ba8b.0204151914.382e126e@posting.google.com><  	 Care All,4  9 Is there any predictability to, or documentation of, how o9 OpenVMS calculates the ISN for TCP/IP packets?  Back whend9 this was a problem for other OS, was OpenVMS historicallyt vulnerable?h  
 93, 93/93,
 Cody Gould   ------------------------------    Date: 15 Apr 2002 21:36:59 +0100( From: joechip31@hotmail.com (Herb Asher)' Subject: irc chat channel #vms on efnets9 Message-ID: <3cbb38da.82684103@news1.uncensored-news.com>w  F Any fellow irc users out there, pop by #vms on efnet and say hello, we* are there most of the time (sad but true).  B If you don't know what irc chat is, then mirc for windows does the< trick or you can install a client on your vms box (ircii for
 instance).   See you there maybe...   Herb (Jiffy on irc)  joechip31@hotmail.com         F ______________________________________________________________________R Posted Via Uncensored-News.Com - Still Only $9.95 - http://www.uncensored-news.comK       <><><><><><><>   The Worlds Uncensored News Source   <><><><><><><><>a      ------------------------------  % Date: Mon, 15 Apr 2002 22:18:10 -0400t) From: "Neil Rieck" <n.rieck@sympatico.ca>. Subject: Itanium Odditiess: Message-ID: <VRLu8.6255$132.1002178@news20.bellglobal.com>   Folks,  E I just stumbled on this great source of 64-bit technical information:i? http://www.extremetech.com/article/0,3396,s=1005&a=22477,00.aspLE One thing that stood was the fact that Itanium (a.k.a. IA-64; Merced;mJ McKinley) has no ADD instruction. According to the above link, in order toI add, you must use a MAC (Multiply and Accumulate) instruction that adds auI value after multiplying by one. Maybe one of the people currently portingJ: OpenVMS (or Tru64) to IA-64 would like to comment on this.  
 Neil Rieck Kitchener/Waterloo/Cambridge,a Ontario, Canada.! http://www3.sympatico.ca/n.rieck/I   ------------------------------    Date: 16 Apr 2002 01:41:01 +0800, From: Paul Repacholi <prep@prep.synonet.com> Subject: Re: Itanium troublesl- Message-ID: <87elhg3ln6.fsf@prep.synonet.com>   - "Main, Kerry" <Kerry.Main@Compaq.com> writes:c  @ > Re: Disaster backup sites overflowing with other Customers ..   eB > You are correct and there is some additional items to consider -  dD > 1. Disaster tolerant "rental" sites typically cost in the range ofB > $200k USD per year (every year) - just to have the equipment andE > space on standby. Now, if implemented, the daily rates are often in.D > excess of $10K USD per day. Think about what the final costs would= > be to go out and build / find a new site - optimistically 3eE > months. Now add the complexity associated with many other Customers 2 > also looking for similar spare datacenter space.  F The lession is to add to the queries and the contract limits on othersF from your 'primary' site. so you are not all piled up at the door when the bell tolls.y  D > 2. Based on issues discovered in point 1.), some Customers are nowD > questioning whether 2 sites is enough i.e. perhaps 3 site clustersE > (active-active-passive?) are the right approach for the future? BigsD > issue is that if the primary site goes away permanently, then thatB > business must run everything on a single site and that is a huge > risk to their business.s  eF > Question for future discussion - does the disaster recovery site you; > are renting from have a backup site? Some do, some don't.   tD > 3. While obviously not for everyone, keep in mind that the cost ofC > multi-site solutions for those that are looking at them typically A > have downtimes that mean even a few hours would pay for all the F > multi-site HW and software. That does not even consider that VendorsC > credibility as an online financial provider. Picture becomes evenHF > more critical when one considers the ever extending times for online > financial trading.   C > 4. While bandwidth and latency are always going to be issues with-F > multi-site clusters, keep in mind that for large centers, Cisco (andC > others in that space) already offer 10Gb WAN solutions today with  > 40Gb demonstrated.  B All true. I have often wondered why more people do not use 'mutual backup' arrangments.  @ I guess there will be another round of lessions soon from Tokyo.   -- n< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.c@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------   Date: 15 Apr 2002 19:43:35 GMT From: phn@icke-reklam.ipsec.nu Subject: Re: Itanium troublesU) Message-ID: <a9fah7$7fg$4@nyheter.crt.se>   4 In comp.os.vms Ketil Malde <ketil+@ii.uib.no> wrote:9 > "Tarjei T. Jensen" <tarjei.jensen@kvaerner.com> writes:l  O >> I have just installed FreeBSD 4.5 and reading the manual pages and web pageseJ >> it becomes quite clear that these people simply do not get it. They areL >> unable to think product and think like a user would (user = non fanaticalO >> devotee). I wonder if I should use the word inept or incompetent to describea >> the effort involved.e  D BSD is intended for experts ( or at least knowlegebal users) If that+ does not suit, then install something else.n  = Personally it was a relief to start using manpages instead ofe VMS help. YMMW     > How about "disinterested"?  G > I mean, why do you expect people to think "product" -- that is, solve H > problems for end users who are unwilling to solve them for themselves,* > when there's no motivation for doing so?  + The "product" is an OS. Not a word-mungler.o  G > (I'm really amazed at the level of support available through internettD > channels from people who really have very little to gain from it.)   > -kzm > -- DJ > If I haven't seen further, it is by standing in the footprints of giants   --   Peter Hkanson          K         IPSec  Sverige      (At the Riverside of Gothenburg, home of Volvo)sJ            Sorry about my e-mail address, but i'm trying to keep spam out.& 	   Remove "icke-reklam" and it works.   ------------------------------  % Date: Tue, 16 Apr 2002 07:46:37 +0930s/ From: Mark Daniel <Mark.Daniel@wasd.vsm.com.au>h* Subject: Re: Java with the WASD webserver.. Message-ID: <3CBB5145.2010605@wasd.vsm.com.au>   Doc.Cypher wrote: 
 8< snip 8<F > Yes, as you'd see from my other message I have installed the updatedI > version and it all works now. I also suspect the web server is a little5 > faster in serving up pages.   J I'm always tweaking! (as those who stumble across the bugs will attest ;^)  E > I saw that you're working on v8, Ken Kalish mailed back to me aftereK > querying you about Tomcat that you were looking into it. Will support foraL > Tomcat be included in the final v8 release of WASD? If so, I'd be happy to > beta it for that function.  I Ahh, Ken!  Yes, he was interested in Tomcat capabilities.  I had spent a wH few hours in February reverse-engineering the AJP v1.(2|?) protocol and G building a basic WASD RTE agent for it.  Then got bogged down in other fL things.  The RTE works but is only at a pre-beta stage.  I'm just trying to J get back to it this week or next (looks more like next at the moment).  I L provided Ken with an archive of the RTE not realizing he wasn't running his I own setup.  It's not at an installable point yet so I don't think you be u able to accomodate him.c  G I am endeavouring to decouple significant scripting environments (e.g. rI Perl, PHP, JSP, Python, etc.) from the mainstream WASD distributions, so iG the release of the WASD Tomcat interface will independent (and I doubt d6 coincident) with that.  I'll put you on the beta list.   ------------------------------    Date: 15 Apr 2002 13:12:26 -0500- From: koehler@encompasserve.org (Bob Koehler) G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)a3 Message-ID: <gPhDHwzT6KND@eisner.encompasserve.org>a  ` In article <o2rlbu49scjepunhsf5cvbvdeg69am4rqp@4ax.com>, Alan Greig <a.greig@virgin.net> writes: > G > Learning not to head down the wrong side of the road at intersections_G > is my problem. Why can't you Europeans and Yanks drive on the correctcA > lefh-hand side of the road the way nature obviously intended :)s >   H    Same reason we don't make such a fuss over tea.  Only folks living on8    an island nation seem to have either of those habits.   ------------------------------  # Date: Mon, 15 Apr 2002 19:15:03 GMTo* From: "Bill Todd" <billtodd@metrocast.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles),@ Message-ID: <XEFu8.51153$K5.4809575@bin5.nnrp.aus1.giganews.com>  3 "Jan-Erik Sderholm" <aaa@aaa.com> wrote in messagee! news:3CBADDE4.7DB7724C@aaa.com...hD > Hm, note that automatic transmissions are more or less a "feature"C > for the leasy US'es :-) In Europe, manual is still the way to go.t@ > Imagine a Lotus or a Ferrari with an automatic transmission... >dG > Learning to manage a clutch and manual transmission, must be the easy  > part of driving.  J It's easy only in retrospect:  when someone is just starting out, it's oneK of the highest initial hurdles to overcome and a significant distraction in 9 traffic when one should be concentrating on other things.-  7  Learning to behave sensible in traffic has always been.# > and will always be the hard part.   G True, but that's not something vehicle design can help much with, while K avoiding the distraction of having to learn completely new, somewhat subtleoK (how many people do you know who can shift a manual transmission without it<I being obvious to a passenger who has their eyes closed?), and now largelyr! useless physical coordination is.h   >sE > Anyway, we'll have to drive less in the future to save the earth....  K Agreed, unless clean fusion energy generation plus a way to transport it inaJ a vehicle appear.  I was an enthusiastic rallyist (all over North America)H in the '60s and '70s, but would have significant reservations about such activity today.    - bill   ------------------------------   Date: 15 Apr 2002 19:19:28 GMT  From: hack@watson.ibm.com (hack)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)c+ Message-ID: <a9f940$c2i$1@news.btv.ibm.com>   0 In article <a9f0qb$7n0$1@pegasus.csx.cam.ac.uk>,) Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:t >fN >In article <a9evit$hn4$1@news.btv.ibm.com>,hack@watson.ibm.com (hack) writes:4 >|> In article <a9e1ka$6dh$1@pegasus.csx.cam.ac.uk>,- >|> Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:uF >|> ...  It prevents you taking an accurate record.  So that, when youG >|> >report a bug, the support people can always say that you must havepH >|> >made a mistake.  This saves developers time because they don't have >|> >to fix most bugs. >|> ? >|> Not necessarily.  If the GUI is indeed just a front end ...   , >Memo to self:  Never post satire to Usenet.  G Oh, the satire was delicious!  But GUIs that provide no means to keep a I record of what they show and what I enter are a very sore point for me...    Michel.c   ------------------------------   Date: 15 Apr 2002 19:42:34 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)o0 Message-ID: <a9fafa$h34$1@pegasus.csx.cam.ac.uk>  M In article <a9f940$c2i$1@news.btv.ibm.com>, hack <hack@watson.ibm.com> wrote:i1 >In article <a9f0qb$7n0$1@pegasus.csx.cam.ac.uk>,F* >Nick Maclaren <nmm1@cus.cam.ac.uk> wrote: >>O >>In article <a9evit$hn4$1@news.btv.ibm.com>,hack@watson.ibm.com (hack) writes:w5 >>|> In article <a9e1ka$6dh$1@pegasus.csx.cam.ac.uk>,i. >>|> Nick Maclaren <nmm1@cus.cam.ac.uk> wrote:G >>|> ...  It prevents you taking an accurate record.  So that, when you'H >>|> >report a bug, the support people can always say that you must haveI >>|> >made a mistake.  This saves developers time because they don't havey >>|> >to fix most bugs.M >>|>  @ >>|> Not necessarily.  If the GUI is indeed just a front end ... > - >>Memo to self:  Never post satire to Usenet.n >eH >Oh, the satire was delicious!  But GUIs that provide no means to keep aJ >record of what they show and what I enter are a very sore point for me...  ? I agree.  I started to fight that one round about the time that A addressable VDUs were replacing teletypes in UK academia, but gotnA nowhere.  There were some SEAS or SHARE Europe resolutions on thec? matter during the CUA period, if I recall, though they may havet& been less formal feedback.  No joy :-(     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679D   ------------------------------  # Date: Mon, 15 Apr 2002 19:39:49 GMT * From: "Bill Todd" <billtodd@metrocast.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)sA Message-ID: <90Gu8.38173$%l3.3808097@bin8.nnrp.aus1.giganews.com>o  5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message * news:a9evam$69k$1@pegasus.csx.cam.ac.uk... > . > In article <3CBAFDB1.B041C9C8@videotron.ca>,1 > JF Mezei <jfmezei.spamnot@videotron.ca> writes:n > |> Bill Todd wrote: C > |> > Cars today are mostly point and go.  *Driving* skills remainh important, butG > |> > required knowledge of the vehicle has been dramatically reduced.r > |> > |> Are you kidding ? > |>I > |> One has to learn how to use that fancy electronic car stereo that isc often asL > |> complex to program as a VCR. One has to learn how to use that fancy GPS moving, > |> map display, the trip computer etc etc. >wC > One doesn't have to, provided that it is shipped with them turned- > off.  K Indeed.  And while it's possible that some high-end car stereos are hard tocD use, I've certainly never had to resort to a manual to use mine (andG appreciate some of its ease-of-use features such as its ability to scaneK upwards or downwards for stations strong enough to provide good reception).c  ;   However, your point is justified - recently, I have had acE > hire care where I had to drive into a garage to ask how to get intoiA > reverse gear, because it was rented to me without a copy of the.F > instruction manual!  The assistant couldn't work out how, either :-)  J Many vehicles solve this problem with a diagram embossed on the shift knob> (an approach which they generalize to most of their controls).   >i6 > That was in Texas, which may or may not be relevant.  ; That depends:  did your vehicle neigh and like sugar cubes?A   - bill   ------------------------------  # Date: Mon, 15 Apr 2002 19:41:35 GMT-* From: "Bill Todd" <billtodd@metrocast.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles) A Message-ID: <P1Gu8.38195$%l3.3809728@bin8.nnrp.aus1.giganews.com>e  : "Johan Schoofs" <jschoofs@compuserve.com> wrote in message7 news:GjDu8.5179$mc3.1926260937@hestia.telenet-ops.be..., >e5 > "Jan-Erik Sderholm" <aaa@aaa.com> wrote in messagei# > news:3CBADDE4.7DB7724C@aaa.com...fF > > Hm, note that automatic transmissions are more or less a "feature"E > > for the leasy US'es :-) In Europe, manual is still the way to go.cB > > Imagine a Lotus or a Ferrari with an automatic transmission... > >lI > > Learning to manage a clutch and manual transmission, must be the easyoK > > part of driving. Learning to behave sensible in traffic has always beenz% > > and will always be the hard part.m > >gG > > Anyway, we'll have to drive less in the future to save the earth...m > >s > > Jan-Erik Sderholm.d > >h > K > Be careful with the examples, Jan-Erik: the Ferrari 456 GTA comes with ant> > automatic gearbox and most other Ferrari's can be had with a semi-automaticK > gearboxes in with the clutch is automatically operated. Shifting gears issI > done using paddles on the steering wheel. The same system can be had ona therL > BMW M3 and a few other cars such as the little Toyota Roadster. All Jaguar; > XK roadsters come with an automatic transmission as well!s  J That reminds me that I think I recall some kind of clutchless shift option( from Porsche appearing recently as well.   - bill   ------------------------------   Date: 15 Apr 2002 20:39:25 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)e0 Message-ID: <a9fdpt$jlg$1@pegasus.csx.cam.ac.uk>  @ In article <5CGu8.51253$K5.4873808@bin5.nnrp.aus1.giganews.com>,) Bill Todd <billtodd@metrocast.net> wrote:s >iJ >The question is whether the alternative help interfaces you favor do suchH >things any better.  If they don't, then that's a wash and we move on to
 >other areas.c  C Oh, they did and they do.  But, given your abuse, I fail to see why1 I should explain further.n  D >> |> > >The GUI does it in a more aestetically pleasing manner than
 >character
 >> |> cellM >> |> > >menu system. GUI and character cell menus save some time because youo >just I >> |> > >tab/point to the field and type in the value instead of havin to 	 >type ther >> |> > >field name and value. >> |> >i0 >> |> > Evidence is that this doesn't save time. >> |>_H >> |> Please cite your authoritative source for this (it's something you >really L >> |> need to do far more often).  Be sure to include any effects related toI >> |> having full context for data entry in front of you in a menu screen- >(where M >> |> individual errors can be easily seen and corrected) vs. (at best) beinguK >> |> prompted one item at a time with command input:  it's easy to suspecta >that M >> |> when the data-entry person is obtaining the information via 'phone thisnK >> |> could make a noticeable difference.  I suppose it's possible that then >pastsG >> |> several decades of use of form-style screens have been completelyA >misguided,n? >> |> but I'm more inclined to suspect that your perception is.A >>C >> N.M. Maclaren, pers. comm.  Based on actual measurements of selfyE >> and other users, timed with a watch.  Data lost due to the passage6# >> of time.  So now you have it :-)t >.M >That's what I suspected.  Leaving aside the question of whether personal andyC >anecdotal experience merits being termed 'evidence' at all, baldlyrK >presenting it without qualification as 'evidence' sufficient to settle theRG >issue is at best extremely sloppy thinking and at worst intellectuallyM >dishonest.K  ? As I said, I took actual measurements.  I have seen very littlel; evidence that other people have done so, though I am prettyr( certain that some people must have done.  C >> Now, please produce your source OF AT LEAST THE SAME QUALITY ANDfE >> COMMERCIAL INDEPENDENCE that demonstrates that GUIs save time over ? >> going the same operations using a well-designed command linei >> interpreter.1 >UJ >Xerox PARC's work in the '70s.  Apple's work in the '80s (as I mentioned,H >they performed wide studies on test groups).  Your own biases may carryJ >equal weight with you but carry nothing like equal weight with me (nor, I1 >would suspect, with any other neutral assessor).s  D You may not be aware that Xerox PARC and the University of Cambridge> had pretty close links.  No, that was NOT one of their claims.A The nearest that they came to that in any document that I saw wass? a claim that it could be quicker to do a task that the user did ? not do regularly with a GUI than a command language.  But I sawi little evidence even of that.a  > You seem to be unaware that Apple had a commercial interest in@ their claims, and it is generally agreed that considerable doubt? needs to be placed on claims when that is the case.  Certainly,cA one claim of that nature which I did see was clearly bogus, as iteB assumed typing speeds that were low even by casual user standards.  D As I said, I took actual measurements on real people in as unbiassedB a fashion as I could manage (and I am a statistician).  You are at8 liberty to think me a liar, but you are wrong if you do.  = You have so far produced NO evidence of your claims, except apA further claim about what Xerox PARC and Apple claimed - for whicht you have provided no evidence.     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 3346790   ------------------------------  # Date: Mon, 15 Apr 2002 20:20:18 GMTt* From: "Bill Todd" <billtodd@metrocast.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)n@ Message-ID: <5CGu8.51253$K5.4873808@bin5.nnrp.aus1.giganews.com>  5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message * news:a9enpd$s4t$1@pegasus.csx.cam.ac.uk... > B > In article <wMwu8.58908$%8.4814023@bin2.nnrp.aus1.giganews.com>,. > "Bill Todd" <billtodd@metrocast.net> writes:: > |> "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message/ > |> news:a9e1ka$6dh$1@pegasus.csx.cam.ac.uk...G > |> > >E > |> > >A GUI system is not much different than a character cell menuaD > |> > > system such as ALL-I-1: it removes the need to remember theB > |> > > commands and what arguments are needed and shows you what > |> > > your options are. > |> >I > |> > In theory.  In practice, you are faced with something every bit as & > |> > inscrutable and no help system! > |>L > |> Windows certainly has comprehensive on-line help available, both in theK > |> context of the current activity and indexed (and cross-indexed) to aid  inJ > |> finding out other operations that might be pertinent, so I'm not sure what  > |> you're trying to say above. >iA > I assume that you mean Microsoft Windows, as X Windows does nott > have such a system!e >oB > Have you ever used a good help system?  The VMS one is adequate,C > though not good, but there were a fair number of good ones arounde > academia in the 1970s.  I I consider Windows help to qualify as 'good', in that it usually providesgK whatever information I need.  I was familiar with VMS's a 15 years ago, andaE it wasn't nearly as generally useful (though if you knew the specificpH command you needed it was good at enumerating options for it); one could% hope that it is better today, though.g  K A help system is one of the few things I'd expect academia might be able toSC implement well, but I thought we were talking actual products here. G Nonetheless, a brief description of how the systems you favor organizedeC their help might be more convincing than your assertions have been.    > A > The point is that the simplistic help systems as implemented ons@ > the better GUIs will tell you what buttons etc. do and provideB > similar information; they may even tell you which actions relate@ > to that facility.  Of course, they interactive abominably withA > active icons (ones where mere selection causes invokation), butb > that is another issue. >r< > What they do NOT help with is in answering questions like: >bA >     I can't find out how to do XXX - is it possible and, if so,n >     how do I do it?b > A >     Every time I do XXX, I get the message "Get knotted" - whato+ >     does it mean and how do I resolve it?i  I The question is whether the alternative help interfaces you favor do suchnG things any better.  If they don't, then that's a wash and we move on tod other areas.   >r- > |>   A good test of whether a GUI interface|J > |> > is badly designed is how much paper you need to use it - needing toH > |> > read a book on it is very bad, but so is having to copy text onto+ > |> > paper to type it into another field.o > |>I > |> The amount of paper documentation available for Windows is no more a  goodC > |> measure of the comparative need for it than the example of thes
 voluminousI > |> Saturn owner's manual recently cited.  In this case at least in partr becausecF > |> a GUI interface is significantly richer in both presenting useful	 on-screenmG > |> context and allowing flexible operations on it than a command linei	 interfacelL > |> is:  consider the ability to view a list of files and select individual  > |> items from it, for example. >-C > Well, I was around at the time when GUIs were first being foistedrC > on an unsuspecting population as the solution to making computerseB > usable by mere mortals.  They were claimed to be simpler to use,@ > as shown by the fact that you didn't need manuals to use them.  L And that claim was correct.  I watched my daughter (perhaps 3 or 4 years oldJ at the time) start using a Mac after being shown how to move and click the* mouse:  try that with a command interface.  J As I said elsewhere, Apple performed a fair amount of formal human testingI when developing its GUI interface.  Produce some equivalent research thatdK refutes the answers they came up with (rather than your own opinion or that-D of others brought up on non-GUI interfaces) if you want to get taken
 seriously.   >uC > I can do the operation you describe perfectly well on the command,  > line - what is the difficulty?  C The difficulties are  1) in learning and remembering the applicablewF commands,  2) in remembering the content of the file list after you'veF started issuing them, and  3) in typing a command for each file (don'tK assume that the file names bear any obvious similarities).  I could 'do theaL operation I describe perfectly well' without an interface at all, by openingH up the disk, measuring its magnetic domains one by one, and interpreting3 them:  that doesn't make it the preferred approach.t  #   I can also do operations like thet > following: >sB >     For all files in a directory that have a name ending in .datE >     which have a corresponding name in another directory but endings >     in .prog, run the commandh@ > execute weeble.wombat -prog <dir>/<name>.prog -data <name>.dat# >     otherwise execute the commands( > execute weeble.numbat -data <name>.dat  5 Scripting (or near-scripting using command piping andsC regular-expression-oriented utilities) is certainly useful for suchwD operations.  It also has little to do with the operations most usersH (remembering that most users are near-neophytes sitting at PCs) perform.   >eC > |> > >The GUI does it in a more aestetically pleasing manner thant	 character 	 > |> celleL > |> > >menu system. GUI and character cell menus save some time because you justH > |> > >tab/point to the field and type in the value instead of havin to type the > |> > >field name and value.- > |> >/ > |> > Evidence is that this doesn't save time.o > |>G > |> Please cite your authoritative source for this (it's something you  reallyK > |> need to do far more often).  Be sure to include any effects related topH > |> having full context for data entry in front of you in a menu screen (whereL > |> individual errors can be easily seen and corrected) vs. (at best) beingJ > |> prompted one item at a time with command input:  it's easy to suspect thatL > |> when the data-entry person is obtaining the information via 'phone thisJ > |> could make a noticeable difference.  I suppose it's possible that the pastF > |> several decades of use of form-style screens have been completely
 misguided,> > |> but I'm more inclined to suspect that your perception is. > B > N.M. Maclaren, pers. comm.  Based on actual measurements of selfD > and other users, timed with a watch.  Data lost due to the passage" > of time.  So now you have it :-)  L That's what I suspected.  Leaving aside the question of whether personal andB anecdotal experience merits being termed 'evidence' at all, baldlyJ presenting it without qualification as 'evidence' sufficient to settle theF issue is at best extremely sloppy thinking and at worst intellectually
 dishonest.   >1B > Now, please produce your source OF AT LEAST THE SAME QUALITY ANDD > COMMERCIAL INDEPENDENCE that demonstrates that GUIs save time over> > going the same operations using a well-designed command line > interpreter.  I Xerox PARC's work in the '70s.  Apple's work in the '80s (as I mentioned,>G they performed wide studies on test groups).  Your own biases may carryaI equal weight with you but carry nothing like equal weight with me (nor, Ir0 would suspect, with any other neutral assessor).   - bill   ------------------------------  # Date: Mon, 15 Apr 2002 20:31:08 GMTm  From: cjt <cheljuba@prodigy.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)N+ Message-ID: <3CBB387F.78704A48@prodigy.net>t   Nick Maclaren wrote: >  <snip>E > Yes.  It prevents you taking an accurate record.  So that, when youeD > report a bug, the support people can always say that you must haveE > made a mistake.  This saves developers time because they don't haven > to fix most bugs.n >  <snip>  L I think I finally understand why Microsoft is so keen on GUIs.  Maybe that's1 also why I've never seen a crash dump on Windows.s   ------------------------------  # Date: Mon, 15 Apr 2002 20:39:29 GMTo# From: "John Smith" <a@nonymous.com>uG Subject: Re: learning how to use a computer (was: Re: Itanium troubles)mE Message-ID: <5UGu8.1525$Gq1.641@news02.bloor.is.net.cable.rogers.com>-  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3CBAFDB1.B041C9C8@videotron.ca... > Bill Todd wrote:K > > Cars today are mostly point and go.  *Driving* skills remain important,1 but3D > > required knowledge of the vehicle has been dramatically reduced. >c > Are you kidding ?  >eL > One has to learn how to use that fancy electronic car stereo that is often asI > complex to program as a VCR. One has to learn how to use that fancy GPSo moving) > map display, the trip computer etc etc.g > I > And it is only a matter of time before cars provide an ethernet port sov thatK > you can plug in into your PC and use your browser to get a report on youre > car's health.  >uL > The driving may hav gotten a bit simpler, but the car itself hasn't and isE > outfitted with many more gadgets and those compicate the car. A carp without 2 > adult electronic toys won't attract male buyers.      H General Motors doesn't have a help line for people who don't know how to drive. Imagine if they did ...  0 ------------------------------------------------  : *HelpLine:* "General Motors HelpLine, how can I help you?"G *Customer:* "I got in my car and closed the door and nothing happened!"rC *HelpLine:* "Did you put the key in the ignition slot and turn it?"e! *Customer:* "What's an ignition?"lF *HelpLine:* "It's a starter motor that draws current from your battery and turns over the engine."eA *Customer:* "Ignition? Motor? Battery? Engine? How come I have too3 know all these technical terms just to use my car?"r  0 ------------------------------------------------  : *HelpLine:* "General Motors HelpLine, how can I help you?"F *Customer:* "My car ran fine for a week and now it won't go anywhere!"$ *HelpLine:* "Is the gas tank empty?"! *Customer:* "Huh? How do I know?"sH *HelpLine:* "There's a little gauge on the front panel with a needle and8 markings from 'E' to 'F'. Where is the needle pointing?"8 *Customer:* "It's pointing to 'E'. What does that mean?"F *HelpLine:* "It means you have to visit a gasoline vendor and purchase: some more gasoline. You can install it yourself or pay the vendor to install it for you."F *Customer:* "What? I paid $12,000 for this car! Now you tell me that I< have to keep buying more components? I want a car that comes with everything built in!"  0 ------------------------------------------------  : *HelpLine:* "General Motors HelpLine, how can I help you?" *Customer:* "Your cars suck!"o *HelpLine:* "What's wrong?"m, *Customer:* "It crashed, that's what wrong!"" *HelpLine:* "What were you doing?"J *Customer:* "I wanted to run faster, so I pushed the accelerator pedal all the ? way to the floor. It worked for a while and then it crashed and  it won't start now!tL *HelpLine:* "It's your responsibility if you misuse the product. What do you expect us to do about it?"I *Customer:* "I want you to send me one of the latest version that doesn't  crash any more!"  0 ------------------------------------------------  : *HelpLine:* "General Motors HelpLine, how can I help you?"L *Customer:* "Hi, I just bought my first car, and I chose your car because it; has automatic transmission, cruise control, power steering,i$ power brakes, and power door locks."< *HelpLine:* "Thanks for buying our car. How can I help you?" *Customer:* "How do I work it?"l' *HelpLine:* "Do you know how to drive?"r$ *Customer:* "Do I know how to what?"' *HelpLine:* "Do you know how to drive?"-G *Customer:* "I'm not a technical person. I just want to go places in mya car!"u   ------------------------------  # Date: Mon, 15 Apr 2002 21:18:09 GMTe* From: "Bill Todd" <billtodd@metrocast.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)i@ Message-ID: <lsHu8.63156$%8.5370837@bin2.nnrp.aus1.giganews.com>  5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in messaget* news:a9fdpt$jlg$1@pegasus.csx.cam.ac.uk...B > In article <5CGu8.51253$K5.4873808@bin5.nnrp.aus1.giganews.com>,+ > Bill Todd <billtodd@metrocast.net> wrote:1 > >hL > >The question is whether the alternative help interfaces you favor do suchJ > >things any better.  If they don't, then that's a wash and we move on to > >other areas.s > E > Oh, they did and they do.  But, given your abuse, I fail to see whye > I should explain further.n  J You always do seem to find some excuse for not backing up your claims withE anything substantial.  That's one major reason you may feel 'abused'.a   ...u  ( > As I said, I took actual measurements.  A Wow!  Now tell us your group sizes, and compositions, and how theaL measurements varied depending on user experience with the GUI interface, the command interface, and both.  J I suspect you have no idea of what it takes to gather evidence of any real significance in such areas.m     I have seen very littlep= > evidence that other people have done so, though I am pretty * > certain that some people must have done.  K Then you just haven't looked.  No surprise there:  you usually just seem toe9 assume that anything you don't know is not worth knowing.t   > E > >> Now, please produce your source OF AT LEAST THE SAME QUALITY ANDoG > >> COMMERCIAL INDEPENDENCE that demonstrates that GUIs save time overfA > >> going the same operations using a well-designed command lineh > >> interpreter.i > >SL > >Xerox PARC's work in the '70s.  Apple's work in the '80s (as I mentioned,J > >they performed wide studies on test groups).  Your own biases may carryL > >equal weight with you but carry nothing like equal weight with me (nor, I3 > >would suspect, with any other neutral assessor).l >yF > You may not be aware that Xerox PARC and the University of Cambridge@ > had pretty close links.  No, that was NOT one of their claims.C > The nearest that they came to that in any document that I saw waseA > a claim that it could be quicker to do a task that the user didTA > not do regularly with a GUI than a command language.  But I sawt > little evidence even of that.g  L I really don't care what you think you saw, Nick.  And while you likely haveG a more than usually sympathetic audience here to the virtues of command,J interfaces over GUIs, I suspect even they would like to see more substance and less assertion.u   >c@ > You seem to be unaware that Apple had a commercial interest inB > their claims, and it is generally agreed that considerable doubt5 > needs to be placed on claims when that is the case.l  L Let's see:  1) you make an assertion that a commercial interest was involvedI in justifying Apple's new interface, when all accounts I've read indicatetL that Apple first investigated and then adopted said interface because it wasG convinced of its value, and then  2) wave your hands about the 'general E agreement' that 'considerable doubt' needs to be given to such claimsgI (failing to realize that you didn't offer any proof of commercial bias in  the first place).   F I have significant reservations about finding myself in agreement withJ Dennis about much of anything, but in this case it's more than tempting toH agree with his assessment of you.  Perhaps I should instead associate myD views with most of the other people who've recently had this kind ofF 'discussion' with you in comp.arch:  while I agree that truth is not aE matter of popularity, I suggest that you should at least consider thenJ implications of a situation in which the major preponderance of opinion ofK reasonably intelligent and informed people seems to be against you across a  wide variety of issues.e     Certainly,C > one claim of that nature which I did see was clearly bogus, as it3D > assumed typing speeds that were low even by casual user standards. >1F > As I said, I took actual measurements on real people in as unbiassedD > a fashion as I could manage (and I am a statistician).  You are at: > liberty to think me a liar, but you are wrong if you do.  L I don't think you're a liar, Nick, but I do think you're either very (and to9 all appearances unjustifiably) arrogant or very confused.n   - bill   ------------------------------   Date: 15 Apr 2002 22:24:10 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles) 0 Message-ID: <a9fjua$oig$1@pegasus.csx.cam.ac.uk>  @ In article <lsHu8.63156$%8.5370837@bin2.nnrp.aus1.giganews.com>,) Bill Todd <billtodd@metrocast.net> wrote:    [ Considerable abuse snipped. ]   ) >> As I said, I took actual measurements.  >>B >Wow!  Now tell us your group sizes, and compositions, and how theM >measurements varied depending on user experience with the GUI interface, them >command interface, and both.l >bK >I suspect you have no idea of what it takes to gather evidence of any reale >significance in such areas.  @ As should be obvious, I do, which is why I never wrote up what I@ did nor bothered to keep the records.  My measurements were goodC enough to disprove the claims of the GUI propagandists of the time, C but I wasn't interested in a religious war.  There is no convincingh dogmatists.,   >  I have seen very little> >> evidence that other people have done so, though I am pretty+ >> certain that some people must have done.u >yL >Then you just haven't looked.  No surprise there:  you usually just seem to: >assume that anything you don't know is not worth knowing.  5 Well, why don't you provide a proper reference, then?     C The background to my taking measurements is interesting, especiallyu. in the light of your, er, emotional outbursts.  B A certain Senior Person fell into the hands of the Apple sect, was@ converted, and became an evangelist.  One of his claims was thatA the interface was so intuitive that it didn't need documentation,oA a help system or support staff.  And another was that a GUI woulds= always be faster for file management and similar tasks (if wet< ignore scripting).  So I took it upon myself to investigate.  @ Well, I couldn't find such claims in Xerox PARC's papers, thoughA he and his ilk claimed that they were there.  Upon querying this,eB I was told that I hadn't looked properly and that such claims were> there - but was not given a reference, despite asking for one.  @ And I investigated the claims, both on myself and on a few usersD and tasks selected for being 'typical.  Well, curiously, both claims? seemed to be the converse of the truth - surprisingly so.  UponeD querying this, I was abused for not having done a fair test - before enquiring what test I had done.c  B Now, my measurements did NOT prove that a GUI was harder to use or@ slower than a CLI, but they did cast considerable doubt over theC claims that a GUI was necessarily better.  Over the next 5 years oriA so, I enquired from a good many people working in this area abouti? the matter (including some from Xerox PARC) and universally goto6 negative answers as to the existence of such evidence.  C I am highly amused to see you spouting the same dogma as the Seniorf> Person, and reacting exactly the same way to both requests for5 evidence of your claims and evidence of the converse.   B As someone who uses both GUIs and CLIs, often for the same task, I@ am firmly of the opinion that there is no systematic difference.@ A good user interface, whether in terms of ease or speed of use,A comes from good design - it is as simple as that.  The GUI versusf1 CLI religious war is irrelevant to those aspects.h     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 334679o   ------------------------------  % Date: Mon, 15 Apr 2002 16:33:29 -0700d& From: name99@mac.com (Maynard Handley)G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)e7 Message-ID: <name99-1504021633290001@handma2.apple.com>   E In article <5UGu8.1525$Gq1.641@news02.bloor.is.net.cable.rogers.com>,v$ "John Smith" <a@nonymous.com> wrote:  < > "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message( > news:3CBAFDB1.B041C9C8@videotron.ca... > > Bill Todd wrote:M > > > Cars today are mostly point and go.  *Driving* skills remain important,t > butaF > > > required knowledge of the vehicle has been dramatically reduced. > >u > > Are you kidding ?s > > N > > One has to learn how to use that fancy electronic car stereo that is often > asK > > complex to program as a VCR. One has to learn how to use that fancy GPSo > moving+ > > map display, the trip computer etc etc.b > >nK > > And it is only a matter of time before cars provide an ethernet port sol > thatM > > you can plug in into your PC and use your browser to get a report on your  > > car's health.e > >iN > > The driving may hav gotten a bit simpler, but the car itself hasn't and isG > > outfitted with many more gadgets and those compicate the car. A cara	 > withoutm4 > > adult electronic toys won't attract male buyers. >   I The implications of the stuff below are bullshit, and the reason they areeJ bullshit is that the car is much more like GUI than the textual interface.J The GUI is not necessarily "immediately" obvious to a novice who has neverF seen a computer before and has no idea what sort of things can be doneC with one. That is not the claim. The claim is that, once things are I explained, they better match the mental constructs of most users and thus-F are more easily remembered. This is, of course, what cars likewise areA trying to acheive. As the example below gives, you do not have an>J intuitive idea at birth of what a fuel guage does, but having it explainedF once makes it obvious and you'll probbably remember it for the rest of* your life. Having to type into a terminal     frq -xw gsl >G to get the amount of gasoline left in your tank is NOT obvious and willu, not be remembered without considerable pain.   Maynard         J > General Motors doesn't have a help line for people who don't know how to  > drive. Imagine if they did ... > 2 > ------------------------------------------------ > < > *HelpLine:* "General Motors HelpLine, how can I help you?"I > *Customer:* "I got in my car and closed the door and nothing happened!"hE > *HelpLine:* "Did you put the key in the ignition slot and turn it?"h# > *Customer:* "What's an ignition?"eH > *HelpLine:* "It's a starter motor that draws current from your battery > and turns over the engine."aC > *Customer:* "Ignition? Motor? Battery? Engine? How come I have to 5 > know all these technical terms just to use my car?"m > 2 > ------------------------------------------------ > < > *HelpLine:* "General Motors HelpLine, how can I help you?"H > *Customer:* "My car ran fine for a week and now it won't go anywhere!"& > *HelpLine:* "Is the gas tank empty?"# > *Customer:* "Huh? How do I know?"dJ > *HelpLine:* "There's a little gauge on the front panel with a needle and: > markings from 'E' to 'F'. Where is the needle pointing?": > *Customer:* "It's pointing to 'E'. What does that mean?"H > *HelpLine:* "It means you have to visit a gasoline vendor and purchase< > some more gasoline. You can install it yourself or pay the  > vendor to install it for you."H > *Customer:* "What? I paid $12,000 for this car! Now you tell me that I> > have to keep buying more components? I want a car that comes > with everything built in!" > 2 > ------------------------------------------------ > < > *HelpLine:* "General Motors HelpLine, how can I help you?" > *Customer:* "Your cars suck!"i > *HelpLine:* "What's wrong?"-. > *Customer:* "It crashed, that's what wrong!"$ > *HelpLine:* "What were you doing?"L > *Customer:* "I wanted to run faster, so I pushed the accelerator pedal all > the3A > way to the floor. It worked for a while and then it crashed ande > it won't start now!iN > *HelpLine:* "It's your responsibility if you misuse the product. What do you > expect us to do about it?"K > *Customer:* "I want you to send me one of the latest version that doesn't  > crash any more!" > 2 > ------------------------------------------------ > < > *HelpLine:* "General Motors HelpLine, how can I help you?"N > *Customer:* "Hi, I just bought my first car, and I chose your car because it= > has automatic transmission, cruise control, power steering,R& > power brakes, and power door locks."> > *HelpLine:* "Thanks for buying our car. How can I help you?"! > *Customer:* "How do I work it?"J) > *HelpLine:* "Do you know how to drive?"t& > *Customer:* "Do I know how to what?") > *HelpLine:* "Do you know how to drive?"tI > *Customer:* "I'm not a technical person. I just want to go places in my  > car!"d   ------------------------------  # Date: Tue, 16 Apr 2002 00:17:38 GMTi/ From: Andrew Reilly <andrew@gurney.reilly.home>fG Subject: Re: learning how to use a computer (was: Re: Itanium troubles)C> Message-ID: <C4Ku8.12667$o66.38192@news-server.bigpond.net.au>  4 On Tue, 16 Apr 2002 06:20:18 +1000, Bill Todd wrote:7 > Scripting (or near-scripting using command piping andvE > regular-expression-oriented utilities) is certainly useful for suchpF > operations.  It also has little to do with the operations most usersA > (remembering that most users are near-neophytes sitting at PCs)e
 > perform.  H Marketing numbers seem to suggest that sales of PCs to first time buyersE is plateauing, and most sales are second or later systems, or variouse portable devices.   D One can only hope that the GUI architects will take the hint that PCH users are now or will soon generally _not_ be near-neophytes, and we can all get some useful work done.  G (Well, I never stopped using my Unix systems, even though they've growntI some shiny graphical applications over the years.  I mean the rest of you?9 who spend significant time inside file managers or IDEs.)o   -- Andrew   ------------------------------  # Date: Tue, 16 Apr 2002 01:15:06 GMTg* From: "Bill Todd" <billtodd@metrocast.net>G Subject: Re: learning how to use a computer (was: Re: Itanium troubles)-= Message-ID: <tWKu8.197$XV5.39813@bin4.nnrp.aus1.giganews.com>   5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in messageo* news:a9fjua$oig$1@pegasus.csx.cam.ac.uk...B > In article <lsHu8.63156$%8.5370837@bin2.nnrp.aus1.giganews.com>,+ > Bill Todd <billtodd@metrocast.net> wrote:t > ! > [ Considerable abuse snipped. ]' >h+ > >> As I said, I took actual measurements.' > >nD > >Wow!  Now tell us your group sizes, and compositions, and how theK > >measurements varied depending on user experience with the GUI interface,y thel > >command interface, and both.t > >yH > >I suspect you have no idea of what it takes to gather evidence of any real > >significance in such areas. >I > As should be obvious, I do,e  G As is likely obvious to everyone *but* you, you have yet to provide anyH indication that you do.   %  which is why I never wrote up what I:B > did nor bothered to keep the records.  My measurements were goodE > enough to disprove the claims of the GUI propagandists of the time,oE > but I wasn't interested in a religious war.  There is no convincingh
 > dogmatists.f  - You've got that right.  Now look in a mirror.n   >t > >  I have seen very little@ > >> evidence that other people have done so, though I am pretty- > >> certain that some people must have done.n > > K > >Then you just haven't looked.  No surprise there:  you usually just seem- to< > >assume that anything you don't know is not worth knowing. >t7 > Well, why don't you provide a proper reference, then?s  J Because when *I* was looking into this was 20 years ago, and I'm not aboutJ to redo the search that you've clearly never done in the first place.  TheK fact remains that research in this area *does* exist, and all you've got toHD offer in return is some unspecified 'measurements' you claim to haveL performed on what appear to be decidedly atypical user populations (at leastL for current computer users; they might have been more typical back then, butL that's not particularly relevant now that the PC has penetrated the masses).   >  >dE > The background to my taking measurements is interesting, especially 0 > in the light of your, er, emotional outbursts.  K The only emotion I'm starting to feel is disgust, Nick.  Perhaps you'd carecH to offer an explanation of why, if GUIs are inferior to command-orientedK control for anything like the majority of users, people aren't still mostlylK using the DOS command window option on Windows systems in preference to theeD graphical interface.  Or why forms-oriented interfaces are virtuallyD universal in things like point-of-sale terminal and other data-entry/ applications.  Is everyone out of step but you?a   - bill   ------------------------------  + Date: Tue, 16 Apr 2002 04:23:40 +0000 (UTC)n7 From: dsiebert@excisethis.khamsin.net (Douglas Siebert)aG Subject: Re: learning how to use a computer (was: Re: Itanium troubles)o+ Message-ID: <a9g90c$lik$1@sword.avalon.net>o  / JF Mezei <jfmezei.spamnot@videotron.ca> writes:s  M >And it is only a matter of time before cars provide an ethernet port so that J >you can plug in into your PC and use your browser to get a report on your >car's health.    G Already there.  OK, not ethernet, but you can buy an adaptor to connecteD a laptop's serial port to the diagnostic port on many new cars alongH with software to read fault codes from the engine computer, etc. as wellE as reset various preferences for the operation of the stereo, keylesss entry, etc., etc.t  E Automakers are going to change the electrical system from 12 volts to D 42 volts to power all the additional stuff they are adding.  There'sG even talk about running fibre optics through cars in the future to meetnG data requirements down the road for more advanced systems.  And at thatw@ time, yes, you'll probably have an ethernet port under the hood.   --H Douglas Siebert                          dsiebert@excisethis.khamsin.net  J A good friend will help you move, a true friend will help you move a body.   ------------------------------  % Date: Mon, 15 Apr 2002 10:33:52 -0700s' From: David Mathog <mathog@caltech.edu>s" Subject: Re: LVD SCSI on PWS 500au+ Message-ID: <3CBB0F00.845C7E3C@caltech.edu>    "Keith A. Lewis" wrote:e > D > >I'm using http://www.acard.com/eng/product/scside/aec-7720uw.html. > >in my PW 500au, Tru64 v51.A. Drive is IBM : > K > That's an interesting product, but why would you use a converter when thet" > 500au has 2 built-in IDE busses?  J Performance maybe?  The DS10, for instance, had remarkably unimpressive IOJ with IDE disks through its built in IDE interface.  The 500au may have the4 same poor performance.  There's a test package here:  A ftp://saf.bio.caltech.edu/pub/software/benchmarks/mybenchmark.zipt  P   which has some simple code for testing read/write speed on VMS disks (and alsoB other disks on Unix systems).  There are some tables in there with5 results from various systems for comparison purposes.   G I suspect that the reason you had to bump up the priority on your othera applicationaO when the ftp server was moving data was that the IDE read/write method was botho slowL and was generating CPU interrupts relatively frequently at a higher priority	 than youraP user level application.  If that's true then if you set up a RAMDISK on your VMSN system you'll probably find that the ftp server load is no longer significant.@ For what it's worth, on our late lamented VMS DS10 with U2W SCSIK the FTP server never seemed to load the system significantly even when veryn larges files were being downloaded.   Regards,   David Mathog mathog@caltech.edu   ------------------------------  % Date: Mon, 15 Apr 2002 21:55:04 +0200  From: Dirk Munk <munk@home.nl>" Subject: Re: LVD SCSI on PWS 500au& Message-ID: <3CBB3018.3030601@home.nl>   Keith A. Lewis wrote:o  M >Thanks to everybody for all the helpful suggestions.  In the end I went withaM >IDE, just because those drives are so cheap (got a 60 GB 7200 RPM Maxtor forwM >$100 new).  It seems to work just fine on VMS, except my real time app had a L >little contention problem when I was trying to upload files and use the appI >at the same time.  I bumped up the application process priorities to 8-9tJ >(leaving the FTP server at 4), and the problem went away but the speed of, >the FTP transfers dropped to around 1 MB/s. >Iw >Bob Marcan <bob.marcan@aster.si> writes in article <3CB84E77.149AB8EA@aster.si> dated Sat, 13 Apr 2002 17:27:51 +0200:- >-C >>I'm using http://www.acard.com/eng/product/scside/aec-7720uw.htmla- >>in my PW 500au, Tru64 v51.A. Drive is IBM :  >> >aJ >That's an interesting product, but why would you use a converter when the! >500au has 2 built-in IDE busses?3 >$F This only works with a second generation PWS system. First generation I PWS systems have a unsupported IDE controller (not even the CD-Rom works t
 with VMS).   >  >>, >--Keith Lewis              klewis$mitre.org? >The above may not (yet) represent the opinions of my employer.p >e   ------------------------------  % Date: Mon, 15 Apr 2002 10:01:24 -0700s# From: "Tom Linden" <tom@kednos.com>s, Subject: Mixing Alphas and PCs on KVM switch9 Message-ID: <CIEJLCMNHNNDLLOOGNJIOEECEMAA.tom@kednos.com>y  : I put two PWS500au and two W2K servers on a Linksys 4 port> KVM switch, but the control key doesn't function properly when> passing through the switch, connected to either of the Alphas,+ e.g. ^M doesn't generate carriage return.  m  9 The PWS's are running 7.1-2 and 7.3 and they both respondrD the same.  Mouse and Video seem OK.  When I disconnect Keyboard fromB switch and connect directly into either PWS, it works as expected.  E The control key seems to work correctly on the W2K boxes when passingo through the switch.e  E I am using a Digital KB, LK97W-A2, which I guess is PS/2 compatibile.t   Anybody else experience this?e   ------------------------------  + Date: Mon, 15 Apr 2002 21:34:55 -0500 (CDT)e From: sms@antinode.org' Subject: MMOV$ALPHAVCR.EXE: bad colors?,) Message-ID: <02041521345506@antinode.org>m  C    For those of you who might care, this problem seems to have beenaE caused by a dubious graphics card.  I'm guessing that 8-bit stuff waslA ok, but 24-bit was scrambled, so all my usual applications looked ? normal, but the MPEG player suffered.  After a power cycle, theeH background pattern was similarly stripey, so I swapped the card with theF one (ZLXp-E3) from my similar Tru64 system, and all appears to be well? (except on the Tru64 system, where the suspect card seems to be 9 comparably flakey, although it is working for the nonce.)   G    I suppose that after seven years it's out of warranty, too.  Geez...t   --------  / From:	ALP::SMS           9-APR-2002 23:14:44.08 < >    Today's mystery: I recently cranked up the MPEG viewer,G > SYS$SYSTEM:MMOV$ALPHAVCR.EXE ("Alpha VCR V2.0(s) Apr 17 2000 23:49:10 J > BL1-1(Rev.1)", it says), on some old files where it had worked some timeI > ago.  Now, however, it acts as if it has a bad color map, or some othervG > odd defect.  The colors are defective, and the picture is marred by aoC > fairly fine vertical stripe pattern.  I know of no changes to theiI > configuration since it worked last.  Whacking MMOV and the X server hadeG > no effect.  Same for specifying "-colormap private".  Anyone have anye > ideas? > / > AlphaStation 200 4/233 running OpenVMS V7.2-1i > GYA0 is a ZLXp-E2 (PBXGA-BA)  H ------------------------------------------------------------------------  C    Steven M. Schweda               (+1) 651-699-9818  (voice, home)lC    382 South Warwick Street        (+1) 763-781-0308  (voice, work) G    Saint Paul  MN  55105-2547      (+1) 763-781-0309  (facsimile, work)i9    sms@antinode.org                sms@provis.com  (work)u   ------------------------------    Date: 15 Apr 2002 22:07:24 -0500+ From: young_r@encompasserve.org (Rob Young)e, Subject: OS400 in similar non-marketing bind3 Message-ID: <paZVSssTzGh0@eisner.encompasserve.org>2   	Interesting find here...4  B http://www.computerworld.com/storyba/0,4125,NAV47_STO70141,00.html  L More than 3,000 users of IBM's venerable iSeries midrange systems - formerlyI known as the AS/400 - will gather in Nashville this week for the biannual3 Common trade show. c   [snip]  B IBM's perceived lack of marketing of the iSeries, which has been aM long-standing complaint among some users, is sure to be high on the agenda as.H well, said Al Barsa, president of Barsa Consulting Group LLC, a Purchase$ N.Y.-based iSeries consulting firm.     K "For iSeries users, there are just two types of computers in the world: thelL iSeries and inferior computers," Barsa said. "Most people are upset that IBM0 isn't doing enough to tell the world about it."     O At the same time, user satisfaction with IBM's iSeries technology remains high,sI he said. "They are bringing in all of the competitive features that otheroH platforms are introducing and, in many cases, are actually leading [more! popular] platforms," Barsa said.      ( 	Let's have some fun and re-write, okay?    J More than 1500 users of Compaq's venerable VMS Operating System - formerlyD known as OpenVMS - will gather in St. Louis this week for the annual CETS trade show. g   [snip]  = Compaq's perceived lack of marketing of VMS, which has been acM long-standing complaint among some users, is sure to be high on the agenda as K well, said Al Buffo, president of Buffo Consulting Group LLC, a Lonk Islande  N.Y.-based VMS consulting firm.     B "For VMS users, there are just two types of OSes in the world: VMSB and inferior OSes," Buffo said. "Most people are upset that Compaq0 isn't doing enough to tell the world about it."     M At the same time, user satisfaction with VMS remains high, he said. "They areo? bringing in all of the competitive features that other OSes areeO introducing and, in many cases, are actually leading [more popular] platforms,"o Buffo said.    ===   ) 	"Oh, but they won't market it properly!"h   	"Boo hooo!"   				Robk   ------------------------------  % Date: Tue, 16 Apr 2002 01:43:31 -0400H- From: JF Mezei <jfmezei.spamnot@videotron.ca>>0 Subject: Re: OS400 in similar non-marketing bind, Message-ID: <3CBBBA02.9791E452@videotron.ca>   Rob Young wrote:  N > More than 3,000 users of IBM's venerable iSeries midrange systems - formerlyK > known as the AS/400 - will gather in Nashville this week for the biannual- > Common trade show.  N There is a big difference between AS400 and VMS. In the case of VMS, it bringsE Digital/Compaq/HP/whatever technologies and capabilities that are not L available on other platforms. It is also scalable from desktop to mainframe.  N AS400 is a midrange that was designed to replace the system 34/36/38 which hadN been designed to be different enough from MVS/370 so that if IBM was forced toP divest of midrange, the buyer woudln't be inheriting any mainframe technologies.  M While they can complain about limited marketing and internal competition fromHO IBM's linux and AIX efforts, do you hear about fears that IBM might nix AS400 ?a   ------------------------------    Date: 16 Apr 2002 02:39:09 +0800, From: Paul Repacholi <prep@prep.synonet.com>1 Subject: Re: OT Best Practises in the IT Industryn- Message-ID: <87662s3iya.fsf@prep.synonet.com>f  + rob.buxton@wcc.govt.nz (Rob Buxton) writes:   * > I've been hit with a report that states:  3 > Minimum 30 day password expiry as "Best Practise"   C I think they mean a 30 day minimum time before the luser can changet their password.t   ; > Well I've seen 30, 60, 120 and 180 depending on the site.   . This would be password expiry periods I think.  tA > Anyone know of any IT Standards Pages that have a list of "Besty > Practises"  D The 'Rainbow Books' are probably about the best, but they need to beD used more as pointers to what the questions are than as prescriptive4 answers. Unless you need mil type security policies.  oE The big bugbear is the 8 char limit on many unix systems for passwordiF length. I've used 12 char minimum for ages, even with young kids. This? almost forces the use a a pass phrase, and is easy to remember.   B I was looking at some ex-gov PC yesterday. Including the 'dial-in'B utility for windows the dept had. Including phone number, account, password...t     -- s< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.-@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Mon, 15 Apr 2002 19:18:18 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca>g1 Subject: Re: OT Best Practises in the IT Industryd, Message-ID: <3CBB5FB9.28406760@videotron.ca>  5 > > Minimum 30 day password expiry as "Best Practise"H  K When I ran an email server linking different organisations, the associationoK (my customer) didn't care, so I what I ended up doing is ask each user what-J their own banks' policy was and set their account to that. They were quiteK happy because they would change their password on my system on the same day + that they changed their office's passwords.2  L Then there were users who would have prefered never changing their passwordsK and really hated that i would impose a 90 day limit. And I could see on therK logs that they would struggle to find an acceptable password, followed by al> phone call telling me my system wouldn't allow their password.   ------------------------------  # Date: Tue, 16 Apr 2002 00:56:39 GMTk) From: rob.buxton@wcc.govt.nz (Rob Buxton) 1 Subject: Re: OT Best Practises in the IT Industrya2 Message-ID: <3cbb7349.2233076774@news.wcc.govt.nz>   On 15 Apr 2002 12:33:25 -0500,C clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote:w  _ >In article <3cba0910.2140347667@news.wcc.govt.nz>, rob.buxton@wcc.govt.nz (Rob Buxton) writes:n >> s+ >> I've been hit with a report that states:y4 >> Minimum 30 day password expiry as "Best Practise"
 >  ^^^^^^^ > ( Oooopps ... that should've read Maximum. Other stuff was  Minimum = 1 day  Password History = 5 passwords  D The report writers had no idea about the Intrusion mechanisms in VMS< and loads of other stuff I'm having to re-explain to them...  F But the main thing that has got me is blanket statements in the report/ indicating that something is "Best Practise" ! n  E Exactly whose idea of "Best Practise" this is, is what I'm now havingc difficulty finding out.u  C Points taken as regards VMS History, Dictionary etc. but again, the@C Report Writers didn't ask about these and made assumptions based on  non-VMS Systems.  = Sigh...... just gotta go back and make things clearer to 'em.i   Thanks for the feedback.   Rob.     >See below.a >d >> n< >> Well I've seen 30, 60, 120 and 180 depending on the site. >> iB >> Anyone know of any IT Standards Pages that have a list of "Best >> Practises" D >> Tried Google but I'm probably looking for a General Policies Site$ >> rather than a this is our Policy. >> sF >> I'm aware of reaons against shorter password expiry e.g. Users moreG >> likely to write it down and leave it in their desk drawer! So, againeA >> I'm not really looking for the pros and cons, just a, probablye! >> mythical, Best Practises site.  >> i >oJ >It looks like you have interpreted this as the longest time that the userH >can go without changing the password, but note that it says minimum and >not maximum lifetime. > L >My guess is that the report is thinking in terms of those operating systemsH >that don't maintain a password history and this is a workaround to stopE >users from changing the password straight back to it's original one.  > J >If that's the case, you can point to VMS's password history capability to) >see if that's an acceptable alternative.l >i >Simon.h >w >-- C >Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP       n, >Microsoft: The Lada of the computing world.   ------------------------------  # Date: Mon, 15 Apr 2002 20:36:03 GMT * From: "Bill Todd" <billtodd@metrocast.net>9 Subject: Re: Prediction:  VMS lives, merger or no merger!y@ Message-ID: <SQGu8.51302$K5.4884907@bin5.nnrp.aus1.giganews.com>  > "Keith Parris" <KeithParris_NOSPAM@yahoo.com> wrote in message7 news:6ec1251e.0204150700.5fb39886@posting.google.com...s   ...   H > I recently happened across a pointer to info that came out of HP right= > at the time of the merger announcement (Sept. 6, 2001) (seeeG > http://www2.tru64.org/phorum/read.php?f=8&i=47&t=32 )  With regard tof" > VMS, HP's Public Relations said:H > "We are committed to the plan laid out for Alpha when Compaq announcedC > its agreement with Intel to standardize its 64-bit servers on theo > Itanium processor family."  K Indeed.  But part of that plan was the port of Tru64 to Itanic, too:  givenlD that they chose to renege on that portion, their commitment to other! portions becomes at best suspect.o  4 Note also two more gems in that same cited document:  < "HP-UX would be extended to support Alpha systems."  Really?  F "MPE will continue to be managed as it is currently."  Since they thenL turned around and announced its cancellation two months later, once again weG have an indication of just how much faith to place in these statements.d   - bill   ------------------------------    Date: 15 Apr 2002 17:39:00 -0700( From: bob@instantwhip.com (Bob Ceculski)9 Subject: Re: Prediction:  VMS lives, merger or no merger!h= Message-ID: <d7791aa1.0204151638.7f90af6b@posting.google.com>   a JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message news:<3CBB0363.4F2BBA15@videotron.ca>...e > Keith Parris wrote:tD > > "Compaq has made a long-term commitment to its important OpenVMSE > > customers and will continue to deliver on its committed roadmap."e > K > That could be interpreted that of the relatively few customers considereduK > "important", Compaq had promised to provide support for X years for theirrJ > existing installations.  The wording is specific enough that it makes on > question their true intents. > N > Had the text said "Compaq has made long term commitment to the VMS operatingO > system and will continue to deliver on its committed roadmap", then the hopeso  > of VMS woudl have been higher. > G > Note also that this mentions only COMPAQ. Not HP. What Compaq says is:8 > irrelevant. It is what the new owner says that counts.  $ but my letter is signed by Carly ...   ------------------------------    Date: 15 Apr 2002 14:41:04 -0700/ From: chris@applied-synergy.com (Chris Scheers)c& Subject: Re: Printserver 32 Error Code= Message-ID: <754a27c1.0204151341.50e74eda@posting.google.com>c   "Webb, William W Raleigh, NC" <wwebb1@email.usps.gov> wrote in message news:<032D86684229CC4AB84588A0BE50FC4002FBBC@rlghncst625.usps.gov>...1 > We have a venerable Printserver 32 no longer oni) > support and long bereft of documentage.a > 4 > Someone in their infinite wisdom power cycled said7 > Printserver on Friday and now we're getting the error. > message 0F10.0080. > 9 > If there's any significance to this beyond "I can't getl7 > a load from my host" (my guess) can anyone enlighten?i    = As Paul Anderson said, this is probably the upper paper tray.   E In particular, there is a ridge that runs along the side of the papereB tray closest to the front of the printer.  This front edge of this@ ridge engages a switch when the tray in inserted in the printer.  C It seems to be fairly common that the front part of this ridge getslA broken off and no longer engages the switch.  This can cause thisV error.  B Either fix the paper tray or get another paper tray and see if the printer will boot.  
 Good luck!   ------------------------------    Date: 15 Apr 2002 13:59:12 -07001 From: KeithParris_NOSPAM@yahoo.com (Keith Parris)s% Subject: Re: Processes in state RWSCSd= Message-ID: <6ec1251e.0204151259.6f5a1a9c@posting.google.com>-  v KeithParris_NOSPAM@yahoo.com (Keith Parris) wrote in message news:<6ec1251e.0204091324.57763ef9@posting.google.com>...F > VMS 7.3 improves lock remastering by at least an order of magnitude,B > so you may not need to use PE1 at all after you're running that.  = 7.2-2 also includes the fast-remastering capability from 7.3.2. ----------------------------------------------. Keith Parris | parris at encompasserve dot org   ------------------------------  % Date: Mon, 15 Apr 2002 15:09:38 -0400e, From: "Frank Sapienza" <sapienza@noesys.com>J Subject: Question about system messages, $SET MESSAGE, MESSAGE utility ..., Message-ID: <a9f8h702cu5@enews2.newsguy.com>  J It's actually not that complicated, I don't think, so here's the question:  C By what mechanism does VMS include and/or select message files from: SYS$MESSAGE?  K Let me expand a bit.  I know all about $SET MESSAGE, but that only works on H a per-process basis.  I'm looking for how I can put a shareable image inI SYS$MESSAGE (or wherever) and have it included it the system message list G automatically, for every user, without doing a $SET MESSAGE in SYLOGIN.g  J I can't find any reference that discusses how system messages are resolvedJ to the appropriate shareable image in SYS$MESSAGE.  It seems to me that asL layered products are installed they can put a shareable image in SYS$MESSAGEE and do *something* to then make the messages available to the system.rK Simply putting the shareable image in that directory doesn't seem to be thet only requirement.E     Thanks in advance.   ------------------------------    Date: 15 Apr 2002 16:27:21 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)sN Subject: Re: Question about system messages, $SET MESSAGE, MESSAGE utility ...3 Message-ID: <mvOosIKroMKb@eisner.encompasserve.org>   [ In article <a9f8h702cu5@enews2.newsguy.com>, "Frank Sapienza" <sapienza@noesys.com> writes:e  L > I can't find any reference that discusses how system messages are resolvedL > to the appropriate shareable image in SYS$MESSAGE.  It seems to me that asN > layered products are installed they can put a shareable image in SYS$MESSAGEG > and do *something* to then make the messages available to the system.   A I do not know believe that is the case, even with facilities from0B VMS itself that do not have their message files built into SYSMSG. I get the following results:   	$ exit %x0a38074m) 	%BACKUP-F-NOMSG, Message number 00A38074s  * That is not a whole lot more helpful than:   	$ exit %x2398074'& 	%LJK-F-NOMSG, Message number 02398074  9 Of course for shared messages, one gets the more helpful:    	$ exit %x0a3109ap- 	%BACKUP-E-OPENIN, error opening !AS as inputk 	$ exit %x239109as* 	%LJK-E-OPENIN, error opening !AS as input  E But BACKUP and LJK are both registered facilities.  I don't think youe? can get the above with unregistered (site-specific) facilities.b  M > Simply putting the shareable image in that directory doesn't seem to be thea > only requirement.t  @ My response above, however, deals only with DCL actions.  If youA want to resolve a message while your image is running (registeredaA facility or not) be sure at least one message is accessed by yourR1 code as an external value.  In Ada that would be:   @ 	LJK_BADCMDCHAR : constant CONDITION_HANDLING.COND_VALUE_TYPE :=, 		SYSTEM.IMPORT_VALUE ( "LJK$_BADCMDCHAR" );  B The obvious "efficiency" mechanism of doing it all in source usingA MESSAGE/SDL will not give any warning if your message stub objectwD module is missing from your image.  It may even be that lack of suchB a reference causes problems if you try to artificially link in theA message stub, but it has been a while since I ran into this, so I0 don't remember the details.i   ------------------------------  # Date: Tue, 16 Apr 2002 01:56:27 GMT-2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)" Subject: Re: RAM for MicroVAX 33009 Message-ID: <fxLu8.37$Nc.1330267@cacnews.cac.cpqcorp.net>   ] In article <3CAB4A1C.12566839@pcde.inka.de>, Dennis Grevenstein <dennis@pcde.inka.de> writes:o  8 :Can RAM from a VAX 4000-200 be used in a MicroVAX 3300?  K   Yes, MS650 (M7621, M7622) works in the MicroVAX 3300 and in the VAX 4000 mJ   model 200, as well as the VAXstation 3200, MicroVAX 3500, MicroVAX 3600,K   MicroVAX 3800, and MicroVAX 3900 -- though the MS650-AA 8MB memory board iL   is restricted to the slowest of these systems, the 16MB and larger boards    do work in other boxes.   : :There would be no problem, if the 4000/200 still netboots; :without DSSI, but if it's dead I would like to put the RAMt :in a 3300.   H   Replace the MicroVAX 3300 with something from the last decade, please?    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Tue, 16 Apr 2002 03:01:48 GMTh2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: remote mailboxs9 Message-ID: <wuMu8.43$Nc.1330267@cacnews.cac.cpqcorp.net>g  [ In article <b3ad37dc.0204090933.2ef17331@posting.google.com>, mrotgeri@web.de (ROM) writes:sF :I have a C++ program that wants to communicate from one node "A" to a  :remote node "B" with a mailbox.  E   Please use sockets, DECnet, or (within a cluster) ICC (on V7.2 and  H   later).  Please do not use mailboxes.  You will find sockets, DECnet,    and ICC easier to use.  B :I need to read simultaneous without closing the connection to the	 :mailbox.g  D   Mailboxes are system-local constructs, with remote access providedG   using RMS and DECnet FAL -- this extra layer really tends to confuse IH   things, and this layering really tends to eliminate access to various I   of the most useful (asynchronous) features of mailboxes.  (This is not sK   to say you could not perserve in your current approach.  And potentially .8   even succeed.  I simply can't recommend the approach.)  D   If you do choose to use mailboxes, please use one mailbox for eachF   direction of a bidirectional link -- do not mix more than one readerF   with zero or more writers.  If you do try to use multiple readers onH   one mailbox, the (required) application turn-around protocol tends to    be the source of most errors.   D   Again, I'd use IP or DECnet $qio calls, sockets, ICC or otherwise.;   There are other communications techniques and tools, too.D  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------    Date: 15 Apr 2002 12:36:34 -0700( From: bob@instantwhip.com (Bob Ceculski)9 Subject: Re: Sue, a great promotion for "unhackable" vms! = Message-ID: <d7791aa1.0204151136.623cc60b@posting.google.com>n  R mustang@ucc.asn.au.invalid wrote in message news:<a8rcsg$v0l$1@enyo.uwa.edu.au>...+ > Bob Ceculski <BOB@instantwhip.com> wrote:yq > : bob@instantwhip.com (Bob Ceculski) wrote in message news:<d7791aa1.0204061322.2b99c1f7@posting.google.com>...i5 > :> remember m.c. hammer and "YOU CAN'T TOUCH THIS"?i6 > :> why not hire him to do a vms commercial?  Instead2 > :> of his line above, he could sing and dance to > :> p > :> "YOU CAN'T HACK THIS"!: > :> t7 > :> that would be an awesome commercial!  how about its. > :> Sue, or anyone else in Q or HP marketing? >  u: > : he and his dance team would be wearing OpenVMS tshirts: > : and you could film it at defcon10, and use some of the6 > : defcon10 people in the clip trying to hack vms but; > : failing ... if you want help directing the whole thing,16 > : I would be happy to oblige for free ... I have the+ > : whole thing laid out, and it's awesome!o >  > MC Hammer? Hello!e > A > The guy declares bankruptcy then goes off to become a preacher.rL > I'm sure there's grist for Bill Todd's metaphor mill in there, but I don'tC > that that's really the image that you -want- associated with VMS.  >  > D.  F wrong, that is exactly the image you want to draw the young linux/unix/ and windoze crowd who think vms is outdated ...    ------------------------------  % Date: Mon, 15 Apr 2002 14:43:28 -0500d+ From: Christopher Smith <csmith@amdocs.com>m9 Subject: RE: Sue, a great promotion for "unhackable" vms! J Message-ID: <7E008308CD77154485FEF878168D078E0178441A@CMIMAIL1.amdocs.com>   > -----Original Message-----8 > From: bob@instantwhip.com [mailto:bob@instantwhip.com]  H > wrong, that is exactly the image you want to draw the young linux/unix1 > and windoze crowd who think vms is outdated ...t  D Um, Bob -- chances are that the "young linux/unix and windoze crowd"C will also think that "MC Hammer" is outdated.  Aside from that, the  "music" was utter garbage ;)  D What VMS needs is a less "silly" commercial, like some guys in funny' looking space suits, and disco music...5   Chris7  ! Christopher Smith, Perl Developery Amdocs - Champaign, IL   /usr/bin/perl -e '? print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");t 'o  m   ------------------------------    Date: 15 Apr 2002 19:54:28 -0000= From: Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]>.9 Subject: Re: Sue, a great promotion for "unhackable" vms!s5 Message-ID: <20020415195428.6180.qmail@gacracker.org>h  9 On 15 Apr 2002, bob@instantwhip.com (Bob Ceculski) wrote:a, >mustang@ucc.asn.au.invalid wrote in message' >news:<a8rcsg$v0l$1@enyo.uwa.edu.au>...a   <snip>   >> MC Hammer? Hello! >> tB >> The guy declares bankruptcy then goes off to become a preacher.M >> I'm sure there's grist for Bill Todd's metaphor mill in there, but I don't-D >> that that's really the image that you -want- associated with VMS. >> t >> D.g >.G >wrong, that is exactly the image you want to draw the young linux/unixs0 >and windoze crowd who think vms is outdated ...  C MC Hammer is outdated. You've got to pick something current that ishK actually going to be around for a long time. That ain't easy. Pop icons aretJ a particularly poor choice as they generally have a very short shelf life.  J Anyway, the day we get TV advertising for VMS the skills shortage will getI worse due to the number of heart-attacks it induces in the VMS community.a     Doc. --  6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                              http://vmsbox.cjb.netf   ------------------------------  # Date: Mon, 15 Apr 2002 23:02:30 GMTeL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")9 Subject: Re: Sue, a great promotion for "unhackable" vms!i8 Message-ID: <00A0C81F.69235DE2@SSRL04.SLAC.STANFORD.EDU>  h In article <d7791aa1.0204151136.623cc60b@posting.google.com>, bob@instantwhip.com (Bob Ceculski) writes:S >mustang@ucc.asn.au.invalid wrote in message news:<a8rcsg$v0l$1@enyo.uwa.edu.au>...t, >> Bob Ceculski <BOB@instantwhip.com> wrote:r >> : bob@instantwhip.com (Bob Ceculski) wrote in message news:<d7791aa1.0204061322.2b99c1f7@posting.google.com>...6 >> :> remember m.c. hammer and "YOU CAN'T TOUCH THIS"?7 >> :> why not hire him to do a vms commercial?  Insteado3 >> :> of his line above, he could sing and dance toi >> :>  >> :> "YOU CAN'T HACK THIS"! >> :> 8 >> :> that would be an awesome commercial!  how about it/ >> :> Sue, or anyone else in Q or HP marketing?w >>  ; >> : he and his dance team would be wearing OpenVMS tshirtss; >> : and you could film it at defcon10, and use some of the 7 >> : defcon10 people in the clip trying to hack vms buti< >> : failing ... if you want help directing the whole thing,7 >> : I would be happy to oblige for free ... I have the , >> : whole thing laid out, and it's awesome! >> f >> MC Hammer? Hello! >> tB >> The guy declares bankruptcy then goes off to become a preacher.M >> I'm sure there's grist for Bill Todd's metaphor mill in there, but I don't D >> that that's really the image that you -want- associated with VMS. >>   >> D.  >tG >wrong, that is exactly the image you want to draw the young linux/unixl0 >and windoze crowd who think vms is outdated ...  H Bob, you're out of touch.  The young people today either don't know who H MC Hammer is (was) or think of him as a has-been of no current interest;I this would only confirm their impression that VMS is outdated.  (Not that  I'm in touch myself.)2  L They'd do better working up something with a quirky band like _They Might BeM Giants_, who'd probably be amused to write a VMS jingle.  (Actually, to reachmL CEOs rather than young hackers, they should get Elton John to rewrite -  no,I not Candle in the Wind again - but "I'm still standing", and make a video N showing huge disasters and VMS clusters staying up.  They could point out thatJ John himself was big in the 70s and is still big, like VMS.)  A lot of theJ people on corporate boards listened to him when they were younger and they still vaguely think he's hip.    -- Alana    O ===============================================================================l0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056gM  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210yO ===============================================================================t   ------------------------------  # Date: Tue, 16 Apr 2002 01:29:23 GMTo1 From: "David J. Dachtera" <djesys.nospam@fsi.net>a9 Subject: Re: Sue, a great promotion for "unhackable" vms!-' Message-ID: <3CBB810C.51F0142F@fsi.net>7   Christopher Smith wrote: >  > > -----Original Message-----: > > From: bob@instantwhip.com [mailto:bob@instantwhip.com] > J > > wrong, that is exactly the image you want to draw the young linux/unix3 > > and windoze crowd who think vms is outdated ...B > F > Um, Bob -- chances are that the "young linux/unix and windoze crowd"E > will also think that "MC Hammer" is outdated.  Aside from that, theu > "music" was utter garbage ;) > F > What VMS needs is a less "silly" commercial, like some guys in funny) > looking space suits, and disco music...r  1 How 'bout the Devo guys in their flower-pot hats?x   Crack that whip!   -- w David J. Dachterae dba DJE Systemse http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/o   ------------------------------  % Date: Mon, 15 Apr 2002 20:25:30 -0400r1 From: "Island (hpaq.net)" <dbturner@islandco.com>h* Subject: Re: Survey on Your Alpha Hardware/ Message-ID: <ubmrri4jvrln04@news.supernews.com>e  	 Bob et alh  I The bottom of the form, which for some reason was not showing, asaked for ' your email address and telephone number5    J We have a lady in this office - all she does is answer the phone and takes messagesJ He new "other function" is to call the telephone numbers and get a mail to addressmK Reason for this is that some companies are VERY QUIRKY about freebies beingnL sent to their employee. A couple of the chocolate companies are particularly7 strict about it, so we wanted a viable delivery address   I As for the survey, we aren't trying to sell Alpha systems immediately; wecE are trying to target the right audience for the hardware that we haveeK available and will also know what to stock when offered clearance deals etce by Compaq and affiliates   David T    -- Island Computers US Corp.u 2700 Gregory Street  Savannah GA 31404e Toll Free: 1-877 636 4332u International: 001 912 447 6622l  Facsimile:      001 912 201 0096 dbturner@hpaq.net  www.hpaq.net        8 Bob Koehler <koehler@encompasserve.org> wrote in message- news:pmxCt0xQViCS@eisner.encompasserve.org...iE > In article <ubjad3ktn5nc64@news.supernews.com>, "Island (hpaq.net)"o <dbturner@islandco.com> writes:n > > Hi everyoneo > > J > > Thought this was a good idea and ergo I am requesting this information > >bL > > We are finding it quite difficult to target specific customer types with our-2 > > specials, so I thought I would ask all of you. >sI >    I tried.  Tried Mozilla 0.9.7 on my Alpha, got an up arrow or a down6E >    arrow in a lot of scrollable windows, but not both so I couldn'tr >    scroll to the best answer.1 >2C >    Tried Netscape 6.2 on my Mac.  Got little tiny scrollbars thata >    wouldn't scroll.o >rF >    Finally, in frustration, tried Explorer on my Mac.  Filled in theH >    questions.  Submitted the survey (funny, how are they going to sendE >    me that T-shirt when they didn't ask for a snail-mail address?).iA >    Got a FrontPage error, now I know why I had to use Explorer.e >oJ >    Somehow I don't think this really the way to attract Alpha customers. >u   ------------------------------  % Date: Mon, 15 Apr 2002 20:26:25 -0400e1 From: "Island (hpaq.net)" <dbturner@islandco.com>a* Subject: Re: Survey on Your Alpha Hardware/ Message-ID: <ubmrsepg5evk15@news.supernews.com>s   Oh  < And the web page is being redone to fix that Frontpage error   DT8 Bob Koehler <koehler@encompasserve.org> wrote in message- news:pmxCt0xQViCS@eisner.encompasserve.org...oE > In article <ubjad3ktn5nc64@news.supernews.com>, "Island (hpaq.net)"  <dbturner@islandco.com> writes:s > > Hi everyone, > >eJ > > Thought this was a good idea and ergo I am requesting this information > >eL > > We are finding it quite difficult to target specific customer types with oure2 > > specials, so I thought I would ask all of you. >cI >    I tried.  Tried Mozilla 0.9.7 on my Alpha, got an up arrow or a downeE >    arrow in a lot of scrollable windows, but not both so I couldn'ti >    scroll to the best answer.r >oC >    Tried Netscape 6.2 on my Mac.  Got little tiny scrollbars that. >    wouldn't scroll.- >-F >    Finally, in frustration, tried Explorer on my Mac.  Filled in theH >    questions.  Submitted the survey (funny, how are they going to sendE >    me that T-shirt when they didn't ask for a snail-mail address?).eA >    Got a FrontPage error, now I know why I had to use Explorer.n > J >    Somehow I don't think this really the way to attract Alpha customers. >o   ------------------------------  % Date: Mon, 15 Apr 2002 14:01:56 -0400$ From: William_Bochnik@acml.com+ Subject: Re: Throughput and additional CPUsu> Message-ID: <OF60F6043B.7B10CBC5-ON85256B9C.00630294@acml.com>  # not to mention disk thruput as wellt      [                                                                                            h[                       Anamika                                                              m[                       <anamika@home.co                To:  Info-VAX@Mvb.Saic.Com            [                       m>                              cc:                                  o[                                                Subject: Throughput and additional CPUs     u[                       04/15/2002 01:49                                                     b[                       PM                                                                   c[                       Please respond                                                       b[                       to Anamika                                                           y[                       <anamika@home.co                                                     u[                       m>                                                                   e[                                                                                             [                                                                                                    Hi, 3   I just want to sound out more people about this :w> We have a machine with a webserver and a webserver application that7 protects "resources". Until last week there was certainl
 throughput7 whenever users accessed these resources from a browser.i; Last week the admins added a second CPU and they expect the-
 throughput toa> double. I said that throughput cannot double because the whole transaction ? has a I/O part, N/W part, CPU part and adding a second CPU will3 only? benefit the CPU part of the transaction. To actually double the@
 throughput> performance of the I/O part, N/W part also has to be increased
 (doubled). It this logic correct ?r  A This is on OVMS 7.x, OSU server (which is not all that relevant).a   -A          F ______________________________________________________________________;  The information contained in this transmission may contain6@ privileged and confidential information and is intended only forA the use of the person(s) name above.  If you are not the intended = recipient, or an employee or agent responsible for deliveringt3 this message to the intended recipient, any review,A@ dissemination, distribution or duplication of this communication? is strictly prohibited.  If you are not the intended recipient,cA please contact the sender immediately by reply e-mail and destroy # all copies of the original message.n   ------------------------------   Date: 15 Apr 2002 18:11:20 GMT1 From: Dale Dellutri <ddelQQQlutr@entQQQeract.com>o+ Subject: Re: Throughput and additional CPUsz+ Message-ID: <a9f548$45h$1@bob.news.rcn.net>   C On Mon, 15 Apr 2002 17:49:11 GMT, Anamika <anamika@home.com> wrote:s5 >   I just want to sound out more people about this :gF > We have a machine with a webserver and a webserver application that E > protects "resources". Until last week there was certain throughput t9 > whenever users accessed these resources from a browser.oL > Last week the admins added a second CPU and they expect the throughput to M > double. I said that throughput cannot double because the whole transaction iG > has a I/O part, N/W part, CPU part and adding a second CPU will only nM > benefit the CPU part of the transaction. To actually double the throughput  K > performance of the I/O part, N/W part also has to be increased (doubled).M > It this logic correct ?t  @ The most important thing to know is what is the bottleneck: IO, E Network or CPU?  If the process is already CPU bound, and it can use oB twice as much CPU without needing more IO or Network, then adding ( another CPU could double the throughput.  ? You don't give any indication of what bottlenecks are occuring.h  G In general, of course, if the application already has a balanced usage  C of the resources, then you would need to double all of them to get r twice the trhoughput.r   --  , Dale Dellutri -- ddelQQQlutr@entQQQeract.com   ------------------------------    Date: 15 Apr 2002 14:37:32 -0500 From: briggs@encompasserve.org+ Subject: Re: Throughput and additional CPUs 3 Message-ID: <h$Awh0yJheRc@eisner.encompasserve.org>c  W In article <Xns91F18C95465D6anamika@207.181.101.12>, Anamika <anamika@home.com> writes:- > Hi,-5 >   I just want to sound out more people about this :OF > We have a machine with a webserver and a webserver application that E > protects "resources". Until last week there was certain throughput f9 > whenever users accessed these resources from a browser.-L > Last week the admins added a second CPU and they expect the throughput to M > double. I said that throughput cannot double because the whole transaction lG > has a I/O part, N/W part, CPU part and adding a second CPU will only  M > benefit the CPU part of the transaction. To actually double the throughput uK > performance of the I/O part, N/W part also has to be increased (doubled).B > It this logic correct ?g   Yes.  No.  Maybe.0  E If the machine has some other CPU load in addition to the interactiveUA web browsing activity then you could conceivably find that adding C a second CPU would more than double throughput for your interactive @ web browsing activity.  Instead of getting 80% of the CPU, maybe you now get 180%.a  A If you are measuring throughput (as opposed to response time, foreB instance) and CPU is indeed your bottleneck and there are no otherB consumers of CPU then doubling CPU performance could indeed doubleB throughput as long as no other resource is currently more than 50%	 utilized.e  A If you are measuring response time then doubling the speed of thee@ CPU component could have a profound effect by virtue of reducingB the "waiting for the CPU" portion of your web browsing experience.  D If some other resource is a bottleneck (or is within a factor of twoB of becoming a bottleneck) then you won't see the expected doubling5 of throughput and may not see any improvement at all.t   	John Briggs   ------------------------------  % Date: Mon, 15 Apr 2002 22:03:07 +0200 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> + Subject: Re: Throughput and additional CPUs ' Message-ID: <3CBB31FB.85D2EB84@aaa.com>    Dale Dellutri wrote: >  > A > The most important thing to know is what is the bottleneck: IO,IF > Network or CPU?  If the process is already CPU bound, and it can useC > twice as much CPU without needing more IO or Network, then addinge* > another CPU could double the throughput.  D If "the process" is CPU bound, *and* if what the process is doing is singleB threaded, adding a second CPU will not help much. Swapping the CPU@ for a single CPU with double speed will/could double throughput.  C So you must have *multiple* processes fighting for CPU, or a singlef processkE built with multi threading to actualy run both CPU's at the same timer with better total perf.  @ Usualy swapping the whole box will normaly give you a faster boxC where all parts (CPU, I/O, mem) still are "in balance", just as you-  sad later in your post, b.t.w...   Jan-Erik Sderholm.<   > A > You don't give any indication of what bottlenecks are occuring.r > H > In general, of course, if the application already has a balanced usageD > of the resources, then you would need to double all of them to get > twice the trhoughput.a >  > --. > Dale Dellutri -- ddelQQQlutr@entQQQeract.com   ------------------------------  % Date: Mon, 15 Apr 2002 22:10:29 +0200y9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>r+ Subject: Re: Throughput and additional CPUsu' Message-ID: <3CBB33B5.AC504647@aaa.com>g   briggs@encompasserve.org wrote:M >  > C > If you are measuring response time then doubling the speed of theyB > CPU component could have a profound effect by virtue of reducingD > the "waiting for the CPU" portion of your web browsing experience.  F Correct, but, the question was not to double the CPU speed, but to addH a second CPU with *the same* speed, which is two quite different things.  A A single CPU with double speed will always give double speed to aoE single (CPU bound) process. A second CPU (Same speed) will not help at bitS= if there always is only *one* CPU bound process, and no other  process waiting for the CPU.   Jan-Erik Sderholm.    ------------------------------  # Date: Mon, 15 Apr 2002 20:49:27 GMT-  From: Anamika <anamika@home.com>+ Subject: Re: Throughput and additional CPUsE4 Message-ID: <Xns91F1AB2537EFEanamika@207.181.101.12>  4 Dale Dellutri <ddelQQQlutr@entQQQeract.com> wrote in$ news:a9f548$45h$1@bob.news.rcn.net:   K There are no bottlenecks as such and that is not the problem. The question oK is, if a transaction (like accessing a file mapped via a URL), consists of nH tasks that are I/O bound, N/W bound and CPU bound, then adding a second F CPU to the system in question will not double the throughput. All the J other parts need to be beefed up as well ie., the N/W part, I/O part etc., Is that right ?a   -A  E > On Mon, 15 Apr 2002 17:49:11 GMT, Anamika <anamika@home.com> wrote:t6 >>   I just want to sound out more people about this :G >> We have a machine with a webserver and a webserver application that aF >> protects "resources". Until last week there was certain throughput : >> whenever users accessed these resources from a browser.I >> Last week the admins added a second CPU and they expect the throughput D >> to double. I said that throughput cannot double because the wholeI >> transaction has a I/O part, N/W part, CPU part and adding a second CPUpH >> will only benefit the CPU part of the transaction. To actually doubleF >> the throughput performance of the I/O part, N/W part also has to be/ >> increased (doubled). It this logic correct ?r > B > The most important thing to know is what is the bottleneck: IO, G > Network or CPU?  If the process is already CPU bound, and it can use  D > twice as much CPU without needing more IO or Network, then adding * > another CPU could double the throughput. > A > You don't give any indication of what bottlenecks are occuring.a > I > In general, of course, if the application already has a balanced usage oE > of the resources, then you would need to double all of them to get s > twice the trhoughput.' >    ------------------------------    Date: 15 Apr 2002 15:35:40 -0500 From: briggs@encompasserve.org+ Subject: Re: Throughput and additional CPUsj3 Message-ID: <4YijS1N31Ib1@eisner.encompasserve.org>h  c In article <3CBB33B5.AC504647@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:n! > briggs@encompasserve.org wrote:u >> s >> sD >> If you are measuring response time then doubling the speed of theC >> CPU component could have a profound effect by virtue of reducingeE >> the "waiting for the CPU" portion of your web browsing experience.t > H > Correct, but, the question was not to double the CPU speed, but to addJ > a second CPU with *the same* speed, which is two quite different things.  > Barring some kind of single-process choke point as you mention< below, the effect on "waiting for the CPU" time is virtually@ indistinguishable between the two processor and the double-speed processor case.=  B For a well-behaved (no infinite-CPU hogs running at priority 1 andC request arrival rate independent of request service rate) workload,d@ the rule of thumb is that CPU queue length is the inverse of CPU idle percentage.  6 90% CPU utilization = 10 computable/current processes.  E Double the CPU speed or double the number of CPUs and you get 45% CPU B utilization = ~2 computable processes on average.  Your time spent0 waiting in queue just improved by a factor of 5.  C > A single CPU with double speed will always give double speed to a G > single (CPU bound) process. A second CPU (Same speed) will not help ai > bit ? > if there always is only *one* CPU bound process, and no others > process waiting for the CPU.  1 Indeed so.  Not a pleasant situation to discover.e   	John Briggs   ------------------------------  % Date: Mon, 15 Apr 2002 23:02:39 +0200e9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>D+ Subject: Re: Throughput and additional CPUs.' Message-ID: <3CBB3FEF.7DB29DF6@aaa.com>e  ? Correctly. And if one single file access is handled by a single-? process, and there is no other activity on the box at the time,h' the second CPU will not be used at all.c  A Switching the single CPU into a faster one, will, of course, makesB the CPU part of the process faster, but not the I/O and N/W parts.  E I'd say you must analyze the system, to find out where the bottleneckeA is today. This is a WEB server, right ? What about the throughputr; *outside* of the box ? The network, and the client itself ?   E Do you say that accessing a file today is both I/O, network *and* CPUsD bound at the same time ? Not very common, I'd say. And if they realy are,F upgrading either of CPU, I/O or N/W will not do any good. A faster CPUE (or two CPU's) will just spend more time waiting for the I/O and N/W.   B Now, this is very general speaking, I havn't *seen* the system :-)   Jan-Erik Sderholm.    Anamika wrote: > L > There are no bottlenecks as such and that is not the problem. The questionL > is, if a transaction (like accessing a file mapped via a URL), consists ofI > tasks that are I/O bound, N/W bound and CPU bound, then adding a secondnG > CPU to the system in question will not double the throughput. All therL > other parts need to be beefed up as well ie., the N/W part, I/O part etc., > Is that right ?p >  > -A >i   ------------------------------  # Date: Mon, 15 Apr 2002 22:37:12 GMToL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")+ Subject: Re: Throughput and additional CPUsg8 Message-ID: <00A0C81B.E02926EB@SSRL04.SLAC.STANFORD.EDU>  W In article <Xns91F18C95465D6anamika@207.181.101.12>, Anamika <anamika@home.com> writes:g >Hi,4 >  I just want to sound out more people about this :E >We have a machine with a webserver and a webserver application that sD >protects "resources". Until last week there was certain throughput 8 >whenever users accessed these resources from a browser.K >Last week the admins added a second CPU and they expect the throughput to tL >double. I said that throughput cannot double because the whole transaction F >has a I/O part, N/W part, CPU part and adding a second CPU will only L >benefit the CPU part of the transaction. To actually double the throughput J >performance of the I/O part, N/W part also has to be increased (doubled). >It this logic correct ?  J Well, you're both wrong.  (Or maybe the admins are close to right, if theyL really  understand the performance characteristics of their system under the load it's actually getting.)  L They may have determined that there's plenty of unused network resource, andL plenty of unused disk bus bandwidth, and that the CPU is the bottleneck.  IfK so, throwing another CPU at the problem will help with overall throughput - J although it won't double because there's some amount of overhead in havingM multiple processors (I think you might get 1.7 or 1.8 times the throughput ofo* one processor with 2, as a rule of thumb).  J By "double the throughput" do they mean "do twice as much work in the sameI amount of time"?  Maybe, if that doesn't saturate the rest of the system. L If they mean "finish each user request in half the current time" just addingH another CPU won't do it _unless_ the jobs are slow because the system isK overloaded - a whole bunch of jobs sitting around in COM state all the timee waiting to be scheduled.   >uB >This is on OVMS 7.x, OSU server (which is not all that relevant). >M  M Actually, it being the OSU server is relevant because that server (also knowniL as the DECthreads server) supports kernel threads on VMS v.7, which means it0 could make effective use of multiple processors.  M (Of course, it's also really good at exposing DECthreads bugs, so I'm runningr> it with kernel threads disabled.  But we were talking theory.)   -- Alan      >-A1  O ===============================================================================n0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210 O ===============================================================================    ------------------------------  # Date: Mon, 15 Apr 2002 22:52:04 GMT L From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")+ Subject: Re: Throughput and additional CPUs 8 Message-ID: <00A0C81D.F3ABBE66@SSRL04.SLAC.STANFORD.EDU>  W In article <Xns91F1AB2537EFEanamika@207.181.101.12>, Anamika <anamika@home.com> writes: 5 >Dale Dellutri <ddelQQQlutr@entQQQeract.com> wrote in % >news:a9f548$45h$1@bob.news.rcn.net:   > L >There are no bottlenecks as such and that is not the problem. The question L >is, if a transaction (like accessing a file mapped via a URL), consists of I >tasks that are I/O bound, N/W bound and CPU bound, then adding a second  G >CPU to the system in question will not double the throughput. All the  K >other parts need to be beefed up as well ie., the N/W part, I/O part etc.,  >Is that right ?  G If there are no bottlenecks, then _nothing_ will double the throughput. M (The _potential_ throughput, sure, but the actual throughput?  By definition, M if there are no bottlenecks, it's already processing everything as fast as ito, can, and adding resources is just wasteful.)  J If the idea is that the load will increase in future, then there's still aJ question of what percentage of utilization each resource is experiencing; O if you're 10% network, 10% i/o, 75% CPU, you'll probably need more CPU to coverc= a doubled load; you probably won't need more network or disk.o  M Also, the particular application load may scale differently than you expect. nL How long does it take to translate a URL to a filename?  The server probablyN has the mapping rules cached in memory, rather than going out to disk all  theI time, so the translation process is all CPU.  If the distribution of URLsUM doesn't change - I mean, if you have twice as many people who are asking for _P the same stuff, rather than doubling the number of _different_ URLs requested - F then you could get twice as many requests satisfied from the webserverL memory cache rather than requiring disk access at all.  They might be betterO off resolving what seems to be a disk i/o problem by buying more RAM and makinghG the webserver cache bigger - but not if the files being served are veryb volatile.  i  O This kind of thing is extremely sensitive to the actual characteristics of yourrO load, and making decisions based on logic without measurement is likely to be al waste of time and money.   -- Alan   O ===============================================================================r0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056nM  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210wO ===============================================================================a   ------------------------------    Date: 15 Apr 2002 18:59:59 -0700! From: hemanir@yahoo.com (Anamika) + Subject: Re: Throughput and additional CPUsm= Message-ID: <5130f039.0204151759.7b0c70d2@posting.google.com>y  / Well, maybe I did not make myself clear enough.hE By "doubling" the throughput, what was also meant, was that, the samet. task would be done in half the amount of time.B Which in my estimation, cannot be done by just adding another CPU,( which will benefit only CPU bound tasks.   -A   winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") wrote in message news:<00A0C81D.F3ABBE66@SSRL04.SLAC.STANFORD.EDU>...Y > In article <Xns91F1AB2537EFEanamika@207.181.101.12>, Anamika <anamika@home.com> writes:P7 > >Dale Dellutri <ddelQQQlutr@entQQQeract.com> wrote ine' > >news:a9f548$45h$1@bob.news.rcn.net:   > > N > >There are no bottlenecks as such and that is not the problem. The question N > >is, if a transaction (like accessing a file mapped via a URL), consists of K > >tasks that are I/O bound, N/W bound and CPU bound, then adding a second eI > >CPU to the system in question will not double the throughput. All the  M > >other parts need to be beefed up as well ie., the N/W part, I/O part etc.,: > >Is that right ? > I > If there are no bottlenecks, then _nothing_ will double the throughput.eO > (The _potential_ throughput, sure, but the actual throughput?  By definition, O > if there are no bottlenecks, it's already processing everything as fast as it . > can, and adding resources is just wasteful.) > L > If the idea is that the load will increase in future, then there's still aL > question of what percentage of utilization each resource is experiencing; Q > if you're 10% network, 10% i/o, 75% CPU, you'll probably need more CPU to coverp? > a doubled load; you probably won't need more network or disk.t > O > Also, the particular application load may scale differently than you expect. hN > How long does it take to translate a URL to a filename?  The server probablyP > has the mapping rules cached in memory, rather than going out to disk all  theK > time, so the translation process is all CPU.  If the distribution of URLs O > doesn't change - I mean, if you have twice as many people who are asking for  R > the same stuff, rather than doubling the number of _different_ URLs requested - H > then you could get twice as many requests satisfied from the webserverN > memory cache rather than requiring disk access at all.  They might be betterQ > off resolving what seems to be a disk i/o problem by buying more RAM and makingiI > the webserver cache bigger - but not if the files being served are veryu
 > volatile.  y > Q > This kind of thing is extremely sensitive to the actual characteristics of yourdQ > load, and making decisions based on logic without measurement is likely to be ai > waste of time and money. > 	 > -- Alani > Q > =============================================================================== 2 >  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUO >  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056rO >  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210 Q > ===============================================================================    ------------------------------  # Date: Tue, 16 Apr 2002 02:17:39 GMTlL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")+ Subject: Re: Throughput and additional CPUs 8 Message-ID: <00A0C83A.AC44FDDB@SSRL04.SLAC.STANFORD.EDU>  a In article <5130f039.0204151759.7b0c70d2@posting.google.com>, hemanir@yahoo.com (Anamika) writes:h  0 >Well, maybe I did not make myself clear enough.F >By "doubling" the throughput, what was also meant, was that, the same/ >task would be done in half the amount of time.dC >Which in my estimation, cannot be done by just adding another CPU,e) >which will benefit only CPU bound tasks.   M And it won't even benefit CPU-bound tasks very much, unless most of the clocktK time is consumed by sitting in COM state waiting for CPU.  As other postersrM have remarked, the way to speed up a CPU-bound task is with a faster CPU, not $ with a second CPU of the same speed.  H (Although the way you get really dramatic speed improvements on the sameO hardware is typically through application redesign, informed, again,  by havingc= some idea of where the system is actually spending its time.)   N So, yes, you're right, a second CPU isn't going to what they say it's going toO do - halve the transaction time.  But neither would just throwing on more disks0M and another network adapter.  In fact, if you have no bottlenecks now, adding:F more of the same kinds of resources isn't going to buy you _anything_.  L (I say "same kinds of resources" because it might buy you something to add aM whole bunch of memory, configure it as a RAM disk, and serve files from that, + avoiding all rotational latency questions.)f   -- Alan   O ===============================================================================o0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056bM  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CA  94309-0210oO ===============================================================================n   ------------------------------  % Date: Mon, 15 Apr 2002 14:13:02 -0700m0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>7 Subject: Re: VMS "unhackable" example just for Andy ...o+ Message-ID: <3CBADFEE.5F2A826@Mvb.Saic.Com>U   JF Mezei wrote:s >  > Mark Berryman wrote:J > > Get it through *your* skull, buying a separate IP stack has never beenH > > an issue at *any* VMS site I have ever dealt with.  VMS is clearly a7 > > different world than the one you are familiar with.i > O > For existing customers, perhaps since existing customers have been trained too& > accept being screwed for more money.  B Nope, many were new customers.  I have software packages from manyD different vendors on my system.  The fact that one of them is the IPD stack makes no difference to me or, quite frankly, to much of anyone else I've encountered.  C If you feel differently, by all means stay with TCPIP Services.  At>F least on VMS I have multiple choices.  On Solaris, if the one from Sun. doesn't meet my needs, I have no other choice.   > O > But many other systems came with a TCPIP stack built-in at no extra cost. TheLK > problem is that Digital had set its sights on OSI and was blindsighted by 	 > TCPIP. a  E Digital didn't set its sights on OSI, several of its major customers,IH including the U.S. government, did.  Only after ISO turned OSI into suchE a complex hulking dinosaur did the expectations for it wane and folks F were left with TCP/IP as the only other real choice.  Had OSI lived upF to its premise it would now be the networking standard and we wouldn't be waiting on IPv6.e  F Hindsight is always great but I wouldn't put too much blame on DigitalE for listening to so many of its customers regarding the need for OSI.a  D Now, letting UCX languish for so long after OSI tanked, THAT you can definitely blame Digital for.e  E > In doing so, it allowed competitors such as Multinet to fill in thedP > gaps. And by the time DEC woke up, there were too many TCPIP stacks, each with# > their own programming interface. t  H They had their own private API it is true, but, except for the free one,A they also supported the UCX API.  A necessity in order to run anyi+ IP-based software provided by VMS's vendor.e  : > TCPIP V5 is about 8-10 years late in the strategic game.  1 Not quite that late, but very late none-the-less.r  O > By now, I think that things are better coordinated between stacks so it isn'tpD > so much of an issue anymore. It just took way to long to get here.  C None of the stacks are "co-ordinated".  The 3rd-party stack vendorspG (well, vendor, now) simply keep their stack's API up to date with TCPIPnC Service's.  I cannot speak to what sort of co-operation they get ineH maintaining this but I doubt there's ever been anything in the form of aF "head's up were about to make the following changes to TCPIP services" from Compaq.  
 Mark Berrymany   ------------------------------  % Date: Mon, 15 Apr 2002 16:14:01 -0500o+ From: Christopher Smith <csmith@amdocs.com>P7 Subject: RE: VMS "unhackable" example just for Andy ...dJ Message-ID: <7E008308CD77154485FEF878168D078E0178441D@CMIMAIL1.amdocs.com>   > -----Original Message-----9 > From: Mark Berryman [mailto:Mark.Berryman@Mvb.Saic.Com]/  H > Hindsight is always great but I wouldn't put too much blame on DigitalG > for listening to so many of its customers regarding the need for OSI.   @ That is also aside from the fact that OSI is a good concept ;)  * Implementations may be lacking, of course.   Chriss    ! Christopher Smith, Perl Developers Amdocs - Champaign, IL   /usr/bin/perl -e '? print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");  'e      ------------------------------  % Date: Mon, 15 Apr 2002 14:00:22 -0500i. From: Catherine Salvendy <csalvend@purdue.edu>* Subject: Re: VMS 7.2-1 to 7.2.2- Follow Up* Message-ID: <3CBB2346.4ED5A685@purdue.edu>  D I upgraded from 7.2-1 to 7.2-2 and have a forrtl kit and did not get1 this error. The upgrade went off without a hitch.l   Cathy    Rob Buxton wrote:s > 	 > Hi All,a > A > Here's a link to a follow up that identifies the issue Dirk hase	 > raised.o> > Important for anyone considering the 7.2-1 to 7.2-2 Upgrade. > # > http://ftp.support.compaq.com.au/h >  > Rob.   ------------------------------  % Date: Mon, 15 Apr 2002 15:39:38 -0400a  From: norm.raphael@jamesbury.com* Subject: Re: VMS 7.2-1 to 7.2.2- Follow Up4 Message-ID: <C2256B9C.006AFBAF.00@jklh22.valmet.com>   This section is undated.B It may not be true after upgrading to Fortran V7.5-1 with its RTL.+ It may have been true for earlier versions.  Just guessing.        - csalvend@purdue.edu on 04/15/2002 03:00:22 PMh  % Please respond to csalvend@purdue.edud   To:   Info-VAX@mvb.saic.comw cc:u+ Subject:  Re: VMS 7.2-1 to 7.2.2- Follow UpW      D I upgraded from 7.2-1 to 7.2-2 and have a forrtl kit and did not get1 this error. The upgrade went off without a hitch.    Cathyt   Rob Buxton wrote:t > 	 > Hi All,p >lA > Here's a link to a follow up that identifies the issue Dirk hasp	 > raised.L> > Important for anyone considering the 7.2-1 to 7.2-2 Upgrade. >,# > http://ftp.support.compaq.com.au/s >e > Rob.   ------------------------------    Date: 16 Apr 2002 00:26:11 -0400+ From: <Make_Money_From_Home_@attglobal.net>a" Subject: we GUARANTEE you a check!" Message-ID: <5068881@MVB.SAIC.COM>  > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD>F <META http-equiv=Content-Type content="text/html; charset=iso-8859-1">5 <META content="MSHTML 6.00.2713.1100" name=GENERATOR>. <STYLE></STYLE>> </HEAD>t <BODY bgColor=#ffffff>" <DIV><FONT face="ACaslon Regular"> <DIV><FONT face=Arial size=2> O <P align=center><B>You are receiving this because we have exchanged e-mails in sP the past and your e-mail address is on file. If you wish to be removed, removal E instructions are at the end of this message.<BR></B><BR><FONT lang=0 fB style="BACKGROUND-COLOR: #ffffff" face=Arial color=#ff0000 size=3 < FAMILY="SANSSERIF"><B>Yes, that's right! We guarantee you a K check!<BR></FONT><FONT lang=0 style="BACKGROUND-COLOR: #ffffff" face=Arial  ? color=#000000 size=2 FAMILY="SANSSERIF"></B><BR>Hello, this is tN Steve&nbsp;in&nbsp;Georgia, USA. <BR>I hope this email finds you and yours in D good health.<BR>I have something for you to read that will interest < you!<BR>Thank You for your time.<BR><BR></FONT><FONT lang=0 B style="BACKGROUND-COLOR: #ffffff" face=Arial color=#ff0000 size=3 J FAMILY="SANSSERIF"><B>Answer four simple questions and we guarantee you a G check!</FONT><FONT lang=0 style="BACKGROUND-COLOR: #ffffff" face=Arial s: color=#ff0000 FAMILY="SANSSERIF"></B> </FONT><FONT lang=0 B style="BACKGROUND-COLOR: #ffffff" face=Arial color=#000000 size=2  FAMILY="SANSSERIF"><BR><A C href="http://www.netmailsolutions.com/internetmarketing101/"><FONT eT color=#0000ff>http://www.netmailsolutions.com/internetmarketing101/</FONT></A><FONT 1 color=#0000ff size=3> </FONT></FONT><FONT lang=0 uB style="BACKGROUND-COLOR: #ffffff" face=Arial color=#000000 size=2 K FAMILY="SANSSERIF"><FONT color=#0000ff size=3><FONT face="ACaslon Regular" "N color=#000000>(private invitation only)</FONT><BR></FONT><BR><BR>A successful Q friend of mine called to inform me that a group called Internet Marketing101 had hH just launched an exclusive recruiting software that would revolutionize P networking as we now know it.&nbsp; He asked me to simply go to the site answer N the four questions and told me that it would cost me <B>zero </B>to do it AND M they guaranteed me a check. I am glad I trusted him because I would normally  N ignore a pitch like that cause its sounded to good to be true. I appeased him P and answered the four questions at the bottom of the site.<BR><BR>I am grateful Q I did! I now have over 120 folks in my powerline in less than a week and growing =N like crazy. They were recruited and placed there by this new software and the J team. It blew my mind and I've already made money without doing anything. L Putting my info and taking the tour was all it took, see for yourself. Your K PowerLine will grow like MAD! You can bet I will learn even more about the =L Internet Marketing 101. This incredible software has only been in use for 8 P weeks!<BR><BR><B>This is by far the best system I have seen in 12 years in this Q business!<BR>Now even those who have never made money before ARE!</B><BR><BR>Now DP that I've done my part by alerting you to this awesome opportunity, the rest is O up to you. You owe it to yourself to go answer the four questions and see what lD happens REGARDLESS OF WHAT PROGRAM YOU'RE PRESENTLY WORKING.<BR></P> <P align=left><BR>K <P align=center><BR>NOTE: The first in gets the most spill in a Powerline. nN Powerline means one under the other down one line. So if 100 people join from P this email and you are first you will have 99 people in your powerline, if your J the 50th you will have 49,if your the 90th you will have 10. So on and so I on!<BR><BR>So needless to say <B>go fast, take the tour,</B>answer the 4 fN questions, get your free spot in the powerline and let this system go to work M for you like it is has been doing for all of us!<BR><B><BR>FOLKS THIS SYSTEM eQ WORKS LIKE A CHARM AND COSTS NOTHING SO PUT IT TO THE TEST!<BR></B><BR><B>TRY IT dL AND YOU WILL BE AMAZED! I SURE AM!</B><BR><BR>Regards,<BR><BR>Steve <BR>Ga, L USA<BR><BR></FONT><FONT lang=0 style="BACKGROUND-COLOR: #ffffff" face=Arial O color=#ff0000 size=3 FAMILY="SANSSERIF"><B>Answer four simple questions and we rM guarantee you a check! </FONT><FONT lang=0 style="BACKGROUND-COLOR: #ffffff" t> face=Arial color=#000000 size=2 FAMILY="SANSSERIF"></B><BR><A C href="http://www.netmailsolutions.com/internetmarketing101/"><FONT aT color=#0000ff>http://www.netmailsolutions.com/internetmarketing101/</FONT></A><FONT L color=#0000ff size=3> </FONT><FONT lang=0 style="BACKGROUND-COLOR: #ffffff" G face=Arial color=#000000 size=2 FAMILY="SANSSERIF"><FONT color=#0000ff iF size=3><FONT face="ACaslon Regular" color=#000000>(private invitation I only)</FONT><BR></FONT></FONT>&nbsp;<BR><BR><STRONG>&nbsp;&nbsp;Internet a/ Marketing 101 </STRONG><BR><BR><BR><BR><BR><BR>oC <P align=left><FONT face="ACaslon Regular" size=3></FONT>&nbsp;</P>t: <P align=left>Steve<BR>Ga, USA<BR><BR></FONT><FONT lang=0 ; style="BACKGROUND-COLOR: #ffffff" face=Arial color=#000000 tM FAMILY="SANSSERIF"><B><FONT size=1>Removal &nbsp;<BR>This message is sent in rM compliance of the new email<BR>bill section 301. Under Bill S.1618 TITLE III oP passed<BR>by the 105th U.S. Congress. This message cannot be<BR>Considered Spam O as long as we include the way to be<BR>removed and the senders info to be true =O and not falsified.<BR>Paragraph (a)(c) of S.1618, further transmissions to you sL by<BR>the sender of this email may be stopped at no cost to you&nbsp;<BR>by J sending a blank email with "removeplease"&nbsp;in the subject <BR>line to 6 </FONT><A href="mailto:removeplease@excite.com"><FONT K size=1>removeplease@excite.com</FONT></A><FONT size=1>&nbsp;. We honor all 9M remove&nbsp;</FONT></B></FONT><FONT lang=0 style="BACKGROUND-COLOR: #ffffff" nQ face=Arial color=#000000 FAMILY="SANSSERIF"><FONT size=1><BR><STRONG>requests at aM once. <BR></STRONG></FONT></P></FONT></FONT></DIV></FONT></DIV></BODY></HTML>s   ------------------------------    Date: 15 Apr 2002 13:19:04 -0500- From: koehler@encompasserve.org (Bob Koehler)o1 Subject: Re: Why VMS is "unhackable" lesson 1 ...o3 Message-ID: <UVz71v8IAU5p@eisner.encompasserve.org>O  } In article <3CBAF242.4050908@sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.remove_this.harrison@sun.com> writes:e > : > Its a bug in both the stack and the ping program neither6 > should allow the creation of an illegally large ICMP	 > packet.e  D    Doens't matter whether the ping program is right or not.  Anybody>    can program to the RFC and get it wrong.  A Software always    assume worst case bad input.s   ------------------------------   End of INFO-VAX 2002.209 ************************