1 INFO-VAX	Tue, 03 Sep 2002	Volume 2002 : Issue 486       Contents: RE: "inview" Article% Re: Copy of OpenVMS hobbyist CD in UK % Re: Copy of OpenVMS hobbyist CD in UK 
 DEC-adence Re: Escape Sequences Re: HELP RMU. RE: Is the HP/Compaq merger really going well? Re: Itanic2 performance " Re: OpenVMS and Hobbyist CDROM Kit OpenVMS documentation  RE: OpenVMS documentation $ OpenVMS v7.3 - setting the time zone Re: Please Sum it up.  problems with PIPE Re: problems with PIPE Re: Remote Site Cluster Member9 Re: Secret New Products (was: Is the HP/Compaq merger...) 9 Re: Secret New Products (was: Is the HP/Compaq merger...) A Re: Sick of the trolls, whiners, and non-customers in this forum? " Re: silliest thing heard this week" Re: silliest thing heard this week Re: Strange F$TRNLNM behaviour?   F ----------------------------------------------------------------------  $ Date: Tue, 3 Sep 2002 07:23:43 -0400' From: "Main, Kerry" <Kerry.Main@hp.com>  Subject: RE: "inview" Article T Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF402660997@kaoexc01.americas.cpqcorp.net>   JF,   5 >>> But how does one counter a Unix argument such as: C "Our version of clustering uses the same paradigm as the best known E fault tolerant system: Tandem, and VMS doesn't use that paradigm".<<<   H As stated before, there are some advantages to a shared nothing cluster.F There are no DLM concerns, but with recent HW and OS enhancements, one$ could also argue these are minor.=20  6 However, you also have to deal with other issues like:  C 1. Every system is allocated to "serve" specific workloads to other E servers - not unlike MSCP serving in OpenVMS clustering. Adding a new A system typically requires re-partitioning the database workload -   typically results in outages.=20  H 2. One can not load balance the same workload across multiple servers inF multiple sites as shared nothing approaches typically assume some formF of "replication" between sites which means a master-slave (one writer,D many readers) arrangement i.e. by design, one system handles all theH load associated with specific workload components. If Server A at Site AG requires some data that Server B at Site B is specifically allocated to G handle, then Server A in Site A must access Server B in Site B for that ? information. Server A can not access that information directly.   F Fwiw, Oracle has typically been associated with UNIX vendors in recentD years and they have adopted the shared everything approach i.e. theyG have licensed with their premier 9i RAC database and future versions of F their product a subset of the Tru64 UNIX cluster code, which, in turn,( is a subset of the OpenVMS cluster code.  G In addition, Oracle (Rdb, 8i/9i) also now supports using OpenVMS Galaxy H and shared physical memory as cluster interconnects on partitioned Alpha GS series servers.  H [sidebar - IBM DB2 on mainframe uses shared everything approach as well,; and they do understand cluster and database scaling issues)   7 Here is an article which presents a bit of an overview: H http://www.infoworld.com/articles/fe/xml/01/12/17/011217feclustertca.xmlC "Clustering should be viewed as a necessary evil for OLTP. Consider E shared-disk clusters for most scenarios; shared- nothing clusters are & best-suited to large data warehouses."  E Which, among other things, is why NSK is great for ZLE type workloads , i.e.. it's a big warehouse type application.   Regards   
 Kerry Main Senior Consultant  Hewlett-Packard Canada! Consulting & Integration Services  Voice: 613-592-4660  Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----7 From: JF Mezei [mailto:jfmezei.spamnot@videotron.ca]=20   Sent: September 2, 2002 10:47 PM To: Info-VAX@Mvb.Saic.Com  Subject: Re: "inview" Article      "Main, Kerry" wrote:E > One might argue that NSK is the gold standard for shared nothing=20 F > clusters while OpenVMS is the gold standard for shared everything=20 > clusters.   E It is often said here that VMS' clustering is superior is part due to  the shared everything paradigm.   1 But how does one counter a Unix argument such as: C "Our version of clustering uses the same paradigm as the best known B fault tolerant system: Tandem, and VMS doesn't use that paradigm".   ------------------------------   Date: 3 Sep 2002 05:06:13 -0700 ) From: christian@karg.org (Christian Karg) . Subject: Re: Copy of OpenVMS hobbyist CD in UK< Message-ID: <7b342c68.0209030406.1ae4ae0@posting.google.com>   Yes - I get up to 2Mbit    ;-)  D Could you email me the details? I have no experience burning VAX/VMSD CD's, but I'll give it a try. Are you aware of any special things toC watch out for? My setup includes a PC running WindowsXP with a CDRW  drive.   Thanks, 	 Christian   h issinoho@slayme.com (issinoho) wrote in message news:<d0141774.0209020458.7a15a48@posting.google.com>... > Christian, > T > Have you broadband? If so, I can put up an image for you to grab on my FTP server. >    ------------------------------  % Date: Tue, 03 Sep 2002 17:26:24 +0400 1 From: "Konstantin Klubnichkin" <kostik@beenet.ru> . Subject: Re: Copy of OpenVMS hobbyist CD in UK6 Message-ID: <pan.2002.09.03.13.26.22.781463@beenet.ru>  3 On Mon, 02 Sep 2002 05:58:57 +0000, issinoho wrote:    > Christian, > T > Have you broadband? If so, I can put up an image for you to grab on my FTP server. > 	 <skipped>   @ May I have it too? Joined Encompass too, having same problems asI Christian. Furthermore delivery to Russia takes too much time and makes a  lot of headache...   Thanks!   
 Best regards,  Konstantin Klubnichkin# P.S. Your e-mail box is unavailable    ------------------------------  * Date: Tue, 3 Sep 2002 06:39:28 -0700 (PDT). From: Fabio Cardoso <fabiopenvms@yahoo.com.br> Subject: DEC-adence @ Message-ID: <20020903133928.84548.qmail@web20205.mail.yahoo.com>   Do you know this site ???    Click   / http://telnet.hu/hamster/decadence/e_index.html    =====  ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazil  fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?+ Yahoo! Finance - Get real-time stock quotes  http://finance.yahoo.com   ------------------------------  % Date: Tue, 03 Sep 2002 09:31:24 +0100  From: Roy Omond <Roy@Omond.net>  Subject: Re: Escape Sequences ( Message-ID: <3D74735C.9A28586@Omond.net>   Paul Repacholi wrote:   # > Roy Omond <Roy@Omond.net> writes:  > G > > Who can remember the effects of setting process name to ^E (control F > > E) when a user did $ show system on a VT100 ?  :-) A long time ago > > ...  > : > Before or after you `improved' his answerback string? :)  E Hmm... maybe my memory is fading (Gartner probability factor: 1.0 :-) D I thought the effect of the ^E was to wedge the VT100 so that only a power 9 off could unwedge it.  What was it then if it wasn't ^E ?   ; There was another one to set the VT100 into self-test mode. + Paul Williams (of vt100.net), what was it ?   	 Roy Omond  Blue Bubble Ltd.   ------------------------------  + Date: Tue, 03 Sep 2002 15:23:02 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>  Subject: Re: HELP RMU ; Message-ID: <01KM2HFH1QCW9QUS7H@sysdev.deutsche-boerse.com>   J > We do use the VMS HELP format (always did and will always will).  We did( > use HTML also on those other systems.   H In case it's not clear to all, what I meant is that normally HELP files G are structured so that among the "additional information available" it  G is obvious which are qualifiers etc and also there is more consistency  E between the use of / to separate tokens in a command and within HELP  @ (making cut and paste easier etc).  (While I'm on the topic, my F favourite HELP is for Fortran where not only are all options given in % the overview, but also the defaults.)    ------------------------------  $ Date: Tue, 3 Sep 2002 10:05:12 -0400; From: "Webb, William W Raleigh, NC" <wwebb1@email.usps.gov> 7 Subject: RE: Is the HP/Compaq merger really going well? K Message-ID: <BA52530E3149734A9BAABDBBFA808E49028BEA60@rlghncst964.usps.gov>    Sorry, I rounded.  What I meant to say was:   AM 00:00:01 => 11:59:59  PM 12:00:01 => 23:59:59   , Here's what the US govt says on the issue of noon and midnight:  5 http://www.boulder.nist.gov/timefreq/general/misc.htm    WWWebb  L ---------------------------------------------------------------------------- ----   # Frequently asked questions (FAQ)      L ---------------------------------------------------------------------------- ----   <snip>    ) Are noon and midnight 12 a.m. or 12 p.m.?   L This is a tricky question. The answer is that the terms 12 a.m. and 12 p.m. ! are wrong and should not be used.   I To illustrate this, consider that "a.m" and "p.m." are abbreviations for  G "ante meridiem" and "post meridiem." They mean "before noon" and "after  noon,"  G respectively. Noon is neither before or after noon; it is simply noon.  A Therefore, neither the "a.m." nor "p.m." designation is correct.  L On the other hand, midnight is both 12 hours before noon and 12 hours after K noon. Therefore, either 12 a.m. or 12 p.m. could work as a designation for  > midnight, but both would be ambiguous as to the date intended.  H When a specific date is important, and when we can use a 24-hour clock, L we prefer to designate that moment not as 1200 midnight, but rather as 0000 L if we are referring to the beginning of a given day (or date), or 2400 if we  1 are designating the end of a given day (or date).   I To be certain of avoiding ambiguity (while still using a 12-hour clock),  F specify an event as beginning at 1201 a.m. or ending at 1159 p.m., for	 example;  G this method is used by the railroads and airlines for schedules, and is  often @ found on legal papers such as contracts and insurance policies.   J If one is referring not to a specific date, but rather to several days, or days  J in general, use the terms noon and midnight instead of 12 a.m. and 12 p.m. For L example, a bank might be open on Saturdays from 8 a.m. to noon. Or a grocery  E store might be open daily until midnight. The terms "12 noon" and "12 
 midnight" $ are also correct, though redundant.    -----Original Message-----" From: "AG" [mailto:ang@xtra.co.nz]( Sent: Saturday, August 31, 2002 10:51 PM To: "Info-VAX@Mvb.Saic.Com" 7 Subject: RE: Is the HP/Compaq merger really going well?     5 "Bill Gunshannon" <bill@cs.uofs.edu> wrote in message 0 news:akll51$1k5o0m$1@ID-135708.news.dfncis.de... > In article@ <BA52530E3149734A9BAABDBBFA808E49028BEA4F@rlghncst964.usps.gov>,? > "Webb, William W Raleigh, NC" <wwebb1@email.usps.gov> writes:  > >  > > Bob is correct.  > > E > > AM = Ante Meridian, which starts at 12:01 AM and ends at 11:59 AM E > > PM = Post Meridian, which starts at 12:01 PM and ends at 11:59 PM  > >   J Just out of curiosity - what do you call the time from 11:59pm to 12:01am?D More curious even, your definition above seems to fly in the face ofH conventional math - you have just said that something can start at 12.01+ and end at 11.59 (using the same scale too)    ------------------------------  % Date: Tue, 03 Sep 2002 11:28:18 +0200 E From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>   Subject: Re: Itanic2 performance+ Message-ID: <3D7480B2.6748B852@mediasec.de>    > And the new 1.25 GHz AlphaG > released last month rates 1327 SPECfp2K_peak (no base figure given),    I What is the source of this information? AFAIK, this is not allowed by the J SPEC CPU2000 run rules: you can quote base without peak, or claim the baseM value as the peak value, but you're not allowed to quote a peak value without  its base value.    	Jan   ------------------------------  * Date: Tue, 3 Sep 2002 10:02:35 +0000 (UTC)% From: John Forkosh <john@invalid.com> + Subject: Re: OpenVMS and Hobbyist CDROM Kit , Message-ID: <al21br$n3i$2@reader1.panix.com>  $ NetBoot <netboot@netboot.com> wrote:E : How do I get aboard?  Can someone post a link so I can get started?  : Thanks  0 http://www.encompassusa.org/membership/join.html2      the "associate" class of memebership is free;0      they'll email you your memeber#, putatively,      in a few days, which you'll need for...    http://www.montagar.com/hobbyist1      ...the links from this page allow you to get 0      vms licenses (you need both vms (for vax or*      alpha) and layered product licenses),0      and allow you to purchase a vms hobbyist cd1      (for vax or alpha, though vax is temporarily       sold out).   / Other useful info and links available from both ! the above, particularly montagar.  --  > John Forkosh  ( mailto:  j@f.com  where j=john and f=forkosh )   ------------------------------  . Date: Tue, 3 Sep 2002 08:47:21 +0200 (MET DST)& From: Rudolf Wingert <win@fom.fgan.de> Subject: OpenVMS documentation6 Message-ID: <200209030647.IAA17578@sinet1.fom.fgan.de>   Hello,  H yesterday I did get the new one OpenVMS documentation for OpenVMS 7.3-1.B But since years I do miss the OpenVMS System Messages and RecoveryG Procedures Reference Manual. The last one I did get is for OpenVMS 6.0. C Since this version a lot of messages did change, or are new and not H explained within this manual. When will we get a well documented OpenVMSG again? Or will OpenVMS go the way as UNIX or Windows (every new version  less documantations).    TIA and regards Rudolf Wingert   ------------------------------  $ Date: Tue, 3 Sep 2002 11:31:16 -0400' From: "Main, Kerry" <Kerry.Main@hp.com> " Subject: RE: OpenVMS documentationT Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4023D956C@kaoexc01.americas.cpqcorp.net>   Rudolf,   G Not sure what documentation kits you receive, but is the following what  you were looking for?   ; http://www.openvms.compaq.com/doc/73final/6023/6023pro.html A "OpenVMS System Messages: Companion Guide for Help Message Users"   + Software Version: OpenVMS Alpha Version 7.3  OpenVMS VAX Version 7.3=20 Order Number: AA--PV5TD--TK    Regards   
 Kerry Main Senior Consultant  Hewlett-Packard Canada! Consulting & Integration Services  Voice: 613-592-4660  Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----0 From: Rudolf Wingert [mailto:win@fom.fgan.de]=20 Sent: September 3, 2002 2:47 AM  To: Info-VAX@Mvb.Saic.Com  Subject: OpenVMS documentation     Hello,  H yesterday I did get the new one OpenVMS documentation for OpenVMS 7.3-1.B But since years I do miss the OpenVMS System Messages and RecoveryG Procedures Reference Manual. The last one I did get is for OpenVMS 6.0. C Since this version a lot of messages did change, or are new and not H explained within this manual. When will we get a well documented OpenVMSG again? Or will OpenVMS go the way as UNIX or Windows (every new version  less documantations).    TIA and regards Rudolf Wingert   ------------------------------   Date: 3 Sep 2002 02:38:56 -0700 6 From: andrew.rycroft@intrinsitech.com (Andrew Rycroft)- Subject: OpenVMS v7.3 - setting the time zone = Message-ID: <58ba0101.0209030138.40258486@posting.google.com>    Hi,   D I am trying to set my time zone on OpenVMS v7.3. I am in a time zone: GMT+2 with no daylight saving - Johannesburg South Africa.  E When I try to set the timezone to GMTPLUS2 with NET$CONFIGURE, I get    4 DTSS$CONFIG-I-LOGS Defining system timezone logicals6 %DTSS-E-INVRULE, Invalid SYS$TIMEZONE_RULE: GMT+0200-2@ %DTSS-E-FAILED, error occured while trying to complete directive  B I can set the other times zones - Egypt, Israel, MET, but they all have daylight savings time.   E Any ideas - I have checked the FAQ, and it seems to be a bit obscure.    Thanks Andrew   ------------------------------   Date: 3 Sep 2002 07:07:47 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: Please Sum it up.3 Message-ID: <fD6FubdMGrKG@eisner.encompasserve.org>   U In article <un98rutfr4kscb@corp.supernews.com>, "Vivek Soni" <visoni@bmc.com> writes: ' > Whats the sum of all the researches ?    42   ------------------------------  . Date: Tue, 3 Sep 2002 08:41:54 +0200 (MET DST)& From: Rudolf Wingert <win@fom.fgan.de> Subject: problems with PIPE 6 Message-ID: <200209030641.IAA17568@sinet1.fom.fgan.de>   Hello,  H I would like to get the DECnet name from a DECnet address or the IP nameD from the IP address. To do so, I did use the following DCL commands:   	$ name = "" 	$ line = "" 	$ IF adr_typ .EQS. "DECnet" 	$ THENaB 	$  PIPE MCR NCP SHOW NODE 'adr' 2> NL: | SEARCH SYS$INPUT 'adr' - 		| READ SYS$INPUT liner= 	$  name = F$ELEMENT (1, "(", F$ELEMENT (0, ")", line)) - "("o 	$ ELSEp> 	$  PIPE UCX SHOW HOST 'adr' 2> NL: | SEARCH SYS$INPUT 'adr' - 		| READ SYS$INPUT linee$ 	$  line = F$EDIT (line, "COMPRESS")= 	$  name = F$ELEMENT (1, " ", F$ELEMENT (0, ",", line)) - " "s 	$ ENDIF+ 	$ IF name .NES. "" THEN		$ GOTO name_founds  D Something is wrong within the PIPE command. The symbol line is emptyF regardless the search did found a line. If the output is more then oneG line, all lines exclude the first one are displayed. Do anybody have an-G idea, how I can get the name without using a file. The problem is, that,A the there are some cases, where I could not produce a file (write C protect, no priviledge), or that are more then one tmp files createB at one time.   TIA and regards Rudolf Wingert   ------------------------------  $ Date: Tue, 3 Sep 2002 10:17:14 +0200, From: "Rainer Giese" <waste.not@welcome.net> Subject: Re: problems with PIPEs6 Message-ID: <al1r6b$1mdb8a$1@ID-138444.news.dfncis.de>  9 "Jan-Erik Sderholm" <aaa@aaa.com> schrieb im Newsbeitragh! news:3D746AF0.EBF27464@aaa.com...e; > The different "parts" of the PIPE command are executed ina9 > subprocesses. Any symbols are lost when each subprocesswB > exits. The line symbol at line 2 and the line symbol in the PIPE( > command are not the *same* line symbol  H Therefore I use Job-Logicals to submit values from a pipe to the callingL procedure.(define/job in the pipe, deassign/job after getting the value with
 f$trnlnm).   Rainer Giese   ------------------------------   Date: 3 Sep 2002 08:23:58 -0700a1 From: keithparris_NOSPAM@yahoo.com (Keith Parris)o' Subject: Re: Remote Site Cluster Memberm= Message-ID: <cf15391e.0209030723.6862408c@posting.google.com>.  s steve.cage@lombardodier.com (Steve Cage) wrote in message news:<f3316682.0209020704.4aedc6a1@posting.google.com>...rD > We currently have two Compaq DS20E's which are clustered utilizingF > twin shared SCSI buses and a Quorum disk. Expected Votes is set to 3% > with each node contributing 1 vote.a  ? You'll have to decide how you want to set the votes for all thetE systems.  Since it is intended that the remote site could continue oneA its own in the event of a failure of the main site, I'd recommendAF setting things up with equal votes for each of the two sites, and thenA either planning to take manual action with DECamds / AvailabilitytB Manager's "quorum fix" capability to allow one of the two sites toE continue in the event of a site outage or an inter-site link failure,nF or else put a system with a tie-breaking vote at a separate 3rd site. A Various possible vote schemes for disaster-tolerant clusters, andcF their pros and cons, are covered in my DECUS seminar on 'VMS Clusters:; Advanced Concepts' at http://www.geocities.com/keithparris/e  I > The system disk is shared and shadowed utilizing the shared SCSI buses.s  D Because your inter-site interconnect is LAN-based, the system at theA remote site will eventually need its own system disk (rather than"C continuing to boot as a satellite node), so that it can continue too run if you lose the first site.t  E > This weekend we are going to add another DS20E to the cluster whichaD > will have its disk drives (same capacity)  located in its internalC > drive cage. This machine will be located at our Disaster RecoveryeB > site, which is approx 1 mile away and connected via a 100MB Full, > Duplex Ethernet connection (Colt Lanlink). > G > 	For the test we will use cluster_config.com to create a new root on = > the shared system disk and add the new node to the cluster.g5 > 	Boot the new node into the cluster over Ethernet. F > 	Add a disk on the new DS20E to the System Disk Shadow Set. Then weA > should be able to boot the new machine of the local hard drive.m  C While the sequence of events you have described up to this point isoF fine, in practice you won't be able to boot from this remote disk as aE member of the system disk shadowset.  This is because once the remote-F DS20E shuts down, its copy of the system disk will become inaccessibleD and be removed from the shadowset by the systems at the first site. B When it tries to reboot, when it gets far enough along in the bootD sequence to join the cluster and discovers that it has booted from aF copy of the system disk that is out-of-date and not currently a memberD of the shadowset that is already mounted in the cluster, it gives upE with a Shadowing Detected Inonsistent State bugcheck.  (Maybe, if VMSs@ were smart enough, at some point, perhaps where it re-mounts theC system disk as read/write, to instead mount the shadowset using the D MSCP-served up-to-date member(s), it might be able to continue.  But it doesn't do that now.)  D To be able to boot from the local disk, the local disk would need toC be kept as a member of the shadowset across a reboot.  If it were aiD disk on Fibre Channel, and you had an inter-site Fibre Channel link,? then the remote copy of the disk could be kept as an up-to-date F shadowset member across a reboot of the remote node.  (Although it canE be done, I actually don't currently recommend that people really bootoA from a common FC system disk across sites, since I see the systemiE disk's contents as a potential single point of failure for both sitese in that case.)  @ In your case, what I would do is rather than a shadow copy, do aF BACKUP/IMAGE operation to the local disk, but then change the label on: the disk and reboot from it as an independent system disk.  ? Because you will have two system disks, you'll want to create aoE separate shadowset and put all the cluster-common stuff on that, liketF SYSUAF, RIGHTSLIST, job controller queue file, etc. and define logicalE names to point to those files.  I typically mount this cluster-commona shadowset in SYLOGICALS.COM.  2 > 	Remove the Quorum disk with cluster_config.com/ > 	Shadow remaining data disks to remote site.n  
 This is fine.e  H > The whole point of adding the third member is to keep a real time copyH > of data on a running system at a remote site in case disaster strikes.  @ Very common strategy these days, and easy to accomplish with VMS	 Clusters.a  G > My main question is how could we adjust expected_votes at remote siteaE > if we lost main site and had to run on a single node (assuming linkwG > was shutdown to avoid partitioning the cluster). I have tried this in 7 > the past and the single node just dies when you boot.g  B When the opposite site dies (or the inter-site link fails), if youB have set the two sites up with equal votes, then the systems pauseE (with all processes in RWCAP state, waiting for the Quorum capabilityt? bit).  While in this state, it is possible to manually tell theo4 system(s) at one site to regain quorum and continue.  F While the console IPL 12 (interrupt priority level "C" in hexadecimal)D software interrupt routine can (in theory) be used to adjust quorum,C since that doesn't work reliably on SMP systems due to some timers,2D it's seldom actually used in practice for disaster-tolerant clustersD anymore -- folks use DECamds or Availability Manager, select a node,C and use the quorum adjustment option under the "Fix" pull-down menu + item to adjust quorum quickly and reliably.s  A If for some reason you choose to reboot the system(s) at one sitewF instead of dynamically recovering quorum, you can use a conversational= bootstrap and at the SYSBOOT> prompt, set EXPECTED_VOTES to amE low-enough value to allow you to boot that site's system(s) alone.  ItF also recommend that you SET WRITESYSPARAMS 0 just before you CONTINUE,A so that this temporary parameter change doesn't become permanent. . ----------------------------------------------. Keith Parris | parris at encompasserve dot org   ------------------------------  % Date: Tue, 03 Sep 2002 11:02:48 +0200BE From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>,B Subject: Re: Secret New Products (was: Is the HP/Compaq merger...)+ Message-ID: <3D747AB8.9BBD12E7@mediasec.de>S  K > All of the above is true. You can still buy a VAX vacuum cleaner. Back iniM > them days, folks could work things out without benefit of lawyers and such.-+ > I guess what we have today is "Progress."   K I'm sure lawyers were involved in that agreement. This is standard fare for-M the trade/registered mark business. I agreed to such a deal just a month ago.    	Jan   ------------------------------   Date: 3 Sep 2002 07:28:42 -0600m- From: koehler@encompasserve.org (Bob Koehler)tB Subject: Re: Secret New Products (was: Is the HP/Compaq merger...)3 Message-ID: <BfCGrn3bmKOS@eisner.encompasserve.org>i  ` In article <dda6nuofqj8vh4m6ki0h7nvm03vtgslhb3@4ax.com>, Alan Greig <a.greig@virgin.net> writes: > G > Lawyers for DEC and the vacuum cleaner manufacturer agreed long, longnH > ago that they were happy as long as DEC never made vacuum cleaners and > VAX never made computers.r >   E    That could be a problem.  Starting with the 11/780, tranditionally $    VAXen have sucked from the floor.   ------------------------------  ! Date: Tue, 03 Sep 02 09:56:01 GMTe From: jmfbahciv@aol.comeJ Subject: Re: Sick of the trolls, whiners, and non-customers in this forum?+ Message-ID: <al255u$ekh$4@bob.news.rcn.net>a  = In article <eVLc9.90780$kp.733167@rwcrnsc52.ops.asp.att.net>,d5    "Terry C. Shannon" <terryshannon@attbi.com> wrote:w >eJ ><jmfbahciv@aol.com> wrote in message news:akvi6j$g9$3@bob.news.rcn.net...0 >> In article <akt9he$9o7@web.eng.baileynm.com>,- >>    peter@abbnm.com (Peter da Silva) wrote:c. >> >In article <3D718EC3.C1579C10@sunset.net>,+ >> >Tom Crabtre  <tccrab@sunset.net> wrote: J >> >> Actually, I'm sick of people who just complain how Compaq and now HP
 >> screwed >> >> them.hD >> >> If DEC had such a good thing, Bill Gates would have bought it. >> > >> >Didn't he? >nJ >I make a Small Joke. To all intents and purposes he seized control of the" >firm's strategy and product plas!  9 He paid for those if I understand the funding part of it.-   >i >> >7 >> Nope.  That would have been an honest business move.a >> He did something sneakier.M >o >Would you expct less? ;-}  : Why go to all the extra effort of doing sneaky things when@ you can cut all the bullshit by doing them in a straight-forward manner?1 >- >7F >PS-- Happy Labor Day to our Yank group members. At least this is less7 >controversial Labor Day than last ear. So far, anyhow!   ? Huh.  Then you haven't been listening to the news in Yank land. % What controversy are you thinking of?s   /BAH  ' Subtract a hundred and four for e-mail.e   ------------------------------  % Date: Tue, 03 Sep 2002 10:49:41 +0200 E From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>r+ Subject: Re: silliest thing heard this week + Message-ID: <3D7477A5.5BF54859@mediasec.de>j  J I would assume that for performance reasons, F$PARSE is in such situationsN reading the the directory (possibly cached) and not performing a header lookupJ when doing the parsing. This saves a lot of I/O in situations where you'reK not going to open the files/filenames that result - e.g., doing a directory.J listing or a wildcard lookup. F$PARSE just assumes that nobody in his saneL mind will create a file names .DIR;1 and not mean it to be a directory file.M I do think that having a file with extension .DIR and a version number higherwH than 1 will hide the directory both from RMS's parsing and from opening.   	Jan   ------------------------------   Date: 3 Sep 2002 08:02:07 -0700-. From: SPAMSINK2001@YAHOO.COM (Alan E. Feldman)+ Subject: Re: silliest thing heard this weekv= Message-ID: <343f30ae.0209030702.1e690d22@posting.google.com>g  [ carl@gerg.tamu.edu (Carl Perkins) wrote in message news:<2SEP200221121352@gerg.tamu.edu>...a4 > spamsink2001@yahoo.com (Alan E. Feldman) writes...d > }Paul Repacholi <prep@prep.synonet.com> wrote in message news:<87it1sxhnj.fsf@prep.synonet.com>... > }> f > } A > }And now, in reference to the "Unix boy" request about checkings > }everything: > I > I'm the one who pointed out this flaw and suggested that checking couldeI > be a good plan. If you think I'm a "Unix boy" then you clearly have nota7 > been paying attention, and you just might be a moron.t    C Relax man. Unix boy was the original poster who asked about findingrC and checking directories. I was referring to John Lewocz. I thoughtm that was pretty obivous.  F Sorry, I misremembered him as checking everything. I now recall (OK, ID re-read his posts) that he just wanted to check if the directory wasC empty. Careless on my part. Sorry. I apologize. But it's not like Iu said "Carl Perkins said...".  C Hey, people make mistakes. Let's not be so quick to pass judgement.t  F Hey, 'sounds like you can become a comedy standup star! You know, likeD those "You just might be a redneck" jokes: You just might be a moronD ... if you name your user files with the form <name>.DIR;1, you just might be a moron.l   > E > }If you want to check if the directory is corrupted, that's another ? > }story. But how much checking do you need to do? How often doSF > }directories get corrupted? I've only seen it happen once myself andI > }I've been using VMS since 1985. And why not check if the disk is stilla* > }mounted? Or if the lights are still on? > F > It isn't really checking to see if the directory is corrupted. It is' > checking to see if it is a directory.3    ! OK, how often does *that* happen?      > $ create foo.dir;1 > foot > ^Z > 9 > This file is not now, nor has it ever been a directory.f >  > It is also not corrupted..    C Well, it's as good as corrupted when a dir-spec is used to refer touD files listed within it. It's not corrupted in the traditional sense,B but it is "effectively corrupted". Upon reading the explanation ofC directories in the user manual, I would think that most users wouldoE get the idea that .DIR;1 (or at least .DIR) is special and should not-% be used as a filetype for user files.- Context matters.    C > It is a perfectly legal file. It just isn't a directory. The file,C > attributes provied the information that indicate that it is not aaG > directory. Yet F$PARSE insists on treating it as a directory (as will F > lots of other things) even though RMS "knows" perfectly well that itB > is not one since the "this is a directory" attribute is not set. > C > I'm also surprised that you would suggest that checking things is D > Unix-like behavior and not VMS-like behavior. In my experience youI > have it exactly backwards. Standard VMS behavior is to check everything(G > that it is practical to check (and maybe some things that aren't veryeI > prcatical) whereas standard Unix behavior is to check only those thingsr5 > that the programmer felt like checking at the time.u    E Well, yes, it's good to check things a lot, to a point. And I have to1> admit, that with my TO.COM, I went to great lengths to make itF bulletproof by considering all sorts of possibilities, both likely and	 unlikely.o  F Also, there's checking, and then there's preemptive checking. You seemD to be talking about preemptive checking. When you run a command withD an invalid argument, VMS doesn't always provide preemptive checking.A How about SHOW ENTRY A? You get the same repsonse as with a valid A entry number that is not currently in use. No preemptive checkinga there, but no harm done.  C And how often do you have this directory "corruption"? How often doeF you have .dir;1 files that are not "true directories"? Where are these< users who have nothing better to do than create files calledB <name>.DIR;1? If someone does do that, you're procedure is screwedF anyway. And DCL is already checking for errors! If you get an error atD that point, trap it and take action accordingly. And allow the errorB message to be displayed so the user can report it. Either way, the; procedure will bomb and manual intervention will be needed.@  B If you still want to check for directory corruption, I guess you'dF have to do something like the DIRECTORY/NOOUT 'test_dir' and check theC value of $STATUS as in your example. Though I'd change one thing inoE your example. You use DIR as the command. Many have DIR defined to be"D something like DIR/SIZE/DATE/PROT or similar. That actually makes itC run slower, even when you suppress the output. So use DIRECTORY and = make sure you don't have DIRECTORY defined as a symbol, whichE shouldn't be anyway.  B And wasn't this DCL code supposed to be used by programmers? Don'tE programmers know not to do such silly things as creating .DIR;1 files A except via the CREATE/DIRECTORY command? Can't programmers handleNE error messages and take action accordingly? Don't programmers have an D even rudimentary understanding of .DIR;1 files? And if they do goof,E explain it to them and you're done! You would have to do that anyway!a  B Also, we're not talking about production code here. And even if we were, it would still bomb.  B Are you worried about rogue processes popping up spontaneously andA littering disks with non-directories that have the .DIR;1 suffix?B  A I've been using VMS for 17 years now and I've only seen an actual-E corruption once (only on one node, it read fine on another) and neversF a problem where I ran into a .DIR;1 file that wasn't a true directory.@ [The corrupted directory problem was bizarre. On one node in theE VAXcluster, it would read fine. On another node, it wouldn't read anymE file past a certain file starting with S. I was a newbie at the time,u" and wasn't able to track it down.]   > ' > }And now, back to our current poster:  > }  > }> Being fixed.l > }  > }??? What's broken?o > }Alan E. Feldman > K > Go back and read the thread again and you might be able to figure it out.  > K > Hint: treating a file that is not now, and never has been, a directory asrK > a directory in spite of the fact that the "this is a directory" attribute02 > is not set, and never was, is not good behavior.    E But how often does it actually cause a problem? In 17 years I haven't3? seen it happen even once. If you want to check for it, use yourd example. Not too hard.  F This would be something for ANAL/DISK to check. And maybe it does, butD I've never had that problem, so I don't know. Maybe I'll check laterA today. ... Well, I just tried it and it doesn't get picked up. Oh E well. But ANAL/DISK (on VMS v6.1) takes long enough to run as is. (!)o$ I suppose it should be added anyway.     Disclaimer: JMHO Alan E. Feldman  spamsink2001 at yahoo dot como   ------------------------------   Date: 3 Sep 2002 02:13:02 -0700A% From: bart.zorn@xs4all.nl (Bart Zorn)r( Subject: Re: Strange F$TRNLNM behaviour?= Message-ID: <9a924482.0209030113.3fc5910a@posting.google.com>a  U Jan-Erik Sderholm <aaa@aaa.com> wrote in message news:<3D7357C8.C7E7B69C@aaa.com>...a0 > I find additional matches after the first even > without the asterisk : >  > $ sh log trace_root29 >    "TRACE_ROOT2" = "DSA200:[TRACE.]" (LNM$SYSTEM_TABLE)l >  >  > $ define trace_root2 sfsfc >  > $ sh log trace_root2/ >    "TRACE_ROOT2" = "SFSF" (LNM$PROCESS_TABLE) 9 >    "TRACE_ROOT2" = "DSA200:[TRACE.]" (LNM$SYSTEM_TABLE)  >  >  > 0 > The asterisk just make it a bit more verbose : >  >  > $ sh log trace_root2*t >  > (LNM$PROCESS_TABLE)t >  >   "TRACE_ROOT2" = "SFSF" >  > (LNM$JOB_80FC4AC0) >  > (LNM$GROUP_000001) >  > (LNM$SYSTEM_TABLE) > % >   "TRACE_ROOT2" = "DSA200:[TRACE.]"u >  > (LNM$SYSCLUSTER_TABLE) >  > (DECW$LOGICAL_NAMES) >  > $ deas trace_root2 >  >  > Best Regards > Jan-Erik Sderholm.  > Larry Kilgallen wrote: > > { > > In article <01KM0PMD4MYQ9OCZIZ@sysdev.deutsche-boerse.com>, Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes:iI > > >> You need to append an asterisk to the logical name in question forh' > > >> SHOW LOGICAL to check all modes:c > > > # > > > I've noticed this also.  Why?k > > D > > I think you will find the difference is not "to check all modes"5 > > but "to find additional matches after the first".t     There are two possible cases:   C 1. There are two definitions of a logical name with the same accessh7 mode in different tables. This applies to your example.p  D 2. There are two definitions of a logical name with different accessA modes in the same table. This one applied to my initial question.e  ) $ SHOW LOGICAL <lnm>   (without asterisk)h  E behaves differently for these two cases, which is unfortunate, to say 
 the least.  	 Bart Zorns   ------------------------------   End of INFO-VAX 2002.486 ************************