INFO-VAX Tue, 18 Nov 2008 Volume 2008 : Issue 624 Contents: Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year Re: 150 year RE: 150 year Re: 150 year Canadian Weather (was Re: In-house testing and patches) Re: In-house testing and patches Re: In-house testing and patches Re: In-house testing and patches Re: In-house testing and patches Re: In-house testing and patches Re: In-house testing and patches Re: Most impressive VAX installations Re: OVMS Integrity BASIC LTU Getting only 1 user at cost of $2400.00??? SWB 1.1.10 -- New quirk? Re: VMS, HP and the recession Re: VMS, HP and the recession ---------------------------------------------------------------------- Date: Mon, 17 Nov 2008 11:48:46 -0800 (PST) From: AEF Subject: Re: 150 year Message-ID: On Nov 17, 11:40=A0am, "Richard B. Gilbert" wrote: > AEF wrote: > > On Nov 17, 3:13 am, "The Mip" wrote: > >> $set ver > >> $ crea foobar.foo > >> $ wr f$file("foobar.foo","bdt") > >> 17-NOV-1858 00:00:00.00 > >> $ wr f$time() > >> 17-NOV-2008 07:53:57.56 > > >> 150 years since i backed up =A0that file :) > > >> - the_mip > > >> Lets celebrate with a good old 1 :http://h71000.www7.hp.com/wizard/wiz= _2315.html > > > So that's fine, but why did VMS adopt this (the Modified Julian Day)? > > It doesn't say. It doesn't answer the question. > > > AEF > > There is an explanation of sorts; something to do with obsoleting star > charts or something like that. =A0There is a detailed explanation at: > > http://h71000.www7.hp.com/wizard/wiz_2315.html > > As the representation used by VMS will not overflow for more than 3,000 > years (it may be 30,000 (I don't recall)) there is no urgent reason to > fix the problem! > > If anybody still cares 3,000 or 30,000 years from now, they can adopt > whatever solution(s) meet their needs. That's the same one that doesn't say why the MJD was adopted by VMS. . . . The mystery remains. AEF ------------------------------ Date: Mon, 17 Nov 2008 11:50:34 -0800 (PST) From: AEF Subject: Re: 150 year Message-ID: On Nov 17, 2:09=A0pm, koeh...@eisner.nospam.encompasserve.org (Bob Koehler) wrote: > In article <6942dd68-2440-4226-b805-399a89a03...@g17g2000prg.googlegroups= .com>, AEF writes: > > > > > So that's fine, but why did VMS adopt this (the Modified Julian Day)? > > It doesn't say. It doesn't answer the question. > > =A0 =A0Prior to first ship. So you're saying they picked MJD because its day 0 pre-dates the first release of VMS? So does Jan 1, 1900. AEF ------------------------------ Date: Mon, 17 Nov 2008 15:06:30 -0500 From: "Richard B. Gilbert" Subject: Re: 150 year Message-ID: <1YCdnenQ2d36U7zUnZ2dnUVZ_obinZ2d@giganews.com> AEF wrote: > On Nov 17, 11:40 am, "Richard B. Gilbert" > wrote: >> AEF wrote: >>> On Nov 17, 3:13 am, "The Mip" wrote: >>>> $set ver >>>> $ crea foobar.foo >>>> $ wr f$file("foobar.foo","bdt") >>>> 17-NOV-1858 00:00:00.00 >>>> $ wr f$time() >>>> 17-NOV-2008 07:53:57.56 >>>> 150 years since i backed up that file :) >>>> - the_mip >>>> Lets celebrate with a good old 1 :http://h71000.www7.hp.com/wizard/wiz_2315.html >>> So that's fine, but why did VMS adopt this (the Modified Julian Day)? >>> It doesn't say. It doesn't answer the question. >>> AEF >> There is an explanation of sorts; something to do with obsoleting star >> charts or something like that. There is a detailed explanation at: >> >> http://h71000.www7.hp.com/wizard/wiz_2315.html >> >> As the representation used by VMS will not overflow for more than 3,000 >> years (it may be 30,000 (I don't recall)) there is no urgent reason to >> fix the problem! >> >> If anybody still cares 3,000 or 30,000 years from now, they can adopt >> whatever solution(s) meet their needs. > > That's the same one that doesn't say why the MJD was adopted by > VMS. . . . The mystery remains. > > AEF I don't see where it uses Julian dates at all. VMS uses the Smithsonian base date as "the beginning of time" and counts the number of 100 nanosecond "ticks" since that date. It doesn't really make any difference that I can see! If you say "SHOW TIME" it gives you the date and time as "17-NOV-2008 16:02:21" which I find entirely adequate. ------------------------------ Date: Mon, 17 Nov 2008 15:14:46 -0500 From: JF Mezei Subject: Re: 150 year Message-ID: <000dadb4$0$12346$c3e8da3@news.astraweb.com> AEF wrote: > That's the same one that doesn't say why the MJD was adopted by > VMS. . . . The mystery remains. When it isn't clearly stated in the documentation, it means that this date was decided at a bar where VMS engineers, after a few beers, did some brainstorming on what sort of "base" dates they should be using, and the one that had absorbed the most beer came up with that one, they laughed and laughed, drank and drank, pissed and pissed, and at the end of the night, couldn't come up with a better one so they decided to use that one. The next morning, they had no clue on WHY they had chosen that one. They just vaguely remembered that they had decided on that one. This is why, my friends, we don't know why they chose 17-NOV-1858 as the base date for VMS. ------------------------------ Date: Mon, 17 Nov 2008 15:26:08 -0500 From: JF Mezei Subject: Re: 150 year Message-ID: <000db06b$0$12274$c3e8da3@news.astraweb.com> Richard B. Gilbert wrote: > It doesn't really make any difference that I can see! If you say "SHOW > TIME" it gives you the date and time as "17-NOV-2008 16:02:21" which I > find entirely adequate. It makes a difference with arithmetic etc. You need to convert from more usable format to the VMS quadword format, and once there, you need to use system routines (especially on VAX which doesn't handle 64 bits well). The Unix system time is just seconds in a 32 bit integer and it simplifies arithmetic being done. Interestingly enough, when GPS system was designed, they did not use the Unix base date, they used 01-JAN-1990. My GPS unit (garmin) uses 01-JAN-1990 is base date. BUT... It has a table of satellite positions. This table is identified by a week number. (so the GPS can know if it has a stale table when you turn it on). This table is based on a date of 05-JAN-1980, and has 10 bits (max value of 1024). On august 22 1999, the GPS date was incremented from 1023 to "0". It was not sure how GPS units back then would react to the week going to 0 ( a number smaller than 1023). The earth did not stop rotating and most units worked fine. ------------------------------ Date: 17 Nov 2008 20:41:13 GMT From: "Bob Eager" Subject: Re: 150 year Message-ID: <176uZD2KcidF-pn2-2sZDsBxBidKE@rikki.tavi.co.uk> On Mon, 17 Nov 2008 16:23:29 UTC, "Tom Linden" wrote: > Here is an interesting addendum, see the comments > on the determination of Easter. > http://www.kednos.com/calendar.html That was interesting! Not having PL/I, I quickly turned it into REXX and it seems to work fine....! -- Bob Eager Use the BIG mirror service in the UK: http://www.mirrorservice.org ------------------------------ Date: Mon, 17 Nov 2008 21:07:33 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: 150 year Message-ID: JF Mezei writes: >AEF wrote: >> That's the same one that doesn't say why the MJD was adopted by >> VMS. . . . The mystery remains. >When it isn't clearly stated in the documentation, it means that this >date was decided at a bar where VMS engineers, after a few beers, did >some brainstorming on what sort of "base" dates they should be using, >and the one that had absorbed the most beer came up with that one, they >laughed and laughed, drank and drank, pissed and pissed, and at the end >of the night, couldn't come up with a better one so they decided to use >that one. Hah! Probably not far from the truth! If it was up to me (and I remained sober enough) I would have chosen the start of the Gregorian calendar. I'm not sure _which_ start, either the overall start in 1582 (no dates before then were in the Gregorian calendar) or perhaps the start of the use of the Gregorian calendar in Britain and her colonies (1752). That date may be good since many dates before then were on the Julian calendar and would be rendered inaccurately by VMS if the 1582 date were chosen. >The next morning, they had no clue on WHY they had chosen that one. They >just vaguely remembered that they had decided on that one. >This is why, my friends, we don't know why they chose 17-NOV-1858 as the >base date for VMS. ------------------------------ Date: Mon, 17 Nov 2008 13:27:58 -0800 (PST) From: AEF Subject: Re: 150 year Message-ID: On Nov 17, 4:06=A0pm, "Richard B. Gilbert" wrote: > AEF wrote: > > On Nov 17, 11:40 am, "Richard B. Gilbert" > > wrote: > >> AEF wrote: > >>> On Nov 17, 3:13 am, "The Mip" wrote: > >>>> $set ver > >>>> $ crea foobar.foo > >>>> $ wr f$file("foobar.foo","bdt") > >>>> 17-NOV-1858 00:00:00.00 > >>>> $ wr f$time() > >>>> 17-NOV-2008 07:53:57.56 > >>>> 150 years since i backed up =A0that file :) > >>>> - the_mip > >>>> Lets celebrate with a good old 1 :http://h71000.www7.hp.com/wizard/w= iz_2315.html > >>> So that's fine, but why did VMS adopt this (the Modified Julian Day)? > >>> It doesn't say. It doesn't answer the question. > >>> AEF > >> There is an explanation of sorts; something to do with obsoleting star > >> charts or something like that. =A0There is a detailed explanation at: > > >>http://h71000.www7.hp.com/wizard/wiz_2315.html > > >> As the representation used by VMS will not overflow for more than 3,00= 0 > >> years (it may be 30,000 (I don't recall)) there is no urgent reason to > >> fix the problem! > > >> If anybody still cares 3,000 or 30,000 years from now, they can adopt > >> whatever solution(s) meet their needs. > > > That's the same one that doesn't say why the MJD was adopted by > > VMS. . . . The mystery remains. > > > AEF > > I don't see where it uses Julian dates at all. =A0VMS uses the Smithsonia= n > base date as "the beginning of time" and counts the number of 100 > nanosecond "ticks" since that date. > > It doesn't really make any difference that I can see! =A0If you say "SHOW > TIME" it gives you the date and time as "17-NOV-2008 16:02:21" which I > find entirely adequate. But if they had chosen Unix time (t_0 - Jan 1, 1970), then $ DIR / SINC=3D17-NOV-1858 would fail. So the question is, Why was MJD day 0 chosen to be VMS time zero? Here's another way to put it: The mapping of dates and times to the binary quadword time would be different if some other time were chosen to be t_0 in VMS. AEF AEF ------------------------------ Date: Mon, 17 Nov 2008 18:31:27 -0700 From: Jeff Campbell Subject: Re: 150 year Message-ID: <1226971540_1181@isp.n> AEF wrote: > On Nov 17, 11:40 am, "Richard B. Gilbert" > wrote: >> AEF wrote: >>> On Nov 17, 3:13 am, "The Mip" wrote: >>>> $set ver >>>> $ crea foobar.foo >>>> $ wr f$file("foobar.foo","bdt") >>>> 17-NOV-1858 00:00:00.00 >>>> $ wr f$time() >>>> 17-NOV-2008 07:53:57.56 >>>> 150 years since i backed up that file :) >>>> - the_mip >>>> Lets celebrate with a good old 1 :http://h71000.www7.hp.com/wizard/wiz_2315.html >>> So that's fine, but why did VMS adopt this (the Modified Julian Day)? >>> It doesn't say. It doesn't answer the question. >>> AEF >> There is an explanation of sorts; something to do with obsoleting star >> charts or something like that. There is a detailed explanation at: >> >> http://h71000.www7.hp.com/wizard/wiz_2315.html >> >> As the representation used by VMS will not overflow for more than 3,000 >> years (it may be 30,000 (I don't recall)) there is no urgent reason to >> fix the problem! >> >> If anybody still cares 3,000 or 30,000 years from now, they can adopt >> whatever solution(s) meet their needs. > > That's the same one that doesn't say why the MJD was adopted by > VMS. . . . The mystery remains. > > AEF It's Julian day 2,400,000. ----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==---- http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups ---= - Total Privacy via Encryption =--- ------------------------------ Date: Mon, 17 Nov 2008 21:18:49 -0500 From: "Richard B. Gilbert" Subject: Re: 150 year Message-ID: JF Mezei wrote: > Richard B. Gilbert wrote: > >> It doesn't really make any difference that I can see! If you say "SHOW >> TIME" it gives you the date and time as "17-NOV-2008 16:02:21" which I >> find entirely adequate. > > > It makes a difference with arithmetic etc. You need to convert from more > usable format to the VMS quadword format, and once there, you need to > use system routines (especially on VAX which doesn't handle 64 bits well) Well, I mostly avoid the 32 bit problem by using an Alpha! . > > The Unix system time is just seconds in a 32 bit integer and it > simplifies arithmetic being done. And if you want/need milliseconds or microseconds?? I don't know about ALL Unix versions but Solaris handles time down to the microseconds. ------------------------------ Date: Mon, 17 Nov 2008 21:28:26 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: 150 year Message-ID: <49222845$0$90265$14726298@news.sunsite.dk> Michael Moroney wrote: > If it was up to me (and I remained sober enough) I would have chosen the > start of the Gregorian calendar. I'm not sure _which_ start, either the > overall start in 1582 (no dates before then were in the Gregorian calendar) > or perhaps the start of the use of the Gregorian calendar in Britain and > her colonies (1752). That date may be good since many dates before then > were on the Julian calendar and would be rendered inaccurately by VMS > if the 1582 date were chosen. The transformation between true time and calendar could have been made locale specific. Or it could have been if it were implemented today. I don't think that way of doing things were common in 1978. Arne ------------------------------ Date: Mon, 17 Nov 2008 21:32:18 -0500 From: "Richard B. Gilbert" Subject: Re: 150 year Message-ID: <38SdnaHKnM1Jtb_UnZ2dnUVZ_tjinZ2d@giganews.com> Michael Moroney wrote: > JF Mezei writes: > >> AEF wrote: > >>> That's the same one that doesn't say why the MJD was adopted by >>> VMS. . . . The mystery remains. > >> When it isn't clearly stated in the documentation, it means that this >> date was decided at a bar where VMS engineers, after a few beers, did >> some brainstorming on what sort of "base" dates they should be using, >> and the one that had absorbed the most beer came up with that one, they >> laughed and laughed, drank and drank, pissed and pissed, and at the end >> of the night, couldn't come up with a better one so they decided to use >> that one. > > Hah! Probably not far from the truth! > > If it was up to me (and I remained sober enough) I would have chosen the > start of the Gregorian calendar. I'm not sure _which_ start, either the > overall start in 1582 (no dates before then were in the Gregorian calendar) > or perhaps the start of the use of the Gregorian calendar in Britain and > her colonies (1752). That date may be good since many dates before then > were on the Julian calendar and would be rendered inaccurately by VMS > if the 1582 date were chosen. >> The next morning, they had no clue on WHY they had chosen that one. They >> just vaguely remembered that they had decided on that one. > >> This is why, my friends, we don't know why they chose 17-NOV-1858 as the >> base date for VMS. In my entire VMS career, I never needed to represent a date/time that VMS could not accommodate! In fact, I don't think I ever needed to represent a date prior to 1984 which was the year I was VAXinated. ------------------------------ Date: Mon, 17 Nov 2008 21:38:47 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: 150 year Message-ID: <49222ab4$0$90276$14726298@news.sunsite.dk> JF Mezei wrote: > Richard B. Gilbert wrote: >> It doesn't really make any difference that I can see! If you say "SHOW >> TIME" it gives you the date and time as "17-NOV-2008 16:02:21" which I >> find entirely adequate. > > > It makes a difference with arithmetic etc. You need to convert from more > usable format to the VMS quadword format, and once there, you need to > use system routines (especially on VAX which doesn't handle 64 bits well). movq addl2 + adwc subl2 + sbwc emul ediv worked fine for me. > The Unix system time is just seconds in a 32 bit integer and it > simplifies arithmetic being done. That is the old way. Unix - 32 bit time - 1970 VMS - 64 bit time - 1978 Windows NT - 64 bit time - 1993 Java - 64 bit time - 1995 .NET - 64 bit time - 2000 Oh - and I think most Unixes already have switched to 64 bit time. Arne ------------------------------ Date: Mon, 17 Nov 2008 21:46:48 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: 150 year Message-ID: <49222c95$0$90268$14726298@news.sunsite.dk> Bob Koehler wrote: > In article , "Richard B. Gilbert" writes: >> As the representation used by VMS will not overflow for more than 3,000 >> years (it may be 30,000 (I don't recall)) there is no urgent reason to >> fix the problem! >> >> If anybody still cares 3,000 or 30,000 years from now, they can adopt >> whatever solution(s) meet their needs. > > IIRC the internal representation is good through year 32767. The > formatting routines are good through year 9999. Fixes to the > formatting routines have been promised prior to year 10000. > > 3000 years from now is not a problem for VMS. The range must be 2^63/10000000 seconds = 922337203685 seconds = 29247 years So it should be good until year 31105. Arne ------------------------------ Date: Tue, 18 Nov 2008 04:25:40 +0100 From: "Martin Vorlaender" Subject: Re: 150 year Message-ID: Michael Moroney wrote: > JF Mezei writes: >> AEF wrote: >>> That's the same one that doesn't say why the MJD was adopted by >>> VMS. . . . The mystery remains. > >> When it isn't clearly stated in the documentation, it means that this >> date was decided at a bar where VMS engineers, after a few beers, did >> some brainstorming on what sort of "base" dates they should be using, >> and the one that had absorbed the most beer came up with that one, they >> laughed and laughed, drank and drank, pissed and pissed, and at the end >> of the night, couldn't come up with a better one so they decided to use >> that one. > > Hah! Probably not far from the truth! The very same thought crossed my mind upon reading it. > If it was up to me (and I remained sober enough) I would have chosen the > start of the Gregorian calendar. I'm not sure _which_ start, either the > overall start in 1582 (no dates before then were in the Gregorian calendar) > or perhaps the start of the use of the Gregorian calendar in Britain and > her colonies (1752). But it wasn't adopted in Russia until after the october revolution (1922) - which is why the celebrations for it later took place in september. And Greece, Turkey and China were even later (1923, 1926, and 1929, respectively). See http://en.wikipedia.org/wiki/Gregorian_Calendar#Timeline What a mess... > That date may be good since many dates before then > were on the Julian calendar and would be rendered inaccurately by VMS > if the 1582 date were chosen. > >> The next morning, they had no clue on WHY they had chosen that one. They >> just vaguely remembered that they had decided on that one. >> >> This is why, my friends, we don't know why they chose 17-NOV-1858 as the >> base date for VMS. cu, Martin -- One OS to rule them all | Martin Vorlaender | OpenVMS rules! One OS to find them | work: mv@pdv-systeme.de One OS to bring them all | http://vms.pdv-systeme.de/users/martinv/ And in the Darkness bind them.| home: martin.vorlaender@t-online.de ------------------------------ Date: Mon, 17 Nov 2008 19:29:02 -0800 (PST) From: David B Sneddon Subject: Re: 150 year Message-ID: <9d1f2a5e-294e-4845-9963-2ee08b991436@t39g2000prh.googlegroups.com> On Nov 18, 11:46=A0am, Arne Vajh=F8j wrote: > > The range must be 2^63/10000000 seconds > =3D 922337203685 seconds > =3D 29247 years > > So it should be good until year 31105. > > Arne It goes negative at 31-JUL-31086 02:48:05.47... Dave ------------------------------ Date: Mon, 17 Nov 2008 22:31:22 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: 150 year Message-ID: <49223706$0$90264$14726298@news.sunsite.dk> David B Sneddon wrote: > On Nov 18, 11:46 am, Arne Vajhøj wrote: >> The range must be 2^63/10000000 seconds >> = 922337203685 seconds >> = 29247 years >> >> So it should be good until year 31105. > > It goes negative at 31-JUL-31086 02:48:05.47... That is probably because I am sloppy and did some roundings and forgot about leap years etc.. Arne ------------------------------ Date: Tue, 18 Nov 2008 14:07:35 +1030 From: "Barratt, Chris (Health)" Subject: RE: 150 year Message-ID: > -----Original Message----- > From: Richard B. Gilbert [mailto:rgilbert88@comcast.net] > Sent: Tuesday, 18 November 2008 13:02 > To: Info-VAX@Mvb.Saic.Com > Subject: Re: 150 year > > > In my entire VMS career, I never needed to represent a date/time that > VMS could not accommodate! In fact, I don't think I ever needed to > represent a date prior to 1984 which was the year I was VAXinated. Never had to represent a date of birth for anyone born before 1984 ? ------------------------------ Date: Mon, 17 Nov 2008 22:48:31 -0500 From: "Richard B. Gilbert" Subject: Re: 150 year Message-ID: Barratt, Chris (Health) wrote: > >> -----Original Message----- >> From: Richard B. Gilbert [mailto:rgilbert88@comcast.net] >> Sent: Tuesday, 18 November 2008 13:02 >> To: Info-VAX@Mvb.Saic.Com >> Subject: Re: 150 year >> > >> In my entire VMS career, I never needed to represent a date/time that >> VMS could not accommodate! In fact, I don't think I ever needed to >> represent a date prior to 1984 which was the year I was VAXinated. > > Never had to represent a date of birth for anyone born before 1984 ? > Nope! Now some of my users did; I had one system that did some HR stuff but *I* never did. ------------------------------ Date: Mon, 17 Nov 2008 14:36:18 -0800 From: Fred Bach Subject: Canadian Weather (was Re: In-house testing and patches) Message-ID: Michael Kraemer wrote: > JF Mezei schrieb: >> While picking up leaves, I got to think... (it happens sometimes). > > I always thought Canada would be buried unter several meters > of snow at this time of the year ... Are you kidding? I live in Canada too, and I use my lawnmower every month of the year. And yes, I do let my friends up in Edmonton know that, as often as I can. Today we're having pleasant spring-like weather, and I intend to cut the grass after work today. .. Fred Bach music at triumf dot c a ------------------------------ Date: Mon, 17 Nov 2008 20:44:34 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: In-house testing and patches Message-ID: JF Mezei writes: >While picking up leaves, I got to think... (it happens sometimes). >In the 1980s, DEC had the largest in-house network. It used its own >hardware and osftware to run the corporation. It used its own ALL-IN-1 >(and mail) to have the in-house email facility. It had its own >applications using FMS and whatever. >As customers, we didn't see many patches flowing to us. >Then comes Bob GQ Palmer who torched the company, ditched DEC's own >products and installed "industry standard" stuff. I would assume Microsoft uses its own OS's, Outlook etc. and therefore its software quality would be excellent, according to that logic. How big is M$'s internal network? >One has to admit that quality control since the 1990s seems to have gone >down. (just look at the TCPIP services product). Just look at UCX, before the current TCPIP. VMS was late getting into the TCPIP game and UCX was bad. >Out of curiosity, is it possible that the high quality of VMS >products/software we saw in the 1980s was due to Digital using that very >software in house and engineers knew that they had to design something >to be used in a 20,000 node environment with high scrutiny over security >etc and the end product , by being ready to run inside Digital ,was also >ready and tested to run on most customer environmnents. I'm sure the DECnet node environment was larger than that at its peak. When I was there, there was announcements that DECnet area N was going to split into Areas N and M, due to imminent exceeding the range of nodes in one DECnet area (1023) in a major facility or region. Then there were 'reverse splits' (everything in a smallish Area N was moved into another smallish area to free up an area for another split. Then they created things like multiple Area 63s for "unimportant" systems like printers that didn't need to reach outside the local DECnet area. Kind of the DECnet equivalent of RFC 1918. There could have been approaching 60,000 nodes if you exclude those in the multiple Area 63's. >But once Palmer started to dismantle the original Digital Network >(completely gone with Compaq and HP), engineers no longer had that large >scale in-house test bed whcih generated much internal feedback on >problems and had to rely on the small labs without much feedback before >the products were signed, sealed and distributed to customers. No, it was the world went with TCP/IP and not DECnet, and VMS was weak in TCP/IP. That is, of course, just the networking part. ------------------------------ Date: Mon, 17 Nov 2008 20:50:12 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: In-house testing and patches Message-ID: <49221f52$0$90274$14726298@news.sunsite.dk> Michael Kraemer wrote: > JF Mezei schrieb: >> Then comes Bob GQ Palmer who torched the company, ditched DEC's own >> products and installed "industry standard" stuff. > > I'd rather think he tried to clean up the mess > the 1980's DEC management had created. The result was not very good for DEC. It is hard to say if it would have been better with another strategy. But I see no proof of him cleaning up any mess. He killed and sold a lot of products. >> One has to admit that quality control since the 1990s seems to have gone >> down. (just look at the TCPIP services product). > > IIRC UCX predates the Palmer era, > and quality wasn't superior already in the early 1990s. So true. I remember UCX 1.3. TELNET and 2 x CTRL/Y and the system crashed. Horrible quality. Arne ------------------------------ Date: Mon, 17 Nov 2008 20:56:09 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: In-house testing and patches Message-ID: <492220b6$0$90274$14726298@news.sunsite.dk> JF Mezei wrote: > In the 1980s, DEC had the largest in-house network. It used its own > hardware and osftware to run the corporation. It used its own ALL-IN-1 > (and mail) to have the in-house email facility. It had its own > applications using FMS and whatever. > > As customers, we didn't see many patches flowing to us. > > Then comes Bob GQ Palmer who torched the company, ditched DEC's own > products and installed "industry standard" stuff. > > One has to admit that quality control since the 1990s seems to have gone > down. (just look at the TCPIP services product). > > Out of curiosity, is it possible that the high quality of VMS > products/software we saw in the 1980s was due to Digital using that very > software in house and engineers knew that they had to design something > to be used in a 20,000 node environment with high scrutiny over security > etc and the end product , by being ready to run inside Digital ,was also > ready and tested to run on most customer environmnents. First of all then security was not that good back then. Remember open DECnet accounts and the DECnet worm ? Was quality in general better back then ? Probably yes. But I don't think it has much to do with the software being used internally. If that were the case then it would mean that the engineers have a horrible disrespect for paying customers. I don't think that is the case. The reasons are more likely to be: - more people in engineering - much less code to maintain - DIY code instead of 5 million lines of C code imported where engineering do not even have time to read the code Arne ------------------------------ Date: Mon, 17 Nov 2008 21:59:19 -0500 From: "Richard B. Gilbert" Subject: Re: In-house testing and patches Message-ID: Arne Vajhøj wrote: > Michael Kraemer wrote: >> JF Mezei schrieb: >>> Then comes Bob GQ Palmer who torched the company, ditched DEC's own >>> products and installed "industry standard" stuff. >> >> I'd rather think he tried to clean up the mess >> the 1980's DEC management had created. > > The result was not very good for DEC. > > It is hard to say if it would have been better with another > strategy. > > But I see no proof of him cleaning up any mess. > > He killed and sold a lot of products. > >>> One has to admit that quality control since the 1990s seems to have gone >>> down. (just look at the TCPIP services product). >> >> IIRC UCX predates the Palmer era, >> and quality wasn't superior already in the early 1990s. > > So true. > > I remember UCX 1.3. > > TELNET and 2 x CTRL/Y and the system crashed. > > Horrible quality. > > Arne It hadn't gotten any better in the late 1990s!! In 1998 I had to manage two VAX/VMS systems at Philadelphia College of Textiles and Science (now Philadelphia University). They were at VMS V5.1 and UCX 2.7E (I think that's the version). E-mail via TCP/IP didn't work worth a damn but one of the systems was supposed to handle E-mail for 2,000 people; students, faculty, and staff. I was advised by tech support to get current so I upgraded, in stages, to 5.5-2 and UCX 3.3. E-mail still didn't work. Apply ECOs up through ECO-13; still had problems with E-mail. ECO-13 was ALMOST good enough but by no means perfect. They didn't get fully reliable E-mail until UCX 4.x where x was greater than zero. I got well acquainted with the E-mail developer. The developer was in hog heaven; he had a test system where anything that COULD go wrong DID go wrong. ------------------------------ Date: Mon, 17 Nov 2008 22:05:13 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: In-house testing and patches Message-ID: <492230e5$0$90264$14726298@news.sunsite.dk> Richard B. Gilbert wrote: > Arne Vajhøj wrote: >> I remember UCX 1.3. >> >> TELNET and 2 x CTRL/Y and the system crashed. >> >> Horrible quality. > > It hadn't gotten any better in the late 1990s!! > > In 1998 I had to manage two VAX/VMS systems at Philadelphia College of > Textiles and Science (now Philadelphia University). > > They were at VMS V5.1 and UCX 2.7E (I think that's the version). E-mail > via TCP/IP didn't work worth a damn but one of the systems was supposed > to handle E-mail for 2,000 people; students, faculty, and staff. > > I was advised by tech support to get current so I upgraded, in stages, > to 5.5-2 and UCX 3.3. E-mail still didn't work. Apply ECOs up through > ECO-13; still had problems with E-mail. ECO-13 was ALMOST good enough > but by no means perfect. > > They didn't get fully reliable E-mail until UCX 4.x where x was greater > than zero. I got well acquainted with the E-mail developer. The > developer was in hog heaven; he had a test system where anything that > COULD go wrong DID go wrong. I was more lucky. PMDF :-) Arne ------------------------------ Date: Mon, 17 Nov 2008 22:41:44 -0500 From: "Richard B. Gilbert" Subject: Re: In-house testing and patches Message-ID: Arne Vajhøj wrote: > Richard B. Gilbert wrote: >> Arne Vajhøj wrote: >>> I remember UCX 1.3. >>> >>> TELNET and 2 x CTRL/Y and the system crashed. >>> >>> Horrible quality. >> >> It hadn't gotten any better in the late 1990s!! >> >> In 1998 I had to manage two VAX/VMS systems at Philadelphia College of >> Textiles and Science (now Philadelphia University). >> >> They were at VMS V5.1 and UCX 2.7E (I think that's the version). >> E-mail via TCP/IP didn't work worth a damn but one of the systems was >> supposed to handle E-mail for 2,000 people; students, faculty, and staff. >> >> I was advised by tech support to get current so I upgraded, in stages, >> to 5.5-2 and UCX 3.3. E-mail still didn't work. Apply ECOs up >> through ECO-13; still had problems with E-mail. ECO-13 was ALMOST >> good enough but by no means perfect. >> >> They didn't get fully reliable E-mail until UCX 4.x where x was >> greater than zero. I got well acquainted with the E-mail developer. >> The developer was in hog heaven; he had a test system where anything >> that COULD go wrong DID go wrong. > > I was more lucky. > > PMDF > > :-) > > Arne But I learned more, met more of the big guns in Colorado (by phone, of course), etc, etc. ------------------------------ Date: Mon, 17 Nov 2008 21:04:15 +0000 (UTC) From: legalize+jeeves@mail.xmission.com (Richard) Subject: Re: Most impressive VAX installations Message-ID: [Please do not mail me a copy of your followup] JF Mezei spake the secret code <000c7fc6$0$12296$c3e8da3@news.astraweb.com> thusly: >Consider the fancy 3D games, those also require a hefty amount of >resources on a PC. For the past few years, multimedia and gaming have been the main things driving PC upgrade cycles. Word processing and typical business applications have not been the driving factor in PC upgrades for some time, unless you count the minimum requirements of Office 2007, etc. If you don't upgrade and don't need the new features, you can get by just nicely with an older machine. -- "The Direct3D Graphics Pipeline" -- DirectX 9 draft available for download Legalize Adulthood! ------------------------------ Date: Mon, 17 Nov 2008 22:07:06 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: OVMS Integrity BASIC LTU Getting only 1 user at cost of $2400.00??? Message-ID: <49223156$0$90264$14726298@news.sunsite.dk> Main, Kerry wrote: >> From: Arne Vajhøj [mailto:arne@vajhoej.dk] >> Main, Kerry wrote: >>>> From: Arne Vajhøj [mailto:arne@vajhoej.dk] >>>> Main, Kerry wrote: >>>>>> From: JF Mezei [mailto:jfmezei.spamnot@vaxination.ca] >>>>>> Jan-Erik Söderholm wrote: >>>>>>> The SPD (Software Product Description) for HP BASIC is >>>>>>> reasonable clear on the licensing options. In IA64, the >>>>>>> "Concurrent Use License" is the only option available, >>>>>> Then $2400 for a single concurrent use is pretty expensive. >>>>>> >>>>>> I know that HP apologists will point to DSPP where compilers are dirt >>>>>> cheap. But for people who do development in-house, they don't qualify >>>>>> for DSPP and forcing them to pay those horrendous prices is not right. >>>>> So I guess the Enterprise Oracle licensing at $40K USD/cpu (not system) >>>>> or BEA at $10K per cpu must really upset you then? >>>>> >>>>> :-) >>>> That is for production server stuff. >>>> >>>> AFAIK then Oracle development tools are completely free. >>> At the end of the day, its one Oracle bill for whoever is paying. >> It is often two bills going to different departments and >> different budgets (if not different companies). > > No, that is not usually the case. > > The CIO usually has both Development and Prod budget responsibilities. > > If not, then that company is in serious trouble as there is no single > throat to choke from a CEO perspective. Sure the CIO is responsible for both. But that does most certainly not prevent the entire IT budget from being chopped up and allocated to specific purposes. Arne ------------------------------ Date: Mon, 17 Nov 2008 23:12:34 -0600 (CST) From: sms@antinode.info (Steven M. Schweda) Subject: SWB 1.1.10 -- New quirk? Message-ID: <08111723123419_202004DB@antinode.info> Trying out the new "SWB 1.1.10 | Mozilla/5.0 (X11; U; OpenVMS COMPAQ_Professional_Workstation; en-US; rv:1.8.1.15) Gecko/20080928 SeaMonkey/1.1.10" on my VMS V7.3-2 system (with VMS732_UPDATE V15.0), I noticed a couple of quirks. At first, the "hp" logo progress indicator (upper right) was going blank instead of rolling around as expected during a Reload, but that problem seems to have disappeared after a browser restart. (I don't know if I was lucky before or now.) Now, the annoyance is a different visual quirk. If I move the cursor onto the "Reload" button by passing over a button directly below it (might be "Bookmarks", might be "The Mozilla Organization", depending on whether "Home" is there on the left), instead of the normally highlighted "Reload" button with its usual curvy blue arrow, I get a slightly corrupted green arrow, like the "Back" arrow, but with a few extra green pixels which look like leftovers from the proper (but blue) "Reload" arrow. It doesn't happen every time, but it's pretty common. (Might be a coin toss.) Anyone else see any of this stuff? I didn't notice any of it with the pre-SeaMonkey SWB. Naturally, (and this _is_ like the pre-SeaMonkey SWB) the FTP code still can't cope with even simple (but long) ODS2 file names as presented by a TCPIP FTP server: ftp://antinode.info/moz_test/ (Each directory there contains three files, as can easily be seen using a simple FTP client program.) I assume that it's also still bewildered by ODS5 extended file names, or at least by the way a TCPIP FTP server deals with them. (But who wouldn't be?) ------------------------------------------------------------------------ Steven M. Schweda sms@antinode-info 382 South Warwick Street (+1) 651-699-9818 Saint Paul MN 55105-2547 ------------------------------ Date: Mon, 17 Nov 2008 21:55:25 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: VMS, HP and the recession Message-ID: <49222e98$0$90264$14726298@news.sunsite.dk> JF Mezei wrote: > Consider that even some of the VMS engineers who participated here last > century did not see the usefulness of marketing/advertising. So it seems > to me that there may be a strong cultural issue within the VMS group > which accepts readily HP's decisions to not try to market VMS and not > fight for advertising budgets. > > If the VMS group as a whole, has an image of not fighting back and just > taking it, it becomes even easier for Stallard/Livermore to shift job > cuts to VMS to reduce impact on the other departments. I would expect VMS engineers to do the job they are paid to do not act as c.o.v ambassadors within HP. If not they would become ex VMS engineers rather quickly. Arne ------------------------------ Date: Mon, 17 Nov 2008 22:04:00 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: VMS, HP and the recession Message-ID: <4922309e$0$90264$14726298@news.sunsite.dk> Richard Maher wrote: > Consider instead a couple of examples of what we know HP/VMS to have deigned > "useful" in the not too distant past: - BridgeWorks (IIRC you're a big fan?) > and SOAP/Toolkit. Tell us how bloody useful they were Arne, and what was the > opportunity cost (let alone the dollar-cost) to a more deserving > architecture(s)? Both seems rather useful to me. And I am not quite sure what you consider a better alternative. > Now (after another, no expense spared, Cecil B Demille effort) we have WSIT > and Axis2; so who the bloody hell is using them? (And will Java6 ever arrive > on VMS?) But I'm sure you're right, HP must know best; VMS certainly didn't > get where it is today by accident :-( Java 6 is late which is not good. > As far as what *I* consider useful? Well Java/Applets, Adobe/Flex, and > Microsoft/Silverlight, have all expressed more than a passing interest in > the technology. Those do not substitute but complement BW/WSIT/etc.. And there are really not that much HP can do in that area. None of those technologies are owned by HP. They will typical not run on a HP platform. All HP has to is supply the server side tools to work with them. And they have supplied all the basic tools. I do not see it as the task of the platform supplier to deliver more specific tools. > (Not to mention WebSockets) But as witnessed in another > thread here (that I will reply to :-) the prevailing Comet wisdom is to have > one server-process per client; so 1000 clients produces 1000 VMS processes > that do nothing much except page in and out now and then. I do not see much future for Comet. > PS. I bet you've really got the hots for Java 1.6_10 and thos Socket Policy > Files! Cross-Domain access? Port-level granularity? I have not even looked at 1.6_10 - those technologies are not really within my areas of interest. 1.7 on the other hand will have some goodies. Arne ------------------------------ End of INFO-VAX 2008.624 ************************