1 INFO-VAX	Tue, 19 Aug 2003	Volume 2003 : Issue 457       Contents: Re: $asctim  Re: $asctim  Re: 306GB drives!  RE: 306GB drives!  Re: 306GB drives!  Re: 306GB drives!  Re: 306GB drives! ( Call for next FAQ edition, next Freeware" Re: CSA/CSE Certifications Request Re: DEC 3000 M500 CXT error ? ( Emulator CDs (console, computer, arcade)% Re: Fast ethernet and DECnet phase IV  Re: FS: VAXstation 4000/90 FS: VAXstation 4000/90 Re: FS: VAXstation 4000/90 HP's image is improving - Re: IBM Beats HP for Computing Deal with Lego 2 Re: Memo:  problem converting data type - Macro 32< Re: Overlapped IO with redirected SYS$OUTPUT for Sub Program< Re: Overlapped IO with redirected SYS$OUTPUT for Sub Program  Re: Querying UAF from MS Windows# Re: running a news SERVER under VMS # Re: running a news SERVER under VMS  Strange programming problem  Re: Strange programming problem  Re: Strange programming problem   F ----------------------------------------------------------------------  # Date: Mon, 18 Aug 2003 21:30:58 GMT # From: hoff@hp.nospam (Hoff Hoffman)  Subject: Re: $asctim1 Message-ID: <mAb0b.2620$3r2.665@news.cpqcorp.net>   p In article <24beb743.0308082234.752a2599@posting.google.com>, hidra2000@yahoo.com (Eyal Sharabi Horwitz) writes:B :im having a bit of difficulty with the $asctim function, for someD :reason, every once in a while it gives me a date that's not in it's2 :usual format, which causes a crash in my program. : C :is there a default return value, or a default error value for this 
 :function?  F   If $asctim were broken -- certainly a situation that is conceptuallyG   possible, of course -- I would assume the error would be widely known D   and reported, as $asctim is used all over the place in OpenVMS andH   customer and partner applications.  Hence I must first assume that the@   problem is within the (unspecified) application code involved.  H   Please consider if you would you like to receive this or similar as anI   error report against one of your own products -- no offense is intended G   here, of course.  But please realize you have given us almost nothing G   to go on here -- not the OpenVMS version(s), not the platform(s), not F   the language, not the error(s), and particularly not the source codeG   involved, etc.  Only a single report of an unspecified and apparently    indeterminate problem.  G   As a completely wild guess with no (present) foundation or supporting F   details, I will guess that you have messed up somwhere with a stringF   descriptor or descriptor-related processing.  If this is coded in C,G   I'm going to further guess errors involving trailing nulls or similar H   assumptions -- please see the OpenVMS FAQ for some related information   on C and string descriptors.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------  % Date: Tue, 19 Aug 2003 03:09:52 -0000 ! From: Z  <zarlenga@conan.ids.net>  Subject: Re: $asctim/ Message-ID: <vk3580r2aanef2@corp.supernews.com>   1 Eyal Sharabi Horwitz <hidra2000@yahoo.com> wrote: C : im having a bit of difficulty with the $asctim function, for some E : reason, every once in a while it gives me a date that's not in it's 3 : usual format, which causes a crash in my program.   % Do you have an example of a bad date?    ------------------------------    Date: 18 Aug 2003 10:22:46 -0500+ From: young_r@encompasserve.org (Rob Young)  Subject: Re: 306GB drives!3 Message-ID: <XhkGNJS3lxte@eisner.encompasserve.org>   _ In article <oDOdnXM5AeWu7N2iXTWJkg@metrocast.net>, "Bill Todd" <billtodd@metrocast.net> writes:  > : > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:zYVV3vv$GJP9@eisner.encompasserve.org... K >> In article <_Nu%a.93731$o27.2153066@twister.rdc-kc.rr.com>, "Mike Naime"  > <mnaime@kc.rr.com> writes: >> >; >> > Rob Young <young_r@encompasserve.org> wrote in message 2 >> > news:Xrz7BErQmnjo@eisner.encompasserve.org...G >> >> In article <%lj%a.93485$o27.2125405@twister.rdc-kc.rr.com>, "Mike  > Naime" >> > <mnaime@kc.rr.com> writes:  >> >? >> >> But you certainly could do a cheaper solution for backups > >> >> (if new) since you have existing infrastructure it makes5 >> >> a lot of sense for you to pop in bigger drives.  >> >>  >> >H >> > Cheaper, yes.  No argument there.  But cheaper usually is inverselyK >> > proportional to faster!  After a test restore from tape.  (The cheaper M >> > solution)   Upper management let my director buy more storage for online : >> > backups.  Restoring from tape is a 3rd or 4th option. >> > >>? >> I wasn't talking about tape.  A cheaper disk based way would E >> be to use ATA.  For instance, Nexsan ATA is about 1.7 cents/MByte. A >> You can get adventurous and get cheaper than that.  But at 1.7 ! >> cents that is $2 or so a GByte  > . > This just isn't your weekend for arithmetic. >    	I guess not.      				Rob    ------------------------------    Date: 18 Aug 2003 10:26:26 -0500+ From: young_r@encompasserve.org (Rob Young)  Subject: RE: 306GB drives!3 Message-ID: <ay0ZrAQmRiSI@eisner.encompasserve.org>   a In article <S8RRaP6jJx80@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: a > In article <CIEJLCMNHNNDLLOOGNJIKEIJHMAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes:  >>   >  >>>>A >>>> I wasn't talking about tape.  A cheaper disk based way would G >>>> be to use ATA.  For instance, Nexsan ATA is about 1.7 cents/MByte. C >>>> You can get adventurous and get cheaper than that.  But at 1.7 C >>>> cents that is $2 or so a GByte, you won't be buying HSG drives  >>>> that cheap any time soon. >>>>- >>>> http://www.fusiondm.net/pages/ataboy.htm  >>  M >> 72GB scsi sca drives are avaialble for about $235 which works out to about I >> 0.3 cents/Mbyte and to this you must add the cost of the cabinetry and ' >> cabling, so say a total of 0.4 cents  >>   >  > 	You missed something: >  > 	72 * 1024 = 73728 MB  > 	 . > 	73728 / 23500 pennies is 3.1+ cents per MB. >    	Ouch.  8 > 	It is easy to find 200 GByte ATA drives that sell for > 	around $200.  >   @ 	Maybe - since I'm not so good at arithmetic this past week end G 	as Bill points out - a better comparison would be $235 for a 72 GByte  H 	SCSI drive, and $200 for a 200 GByte ATA drive.  And of course cheaper E 	than 3 cents a MByte.  What threw me was the Nexsan price (no excuse ? 	for blowing out the math) at 1.7 cents / MByte.  That includes 0 	infrastructure.  Actual ATA storage is cheaper.   				Rob    ------------------------------  # Date: Mon, 18 Aug 2003 21:23:19 GMT % From: "Bob Lail" <Robert.Lail@hp.com>  Subject: Re: 306GB drives!2 Message-ID: <btb0b.2619$3s2.1400@news.cpqcorp.net>  5 "Bill Todd" <billtodd@metrocast.net> wrote in message , news:qZGcnSX5QK6IZ92iU-KYuQ@metrocast.net... > : > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:ay0ZrAQmRiSI@eisner.encompasserve.org...  >  > ...  > 2 > a better comparison would be $235 for a 72 GByte3 > > SCSI drive, and $200 for a 200 GByte ATA drive.  > F > That's almost getting into competitive pricing territory, though.  A 10KrpmK > SCSI drive may still cost a bit more than 3x as much per MB as a 7200 rpm L > ATA drive, but it gives you  a) a higher-quality drive that is more robustE > under less-than-ideal operating conditions (Google up "More than an K > interface - SCSI vs. ATA", a recent paper which I haven't yet had time to I > read but which reportedly is worth it),  b) noticeably better per-drive K > performance (only about 2/3 the average random-access latency),  c) 2 - 3 J > times as many seeking arms for the same capacity (i.e., the per-arm cost isK > fairly close to that of ATA drives, so if your application is arm-limited F > rather than capacity-limited it's close to a wash for cost), and  d)H > consistent availability of performance features such as tagged-commandI > queuing (they may become more standard with the new SATA drives, but so  far I > those new drives aren't as inexpensive as their parallel ATA siblings).  OfK > course, SCSI controllers still cost more than ATA controllers (especially K > parallel ones), and you may want one for each disk (though the likelihood  ofH > a disk failure paralyzing its SCSI bus is smaller than for ATA disks). > J > I had started to think SCSI was on the way out until I realized just howK > inexpensive the lower-end SCSI drives had become.  That may yet happen if K > the new SATA drives turn out to be nearly as robust, but if not there may < > remain an upper-tier niche for people who would rather use singly-mirrored,H > more highly reliable (and faster) drives than doubly-mirrored ATAs (or > multi-parity groups).  >  > - bill >  >  >        Bill  I     Read the paper. It will enlighten you to the real differences between F ATA and SCSI disk drives. They are VERY different and no one should beD advocating replacing SCSI drives with ATA drives in Mission CriticalJ applications unless they know and understand the risks they will be taking with their data.      \Bob Lail       --   Robert G. Lail Senior Solution Architect  Corporate Accounts Hewlett-Packard Company  Voice: 603.424.6272  Email: RobertDOTLail@hpDOTcom  (replace the DOT with ".")   ------------------------------    Date: 18 Aug 2003 16:59:30 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: 306GB drives!3 Message-ID: <PQbhlnMRlYc4@eisner.encompasserve.org>   Z In article <btb0b.2619$3s2.1400@news.cpqcorp.net>, "Bob Lail" <Robert.Lail@hp.com> writes:  
 >     Bill > K >     Read the paper. It will enlighten you to the real differences between H > ATA and SCSI disk drives. They are VERY different and no one should beF > advocating replacing SCSI drives with ATA drives in Mission CriticalL > applications unless they know and understand the risks they will be taking > with their data.   I am not Bill, but what paper ?    ------------------------------  % Date: Mon, 18 Aug 2003 17:12:53 -0700 * From: "Jack Peacock" <peacock@simconv.com> Subject: Re: 306GB drives!, Message-ID: <zN2dnXVFtvYb9tyiXTWJhw@fdn.com>  0 "Bob Lail" <Robert.Lail@hp.com> wrote in message, news:btb0b.2619$3s2.1400@news.cpqcorp.net... > K >     Read the paper. It will enlighten you to the real differences between H > ATA and SCSI disk drives. They are VERY different and no one should beF > advocating replacing SCSI drives with ATA drives in Mission CriticalL > applications unless they know and understand the risks they will be taking > with their data. > K Yes SCSI does have better reliability and queueing, but the real comparison L should be between ATA drives and high (and expensive) capacity tape systems.F When tape cartridges start to approach $100 each a quantity of big ATAL drives begins to look attractive as backup media to supplement archive tape.K For straight sequential transfers in a single task (i.e. the typical backup K while idle scenario: no advantage in cache size or elevator seeks, transfer L rate limited by disk speed instead of interface) ATA drives match Ultra2 andE come pretty close to U160 SCSI in effective transfer rate.  I've been L running a few simple tests on W2K using the Symantec GHOST utility to backupK system images during upgrades and it's far faster than a tape copy to DDS4. L Difference is 500-600 MB/min on U320 to ATA vs. 60-100MB/Min on U320 to DDS4* (separate SCSI channels) with compression.  L Unfortunately the only VMS system with an IDE interface I have at hand is anL old Alphastation 400/233, with an ATA transfer rate that is very competitiveH with TK50s but not much else.  It's a shame DEC/COMPAQ/HP never came outF with a good VMS supported IDE PCI controller board, one that supported DMA-66 and above.     Jack Peacock    ------------------------------  # Date: Mon, 18 Aug 2003 21:38:53 GMT # From: hoff@hp.nospam (Hoff Hoffman) 1 Subject: Call for next FAQ edition, next Freeware 1 Message-ID: <NHb0b.2621$3r2.265@news.cpqcorp.net>   I   I will be pulling together the various and sundry OpenVMS FAQ updates I L   have accumulated since June, and plan to create an edition of the FAQ for J   general release and particularly for inclusion onto the upcoming OpenVMSI   Freeware release -- put another way, if you have or know of any OpenVMS =   FAQ updates, please get them submitted to me by 8-Sep-2003.   I   And if you have any Freeware updates, please get them in by 8-Sep-2003.   ;   OpenVMS Freeware submission information is available via:   *     http://www.hp.com/go/openvms/freeware/  N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------  % Date: Mon, 18 Aug 2003 07:59:00 -0700   From: Jon <jsmyth69@hotmail.com>+ Subject: Re: CSA/CSE Certifications Request 8 Message-ID: <2dq1kvsj971dqe81gbpg9cgdekkravheap@4ax.com>  	 Thanks!     ( "Main, Kerry" <kerry.main@hp.com> wrote:   >>   >> -----Original Message----- / >> From: Mike [mailto:Mike_member@newsguy.com]  ! >> Sent: August 18, 2003 12:17 AM  >> To: Info-VAX@Mvb.Saic.Com >>   >> Hi, >>  @ >>    I'm going to be getting my CSA / CSE for VMS testing soon ? >> and was wondering if anyone could point me to some training  = >> material.  I've got the VMS manuals, but I was hoping for  ; >> some documentation like what they have for the MCSE's.    >>   >> Many thanks!    >>   >> Mike  >>   >>   >  >Mike, > ( >Some training / certification pointers:G >http://h10017.www1.hp.com/certification/region/na/cse/bcs/openvms.html  >  >More background:  > 6 >http://h10017.www1.hp.com/certification/na/index/html: >http://h10017.www1.hp.com/certification/na/migration.html2 >http://www.hp.com/education/sections/openvms.html >http://h71000.www7.hp.com/doc/  >  >Regards >  >Kerry Main  >Senior Consultant >Hewlett-Packard (Canada) Co. " >Consulting & Integration Services >Voice: 613-592-4660 >Fax   : 613-591-4477  >Email: kerryDOTmain@hpDOTcom . >    (remove the DOT's and replace with "."'s)  >OpenVMS DCL - the original .COM   ------------------------------  # Date: Mon, 18 Aug 2003 22:54:59 GMT 1 From: Bob Blunt <robert.blunt@hp.donotspamme.com> & Subject: Re: DEC 3000 M500 CXT error ?1 Message-ID: <7Pc0b.2645$rB2.919@news.cpqcorp.net>    opcom wrote: > Well here's a good one for all the hardware wizards.. And I know someone in this most wise and august group has got to know the answer.  >  > This questions' been floating around out here for a long time.. never found an answer to this -only unanswered questions going back to 1998! So it should be worthy of a challenge, finally put it to rest.  >  > -- > \ > Firing up Ye Olde DEC3000-500 (Flamingo/Pelican), the self test is fine except this error: >  > CXT     ?? 000 0064  >  > I am trying to get the DEC3000 M500 going for free to give to a new VMS hobbyist, so any info on what this cryptic error means, and where I can learn more, would be greatly appreciated. The machine seems to run in spite of this,_ > but I have only just installed VMS, no licences yet etc, so I can't say what it might affect.  >  > I can't find the information on this anywhere, and don't have the manuals on the old turbochannel machine. I don't really know what the CXT is! Hate to have to pitch it, it's got a 19" CRT and all. And this guy NEEDS VMS. He's! > using a MAC for goodness sakes!  >  > -- > M > Please do also reply to me directly as I can't always get to news reliably. T > Please BE SUre to take off the "nounwantedmail_" part of the address to e-mail me. > 
 > Thank you,   > Patrick Jankowiak, > " > VMS Hobbyist Program licensee #13 > Director, The Old Vax Retirement Home, Dallas, TX   I Patrick, as all have said, this error is related to the graphics adapter  I embedded on the Flamingo/3000-500.  I had two "bird" systems (one a 500,  E the other a 500X) that got these errors and ran for a LONG time.  As  C long as you're getting console output to the monitor and you get a  I decent DECwindows startup when booting VMS, you can limp along with this  D for a while.  The biggest nuisance is the system won't boot without D intervention after a powerfail, so someone has to force it past the  error to get it booted.   C You might also try upgrading the firmware on the system, some more  I recent versions mentioned that some CXT errors were fixed there.  If you  F have spare cables and/or monitors that can be swapped, you might give  those a try too.   bob    ------------------------------    Date: 18 Aug 2003 16:31:27 -0700" From: ebayaakkaa@yahoo.com (David)1 Subject: Emulator CDs (console, computer, arcade) = Message-ID: <6fa5abc8.0308181531.6005f1e2@posting.google.com>   E I'm selling Emulator CDs (price 10 euros each). If you are interested @ let me know, I have about 100 CDs, complete collections of every console, computer and arcade.  E-mail to ebayaakkaa@yahoo.com   ------------------------------  % Date: Mon, 18 Aug 2003 13:33:41 -0400 ( From: David Froble <davef@tsoft-inc.com>. Subject: Re: Fast ethernet and DECnet phase IV, Message-ID: <3F410DF5.9010206@tsoft-inc.com>   Dirk Munk wrote:   > Hans Vlems wrote:  > H >> Situation: one Digital Server 5305  with a DE500 combined with a SCSIL >> adapter and one Digital Server 3300 with a single DE500. Both Alpha's run >> VMS 7.3. L >> The adapters were connected to a 10 Mb/s repeater. DECnet, LAT and IP ranD >> fine, never a problem. The 5305 is used to store .ISO files so I  >> bought anH >> Edimax 10/100 switch. The ewa0 devices were changed to fastFD. LANCP 
 >> on bothJ >> nodes reports that the adapters are running at 1000 MHz in full duplex  >> mode.@ >> FTP transfers run flawless, though a little slow, at 30 Mb/s.C >> The problem is with DECnet copies. Even on small files (like the / >> SYS$MANAGER:*.COM files) NCP reports errors. L >> The link is closed, in fact the source node loses connection to the otherI >> machine (adjacancy down). NCP SHOW LINE EWA-0 COUN shows 2 block check 
 >> errors. >> >> Hans  >> >> PS J >> both machines are now down owing to the temperature in the attic; I'll  >> post & >> more details when they're up again. >> >  > K > Cheap unmanaged switches will always default to half-duplex operation if  E > they can not autonegotiate/autosense full duplex with the adapters.  > F > If you buy a chaep switch, try to get one with three leds per port. G > These leds will show how the port operates: speed, data, and full or  J > half duplex operation. The latter led will light up permanently in full J > duplex mode, or flash in case of collisions when running in half duplex  > mode.  >   N An easy check is to set everything to half duplex and monitor things.  If so, Q then either spend the bucks for decent hardware, or accept the significant gains  K in 100baseT in half duplex over 10baseT.  It will depend upon your traffic.   2 Dave, who has fed this particular bug in the past.   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Mon, 18 Aug 2003 20:17:11 +0200 6 From: =?ISO-8859-1?Q?Bj=F6rn_Berg?= <clergyman@gmx.de># Subject: Re: FS: VAXstation 4000/90 9 Message-ID: <bhr58a$29k34$2@ID-144190.news.uni-berlin.de>    Roy Omond wrote: > Bjrn Berg hat geschrieben: ? >> I'm selling my old VAXstation 4000/90. It has the following   > > > You'd probably be best posting this on comp.os.vms where I'm< > sure you'll find someone pretty close to you (I'm assuming& > you're in Germany) with an interest.  E Thank you for that hint and yes you are right I am living in Germany.    Regards, Bjrn    ------------------------------  % Date: Mon, 18 Aug 2003 20:22:35 +0200 6 From: =?ISO-8859-1?Q?Bj=F6rn_Berg?= <clergyman@gmx.de> Subject: FS: VAXstation 4000/90 9 Message-ID: <bhr5ie$29k34$4@ID-144190.news.uni-berlin.de>   
 Hello all,  J I'm selling my old VAXstation 4000/90. It has the following specification:   Processor:	KA49 - 71 MHz Graphics Card:	LCSPX RAM:		64 MB (4 x 16 MB) ) Hard Disk:	RZ26x 1,05 GB (SCSI subsystem)   3 with keyboard and mouse and other necessary cables.   E The VAX comes with no operating system and is technically error-free.   6 You can see pictures of the VAX following these links:( http://www.anubisnet.de/dec/vax_back.jpg) http://www.anubisnet.de/dec/vax_front.jpg   2 I thought about 150,-- EUR (basis for negotiation)  0 If you are interested, please send me an e-mail.   Regards,
 Bjrn Berg   ------------------------------  % Date: Mon, 18 Aug 2003 17:03:23 +0100 - From: Roy Omond <Roy.Omond@BlueBubble.UK.Com> # Subject: Re: FS: VAXstation 4000/90 / Message-ID: <vk1u7kh3bpikaa@corp.supernews.com>    Bjrn Berg hat geschrieben:  > L > I'm selling my old VAXstation 4000/90. It has the following specification: >  > Processor:    KA49 - 71 MHz  > Graphics Card:    LCSPX  > RAM:        64 MB (4 x 16 MB) . > Hard Disk:    RZ26x 1,05 GB (SCSI subsystem) > 5 > with keyboard and mouse and other necessary cables.  > G > The VAX comes with no operating system and is technically error-free.  > 8 > You can see pictures of the VAX following these links:* > http://www.anubisnet.de/dec/vax_back.jpg+ > http://www.anubisnet.de/dec/vax_front.jpg   < You'd probably be best posting this on comp.os.vms where I'm: sure you'll find someone pretty close to you (I'm assuming$ you're in Germany) with an interest.  . The 4000/90 is actually a *very* nice machine.  	 Roy Omond  Blue Bubble Ltd.   ------------------------------  % Date: Mon, 18 Aug 2003 17:52:06 -0400  From: norm.raphael@metso.com  Subject: HP's image is improving? Message-ID: <OF56C69ED0.C1EA458D-ON85256D86.0077E9EF@metso.com>    HP's image is improving K Hewlett-Packard should post its first year-to-year sales increase since the  Compaq merger closed.  August 18, 2003: 1:43 PM EDT- By Paul R. La Monica, CNN/Money Senior Writer     4 NEW YORK (CNN/Money) - Hewlett-Packard is on a roll.    J The stock is up more than 30 percent since the company reported its fiscalK second-quarter earnings in May. HP has won several high profile outsourcing I contracts in recent months. And last week's announcement of more than 150 F new products, many of them printer-related, was widely praised on WallH Street as a sign that HP is not losing its focus on its biggest and most profitable business.    H But can the company deliver a solid fiscal third-quarter report Tuesday?K Hopes are relatively high for the Dow component. Analysts expect HP to post G a profit of 26 cents a share, up from 14 cents in the same quarter last E year, as it continues to cut costs related to its merger with Compaq.     J On the revenue side, Wall Street is predicting that this will be the firstK quarter of year-to-year revenue growth for HP since the Compaq purchase was H completed in May 2002. The consensus sales estimate is $17.5 billion, up/ nearly 6 percent from $16.5 billion a year ago.     K John Rutledge, manager of the Evergreen Technology fund, says that HP could I very well beat the consensus earnings estimate but he does not expect the H company to surpass the $17.5 billion top line projection, mainly becauseD demand in HP's enterprise division, which sells servers, storage andE software to big businesses, remains somewhat sluggish. The enterprise H segment accounted for 21.5 percent of total sales in the second quarter.    ? Despite that, Rutledge does expect continued improvement in the F profitability of the enterprise division thanks to cost cutting. AfterH losing $83 million in the first quarter, the division lost $7 million inI the second. Rutledge hopes the division at least broke even in the third.     B HP is the second-largest holding in the Evergreen Technology fund.    2                          Services and PCs also key    E Wall Street also will be taking a close look at the results from HP's I services division, which accounted for 17 percent of total revenue in the @ second quarter and 25 percent of the company's operating profit.    F Rutledge said that he's impressed by the contract wins, most notably aG 10-year $3 billion outsourcing deal with Procter & Gamble. But he hopes F that HP is not being too aggressive in terms of bidding for deals. "HP/ needs to make money on this business," he said.     I It will be worth noting whether or not operating margins for the services J division in the third quarter dip from the second quarter. Margins were 10J percent in the second quarter, down from 11.6 percent in the first quarterD even though sales increased in the second quarter. HP blamed pricing, pressure in part for the decline in margins.    C On the PC side, HP has posted two consecutive quarters of operating E profitability although it lost the global PC market share lead in the K second quarter to rival Dell, which reported a strong fiscal second quarter C last week. HP's personal systems division, which includes desktops, K notebooks and handhelds, accounted for 28.5 percent of total revenue in the  second quarter.     K Looking out to HP's fiscal fourth quarter, it will be interesting to see if J the company gives concrete guidance for sales in earnings. The company didG not give specific guidance for the third quarter in May, saying instead K that it was on track to post sales of $36.4 billion and earnings, excluding ; charges, of 62 cents for the last half of this fiscal year.     H Analysts expect HP to report revenue of $19 billion in the fiscal fourthK quarter and earnings of 36 cents a share. The company did not say last week I how much it thinks sales of its new products, or for that matter the $300 F million expense from the planned advertising campaign to promote them,/ would affect fourth-quarter sales and earnings.      Find this article at: 7 http://money.cnn.com/2003/08/18/technology/hp/index.htm    ------------------------------  % Date: Mon, 18 Aug 2003 16:31:24 +0100 O From: Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> 6 Subject: Re: IBM Beats HP for Computing Deal with Lego0 Message-ID: <bhqrgc$8qc$1@new-usenet.uk.sun.com>   Mike Naime wrote: . > John Smith <a@nonymous.com> wrote in messageE > news:6xVZa.155972$hOa.31173@news02.bloor.is.net.cable.rogers.com...  > H >>http://story.news.yahoo.com/news?tmpl=story&u=/nm/20030811/bs_nm/tech_
 >>ibm_dc_1 >> >>+ >>IBM Beats HP for Computing Deal with Lego  >>Mon Aug 11,12:55 AM ET >>G >>NEW YORK (Reuters) - IBM Monday said it signed a technology deal with A >>Danish toy maker Lego Co. in which it will replace systems from E >>competitor Hewlett-Packard Co. with pay-as-you-go computing that it 7 >>said will cut Lego's costs during the holiday season.  >>E >>International Business Machines Corp. did not disclose terms of the F >>deal, which involves about 30 computer servers and some data storage
 >>systems. >>A >>Competition between HP and IBM heated up after HP bought Compaq B >>Computer last year and as corporations have cut their technology" >>budgets due to the weak economy. >>E >>HP said this was **one of a "handful" of contract losses related to A >>the Compaq Alpha server**, which is based on proprietary Compaq @ >>technology that was being phased out even before HP bought the
 >>company. >>F >>"We win some, they win some," an HP spokeswoman said, adding that HPG >>recently won a computer server contract over IBM with the China State  >>Tax Authority. >>F >>IBM and Lego in a joint statement said the toy company would replaceF >>230 Hewlett-Packard computers with six large computers from IBM thatD >>run on IBM chips and the popular Unix (news - web sites) operatingD >>system, 24 computers based on microprocessors from Intel Corp. and( >>four very large data storage machines. >>F >>IBM has been pushing its "on-demand" computing model since last yearF >>when Samuel Palmisano took over as chief executive officer. IBM saysG >>companies can save money by paying for only as much computing as they  >>use. >> >>-------------------  >>H >>So the question is......how many new Alpha customers and Alpha systemsC >>sold did HP add to their customer base today to replace Lego as a F >>customer and install count?....oh shit...back to the advertising and >>marketing issue again....  >> >>B >>...just waiting for Fred and Rob to jump on my back for this one >> >> >  > K > Ok, So just exactly what HP product(s) are they phasing out here???  From I > the detail that is not included here, it could be 230 desktop/server PC N > systems that where being replaced.  We are not necessarily talking about VMS > servers here.  > H > I seriously doubt that they where replacing 230 VMS systems with 6 IBMA > systems.  Unless we are talking about some antique VAX systems.  >  >   ? Why. IBM are supplying Lego with 2 P690's these are not desktop = systems. In fact using the measure expoused by HP and many of D its appologists on this newsgroup (TPC-C) the P690 has more capacity: than any current HP server from any of its product ranges.  E Lego were a long time Digital/Compaq customer and it is highly likely F that a reasonable % of the Alpha servers they have are small by modernD standards. The P690's support LPARS and could host a number of small1 applications from relatively small Alpha servers.   F That is assuming Lego aren't also updating their applications as well.  C I am working on a project which is 50% of the way through replacing A ~400 applications (counts vary) running mostly on HP servers with " ~70 applications running on Sun's.   Regards  Andrew Harrison    ------------------------------    Date: 18 Aug 2003 14:17:38 -0500 From: briggs@encompasserve.org; Subject: Re: Memo:  problem converting data type - Macro 32 3 Message-ID: <F$xctWzVAZHw@eisner.encompasserve.org>   d In article <OF37E8BCA9.021C7CEA-ON80256D86.0051F1C2@systems.uk.hsbc>, paul.beaudoin@hsbc.com writes: >  > Folks  > < > What should be a simple exercise, in test appears to fail: > M > I have a need to convert .quad data (times) to f-floats in order to further M > process some data. The VAX instruction set does not provide a converter nor K > is there a specific system routine to do this (that I can find) and so am  > using    LIB$CVT_DX_DX.  > H > In order to test this (code below) I am adding a constant to the inputH > .quad, converting it through the routine and using the VAX instructionL > CVTRFL to convert back to .long and comparing the result with the originalK > input through the debugger.  It doesn't always match. The error is not 32  > bit overflow.   F As you know, quadword integer is capable of representing values in theH range -2^63 to 2^63 - 1 (or 0 to 2^64 - 1 depending on sign convention).B And it is capable of representing all integer values in its range.  ? And, of course, longword integer is capable of representing all ? integer values in its range: -2^31 to 2^31 - 1 (or 0 to 2^32-1)   > F floating is different.  Its range goes from around -10^38 to@ +10^38.  But there's a catch.  It is NOT capable of representing? all integer values in its range.  The F floating mantissa is 24 ? bits in width (counting the free bit from normalization).  That B means that it can exactly represent all integers in the range fromB -2^24 through +2^24.  It can exactly represent all _even_ integers? in the range -2^25 through +2^25.  It can exactly represent all 8 multiples of four in the range from -2^26 through +2^26.  < The first unrepresentable integer is 16,777,217   (2^24 + 1)  9 16,776,855 is exactly representable.  It's less than 2^24 C 16,789,200 is exactly representable.  It's even and less than 2^25. H 16,801,545 is not exactly representable.  It's odd and greater than 2^24   > TST_Q2F\LOUT:   16801546 > TST_Q2F\FRACT:  16801545  F So your conversion from quadword integer to F floating rounds 16801545C up to 16801546 and the conversion from F floating to long is exact. ( The result is that 16801546 != 16801545.  D If you had converted from quad to D floating to long, you would haveC had exact results all the way up to 2^31 (or 2^32 depending on sign - convention).  Then you'd get 32 bit overflow.0  B The approach that I would take with conversions of this sort would@ be to use LIB$ADDX (or lib$SUBX) to subtract some base date fromB the quadword time values.  Now you have a number of 100 nanosecond. "clunks" since your base date in the quadword.  ? Now, use LIB$EDIV to divide this number of clunks by some handyd< multiple.  Maybe 10,000,000 to give you a result in seconds.  I Or, if that result might overflow 31 bits, you could divide by 20,000,000mA and get a result with 2 second resolution.  Or by 100,000,000 and ' get a result with 10 second resolution.   D 1 second resolution will take you about 63 years past your base dateH sign signed arithmetic.  10 second resolution gets you out to 630 years.@ 100 second resolution gets you out to 6300 years.  Pick a number= based on your competing requirements for range and precision.   B You have to use care.  LIB$EDIV requires that the divisor fit into> 32 bits.  So if you want a result in days, you may want to use; LIB$EDIV to divide by 1,000,000,000 and then do an ordinaryfA 32 bit divide by 864.  That still leaves you with overflow issuese# 6300 years out past your base date.7   	John Briggs   ------------------------------  % Date: Mon, 18 Aug 2003 13:49:16 -0400e( From: David Froble <davef@tsoft-inc.com>E Subject: Re: Overlapped IO with redirected SYS$OUTPUT for Sub Programe, Message-ID: <3F41119C.7010101@tsoft-inc.com>   Michael D. Ober wrote:  F > Does anyone have a code snippet (C or BASIC) showing how to redirectG > SYS$OUTPUT for a sub-program called by LIB$SPAWN and have the callinghK > program continue processing.  I can redirect the output and then open theeN > intermediate file after the sub-program completes, but I want to process the= > sub-program's output in parallel to it creating the output.a   No code samples, sorry.s  J I'd need to know a bit more about volumn and what you're doing to be real Q specific, but some concepts would be to use mailboxes, or socket communications, oM or other interprocess communications methods.  A file is so slow compared to iQ those mentioned, but may be appropriate for you requirements.  Regardless, other ,. communications methods would be more flexible.  E If you decide upon mailboxes, I might have some code that would help.y   Dave   -- t4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Mon, 18 Aug 2003 12:33:08 -0600r6 From: "Michael D. Ober" <obermd-@-alum-mit-edu-nospam>E Subject: Re: Overlapped IO with redirected SYS$OUTPUT for Sub Program 1 Message-ID: <FZ80b.555$_U1.53594@news.uswest.net>-  I The subprograms in this case "PRINT" their output, which is why I've beennI using an intermediate file.  Regardless of method, the sub-programs still-G needs to use the basic PRINT statement so I don't have to rewrite them. E Also, the calling program needs to receive the results back in a line-L oriented fashon (LINPUT).  Virtual, Session Table based Mailboxes would haveJ the additional benefit of not having to delete the temporary file when the  calling program is done with it.   Mike.z  5 "David Froble" <davef@tsoft-inc.com> wrote in messagei& news:3F41119C.7010101@tsoft-inc.com... > Michael D. Ober wrote: > H > > Does anyone have a code snippet (C or BASIC) showing how to redirectI > > SYS$OUTPUT for a sub-program called by LIB$SPAWN and have the calling-I > > program continue processing.  I can redirect the output and then open@ thepL > > intermediate file after the sub-program completes, but I want to process then? > > sub-program's output in parallel to it creating the output.o >t > No code samples, sorry.d >tK > I'd need to know a bit more about volumn and what you're doing to be realmB > specific, but some concepts would be to use mailboxes, or socket communications,oK > or other interprocess communications methods.  A file is so slow compared: toL > those mentioned, but may be appropriate for you requirements.  Regardless, othern0 > communications methods would be more flexible. >%G > If you decide upon mailboxes, I might have some code that would help.e >. > Dave >t > -- >6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com8 > T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486 >r   ------------------------------  % Date: Mon, 18 Aug 2003 14:45:07 -0400 0 From: "Brian Tillman" <Tillman@sparkingwire.com>) Subject: Re: Querying UAF from MS Windows $ Message-ID: <3f411eb3$1@news.si.com>  A >I thought he only mentioned seeing the password expiration date.    Doh!  I guess I can't read.d -- aI Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot comn5 Smiths Aerospace  Addresses modified to prevent SPAM.tD 3290 Patterson Ave. SE, MS 1B3 Replace "at" with "@", "dot" with "." Grand Rapids, MI 49512-1991a8        This opinion doesn't represent that of my company   ------------------------------   Date: 18 Aug 2003 13:11:19 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)6, Subject: Re: running a news SERVER under VMS0 Message-ID: <bhqj9n$7it$1@n.ruf.uni-freiburg.de>  w In article <01KZLNXMJ81SAM7Y4A@sysdev.deutsche-boerse.com>, Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes:o7 >Does anyone here actually run a news server under VMS?    Me...>   >If so:a >n& >What news-server software do you run?  $ ANU-News (AFAIK the only one on VMS)  / >How much time is needed for the initial setup?c  I Don't remember, but it was not that much. Let's say a few hours includings compilation.  1 >How much time is needed for routine maintenance?   @ Almost none (wrote a few DCL-procedures to do mainenance tasks).  . >What versions of VMS is this compatible with?  ) Ran under VAX-VMS 6.2 and now under 7.3 .c   >Does it run on VAX and ALPHA?   Yes, no problem.  J >What hardware resources (CPU, memory, disk space, network bandwidth) are ! >needed for successful operation?7  K This depends on the number of newsgroups you want to server. A full feed is E quite consuming (bandwith and disk-space, not so much memory or CPU).    Regards,    Christoph Gartmanno   -- rE  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieA  Postfach 1169                 Internet: gartmann@immunbio.mpg.den  D-79011  Freiburg, Germany 9                http://www.immunbio.mpg.de/home/menue.htmls   ------------------------------  % Date: Mon, 18 Aug 2003 13:25:34 -0700a0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>, Subject: Re: running a news SERVER under VMS' Message-ID: <3f40d3ce$1@cpns1.saic.com>B   Phillip Helbig wrote:c8 > Does anyone here actually run a news server under VMS?   I do.n   > If so: > ' > What news-server software do you run?w  F DNEWS.  Unfortunately, the vendor decided to drop support of VMS some G time ago so there are no new features, support, or bug fixes available.@  F That being said, I also haven't had any issues with it in a long time.  0 > How much time is needed for the initial setup?  F I have run this software for several years so don't remember exactly.  Perhaps an hour or less.  2 > How much time is needed for routine maintenance?   Less than 1 hour a week.  / > What versions of VMS is this compatible with?a  H I started running this on V6.2 and am now running it on V7.3-1.  I have ! only used Alpha systems, however.f   > Does it run on VAX and ALPHA?   B It did.  However, I have just checked their FTP site (something I > haven't done in a long time) and I find that all of their VMS G directories have disappeared.  This means one can no longer get either kH distributions or license keys.  Presumably, this means that my reply is ' also going to be rather useless to you.a  K > What hardware resources (CPU, memory, disk space, network bandwidth) are o" > needed for successful operation?  F It is a very efficient news server although it is something of memory E hog.  It is currently taking about 11,000 pages on my system (that's  G pages, not pagelets).  CPU consumption is minimal (less than 10% on my rG DS10) and the full news feed I take currently consumes 8-10 GB of disk  E space.  DNEWS also supports what it calls a "sucking" feed i.e. only oG take those newsgroups that someone is actually reading.  If configured d/ this way, maybe 200 MB of disk space is needed.t  H It is actually a very good, full featured news server.  Maybe if enough G VMS sites started asking the vendor (www.netwinsite.com) to support it  % on VMS again then perhaps they would.a  
 Mark Berrymanp   ------------------------------   Date: 18 Aug 2003 13:23:13 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)m$ Subject: Strange programming problem0 Message-ID: <bhqk01$7it$2@n.ruf.uni-freiburg.de>   Hello,  L I have a large Pascal program that stopped working, probabely because of theM system upgrade to OpenVMS 7.3 . If I compile the program using /NOCHE/NODEBUG I it simply doesn't run. I get no error message because it is called from anK DECnet task. If I omit these qualifiers the program is working as expected. N Now the strange thing: I added a logfile to the program code (nothing special,I just opening a file and writing out some variables at various steps, then L closing the file again). This modified program doesn't show the problem evenH when compiled with /NOCHECK/NODEBUG. Using /NOOPTIMIZE doesn't help, theO program works only either without the two qualifiers or with the qualifiers andaM the logfile statements. Thus, what could be the cause? What else could I try?w   Regards,    Christoph Gartmannd   -- pE  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452C  ImmunbiologieA  Postfach 1169                 Internet: gartmann@immunbio.mpg.deg  D-79011  Freiburg, Germanyi9                http://www.immunbio.mpg.de/home/menue.htmle   ------------------------------  + Date: Tue, 19 Aug 2003 02:19:11 +0000 (UTC)a7 From: moroney@world.std.spaamtrap.com (Michael Moroney),( Subject: Re: Strange programming problem( Message-ID: <bhs1ev$tsf$1@pcls4.std.com>  5 gartmann@immunbio.mpg.de (Christoph Gartmann) writes:.   >Hello,b  M >I have a large Pascal program that stopped working, probabely because of therN >system upgrade to OpenVMS 7.3 . If I compile the program using /NOCHE/NODEBUGJ >it simply doesn't run. I get no error message because it is called from aL >DECnet task. If I omit these qualifiers the program is working as expected.  G Most likely you have a lingering bug that has either no effect or a badaF effect depending on the environment (often state of the stack).  OftenB called 'heisenbugs'.  They can be rather tough to find.  Sometimes1 increasing the size of arrays can chase them out.t -- a -Mikes   ------------------------------  % Date: Tue, 19 Aug 2003 03:31:10 -0000t! From: Z  <zarlenga@conan.ids.net>e( Subject: Re: Strange programming problem/ Message-ID: <vk36fu1k4v081e@corp.supernews.com>u  4 Christoph Gartmann <gartmann@immunbio.mpg.de> wrote:N : closing the file again). This modified program doesn't show the problem evenJ : when compiled with /NOCHECK/NODEBUG. Using /NOOPTIMIZE doesn't help, theQ : program works only either without the two qualifiers or with the qualifiers andiO : the logfile statements. Thus, what could be the cause? What else could I try?t  4 You almost certainly have bad code and a latent bug.  B Why not just compile it with the qualifiers that work and considerC it fixed?  Else, you need to get neck deep in the code and find thee problem.   ------------------------------   End of INFO-VAX 2003.457 ************************