1 INFO-VAX	Sun, 28 Apr 2002	Volume 2002 : Issue 234       Contents: *FREE* Computer Tips!  Re: Best DIGITAL product ever? Re: Best DIGITAL product ever?: Re: DECUS/EncompassUS Survey -- no longer assuming Windows RE: download VMS Re: download VMS Re: download VMS Re: download VMS' Re: EDIT/FDL changes colours in DECterm  Re: EDT or EVE Re: EDT or EVE Re: EDT or EVE Fix for EDT emulation in EVE  Re: Fix for EDT emulation in EVE- RE: I was right!  Alpha will live in Itanium! - Re: I was right!  Alpha will live in Itanium! / Re: Itanium 2 hits ... Itanium 3 will be Alpha!  Re: Itanium troubles MAIL on VMS and attachments.  Re: MAIL on VMS and attachments.  Re: MAIL on VMS and attachments. Re: Netscape and Mozilla Re: Netscape and Mozilla RE: Openvm 7.3 and DECevent  Re: Purify or equiv. on OpenVMS  Re: Purify or equiv. on OpenVMS P Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For  New Chip P Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For  New Chip P Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For  New Chip P Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip FP Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip FP Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip FP Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip F Re: Security question...  Re: Suggestion: SET DEVICE/RESET test' Re: VMS on http://www.windows-sucks.org C [OT] Fixed Wireless connections to the Internet (was: download VMS)   F ----------------------------------------------------------------------  % Date: Sun, 28 Apr 2002 07:36:57 -0400 3 From: "adrian d'souza" <adrian.dsouza@sympatico.ca>  Subject: *FREE* Computer Tips!7 Message-ID: <P7Ry8.287$sh6.33081@news20.bellglobal.com>   I Hello Everyone.  I want to let you know about a very successful way to do K business on the Internet.  I own my own business and I have discovered that J the best way to market my product is by simply letting my customers see itJ before they buy it.  This concept is called shareware and is available nowH thanks to the Internet.  However,  shareware is only cost-effective whenL your product is information-based.  If you would like to see how this methodH of marketing works you can check out a site that offers FREE information! using this concept.  The site is:   7 http://www.secrets2success.com/special203880/index.html    ------------------------------  # Date: Sun, 28 Apr 2002 07:01:05 GMT * From: "Bill Todd" <billtodd@metrocast.net>' Subject: Re: Best DIGITAL product ever? C Message-ID: <R6Ny8.210644$GS6.19867607@bin3.nnrp.aus1.giganews.com>   9 "Jeremy Barker" <jeremybarker@email.com> wrote in message 7 news:5b86b9ee.0204271538.4f2aeb3f@posting.google.com... # > danco@pebble.org wrote in message ) news:<slrnabi0tl.ijc.danco@pebble.org>... A > > In article <3CB86DA6.3ED3D33A@Free.fr>, Didier Morandi wrote: G > > > I just realized that, since 1957, DEC has produced *many* amazing 	 products:  > > ? > > The Basic+ Run-Time System on RSTS/E.  That was a real gem.  > B > Except that although DEC sold it they didn't produce it.  It was@ > largely produced by an outside company to a DEC specification.  L Hmmm.  I'm pretty sure I remember people in the RSTS group being responsibleK for it in the late '70s, but I suppose the original product could have been  developed elsewhere.   - bill   ------------------------------  % Date: Sun, 28 Apr 2002 06:14:42 -0400   From: John Santos <JOHN@egh.com>' Subject: Re: Best DIGITAL product ever? 6 Message-ID: <1020428055633.26688A-100000@Ives.egh.com>  % On Sun, 28 Apr 2002, Bill Todd wrote:    > ; > "Jeremy Barker" <jeremybarker@email.com> wrote in message 9 > news:5b86b9ee.0204271538.4f2aeb3f@posting.google.com... % > > danco@pebble.org wrote in message + > news:<slrnabi0tl.ijc.danco@pebble.org>... C > > > In article <3CB86DA6.3ED3D33A@Free.fr>, Didier Morandi wrote: I > > > > I just realized that, since 1957, DEC has produced *many* amazing  > products:  > > > A > > > The Basic+ Run-Time System on RSTS/E.  That was a real gem.  > > D > > Except that although DEC sold it they didn't produce it.  It wasB > > largely produced by an outside company to a DEC specification. > N > Hmmm.  I'm pretty sure I remember people in the RSTS group being responsibleM > for it in the late '70s, but I suppose the original product could have been  > developed elsewhere. >  > - bill  E I wasn't here the time, but according to local oral history, a lot of D it was split up between DEC and an outside vendor :-)  For example, B the DH11 driver (part of the monitor) and the lexical analyzer andD the push-pop interpreter, and a little later record I/O (gets, puts,A and field statements) and still later string arithmetic were down > by the outside vendor, while most of the rest was done by DEC.  ! This was in 1973-1975, I believe.   @ DEC/Compaq (or Mentec now) have and have always had the completeD sources and documentation, and have maintained it since then.  SinceF that time, BASIC+ has been pretty stable, with only minor improvementsC and bug fixes, but I think this was mainly due to the fact that any E appreciable new code would have pushed it over 16K, and left only 12K  for user programs.  < It would have been possible to overlay the lexical analysis ? (compile-time) and push-pop interpreter (execution) sections of 7 the RTS to allow a little growth, but there is a lot of = shared code so there would have been a lot of work for little D gain.  Since they were already working on BP2, this must have seemed@ pointless.  Still, it would have been nice if DEC had sprung for@ the couple of $K needed to implement direct fielding of floatingB point and integer variables, since this could have lead quickly to@ a primitive version of records (as per BP2/VAX BASIC/DEC BASIC),& or in other words, user-defined types.   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   Date: 28 Apr 02 13:25:22 +0200) From: p_sture@elias.decus.ch (Paul Sture) C Subject: Re: DECUS/EncompassUS Survey -- no longer assuming Windows ) Message-ID: <htlPMM+msmoy@elias.decus.ch>   c In article <IEnbX4I5irvC@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: B > DECUServe discussion from Richard Copeland quoted some notice ofE > a survey, and for once I liked what I saw, both for privacy and for  > question design: > J >          <<< EISNER::DRA1:[NOTES$LIBRARY]CHAPTERCRACKERBARREL.NOTE;1 >>>7 >                            -< ChapterCrackerBarrel >- P > ==============================================================================P > Note 203.3             New stuff on Encompass web site                  3 of 3P > EISNER::KILGALLEN "Larry Kilgallen, LJK Software"  18 lines  26-APR-2002 07:143 >                                 -< Well Done ! >- P > ------------------------------------------------------------------------------ > 9 >> Please take the Encompass survey regarding the pending ! >> Hewlett-Packard/Compaq merger: / >> http://www.encompassus.org/surveyintro.html   >  > They are getting better: >   D Indeed. It even worked well from Netscape 3.03 on VMS, with Java and Javascript disabled.  N >>>> Please note that instead of using cookies, your ip address will be loggedK >>>> upon taking the survey. This is done for your convenience and privacy.  > D > Furthermore, I tried the survey and did not find any questions for% > which there was no suitable choice.  > @ > I was particularly impressed by the questions that offered the	 > choice:  > 	 5 > 	I am not a current or prospective Windows customer   ' I _really_ enjoyed ticking that one :-)    ------------------------------  % Date: Sun, 28 Apr 2002 06:05:23 -0700 # From: "Tom Linden" <tom@kednos.com>  Subject: RE: download VMS 9 Message-ID: <CIEJLCMNHNNDLLOOGNJICELAENAA.tom@kednos.com>   D David,  have you looked into the Nokia rooftop wireless product.  ItF can go 50 km with directional antenna to the access point.  Moreover, I it uses a cell topology so you can share with your neighbors with simple  ? antennas. I looked into it here but we have way too many trees.    >   -----Original Message-----: >   From: David J. Dachtera [mailto:djesys.nospam@fsi.net]* >   Sent: Saturday, April 27, 2002 8:38 PM >   To: Info-VAX@Mvb.Saic.Com  >   Subject: Re: download VMS  >    >    >   JF Mezei wrote:  >   > [snip]? >   > I do not know what the actual numbers are, but broadband   >   internet access in@ >   > canada has significantly higher penetration rate. Perhaps  >   because a greater D >   > percentage of our population lives in more urban areas served  >   by Cable and the >   > telco's DSL. >   ! >   The problem here is two-fold:  >   I >   1. The cable infrastructure was not installed as two-way capable. The J >   current vendor/owner (AT&T who did not install it) has no incentive to >   upgrade. >   E >   2. Likewise the telco service. There is currently no incentive to ' >   provide remote terminal and DSLAMs.  >   K >   Sprint Broadband Direct is not accepting new customers, AFAIK, which is H >   just as well. This far from Sears tower, coverage is very spotty, at	 >   best.  >   I >   Based on demographic info. available from local municipal websites, I H >   guesstimate there are some 40,000 households waiting for a broadbandG >   option other than satellite in the area surrounding my home (5 mile  >   radius). >   I >   So, unless someone can negotiate with the local municipalities to get F >   fixed wirless towers up and running, satellite is the only option. >    >   --   >   David J. Dachtera  >   dba DJE Systems  >   http://www.djesys.com/ >   , >   Unofficial Affordable OpenVMS Home Page:# >   http://www.djesys.com/vms/soho/  >      ------------------------------  % Date: Sun, 28 Apr 2002 08:19:51 -0600 ( From: emanuel stiebler <emu@ecubics.com> Subject: Re: download VMS + Message-ID: <3CCC0507.83D2AFEC@ecubics.com>    Tom Linden wrote:  > F > David,  have you looked into the Nokia rooftop wireless product.  ItG > can go 50 km with directional antenna to the access point.  Moreover, J > it uses a cell topology so you can share with your neighbors with simpleA > antennas. I looked into it here but we have way too many trees.   : Anybody out here uses the internnet service from DirecTv ? Just curious ...   ------------------------------  + Date: Sun, 28 Apr 2002 16:05:04 +0000 (UTC)  From: david20@alpha2.mdx.ac.uk Subject: Re: download VMS + Message-ID: <aah6jg$126$1@aquila.mdx.ac.uk>    In article <rdeininger-2604022104530001@1cust81.tnt1.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes: 6 >In article <3CC9A56D.828B7C2B@videotron.ca>, JF Mezei& ><jfmezei.spamnot@videotron.ca> wrote: >  >  > D >BOOTP is probably feasible.  Existing systems would need a firmwareI >upgrade (or sidegrade, depending on POV) to support BOOTP.  Once you can I >load the initial boot loader, you can make it do whatever you want.  VMS F >can already use multiple load devices during a single boot.  (VariousJ >components have shipped using the 3rd-party boot mechanism, but it is not
 >well known.)  > D >The hard part does not appear to be booting remotely using some newI >protocol.  Installing or upgrading VMS from a non-random-access, non-ODS H >device would require an overhaul of the current installation mechanism,J >and that is NOT simple.  Again, what does the user community want to give" >up to allow this work to be done?  K A non-random-access , non-ODS device like for instance a TK50 or other tape K device which were the standard way to Install VMS before the widespread use  of CDs ????     
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  # Date: Sun, 28 Apr 2002 16:35:57 GMT 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: download VMS ' Message-ID: <3CCC27CD.59016494@fsi.net>    Tom Linden wrote:  > F > David,  have you looked into the Nokia rooftop wireless product.  ItG > can go 50 km with directional antenna to the access point.  Moreover, J > it uses a cell topology so you can share with your neighbors with simpleA > antennas. I looked into it here but we have way too many trees.   0 Equipment and LOS are only parts of the problem.  E A good chunk of it is getting a wireless ISP up and running. Dunno if G the towns are the stone walls or what. Carol Stream in particular has a H structural height limit of 55 feet - NLOS ala IoSpan would seem the best solution here.  # URL: http://www.iospanwireless.com/    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 28 Apr 02 12:36:25 +0200) From: p_sture@elias.decus.ch (Paul Sture) 0 Subject: Re: EDIT/FDL changes colours in DECterm) Message-ID: <ACem8FAk8zNp@elias.decus.ch>   I In article <3CC8B6E1.906@tgsmc.com>, Brad Hughes <brad@tgsmc.com> writes:  > Phillip Helbig wrote:  > K >> A standard-colour DECterm changes to black background, white foreground  J >> during EDIT/FDL and is not reset after exiting.  Is this a known bug?  + >> (VMS 7.3, most stuff rather up to date.)  >  >  > Try $ set term/noregis first.  > $ I couldn't reproduce Phil's problem.   ------------------------------   Date: 28 Apr 02 13:08:21 +0200) From: p_sture@elias.decus.ch (Paul Sture)  Subject: Re: EDT or EVE ) Message-ID: <ZpFP3lP0UHEu@elias.decus.ch>    In article <00A0D04D.3824112E@SSRL04.SLAC.STANFORD.EDU>, winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") writes: ^ > In article <3CC8EC01.DB472760@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > @ >>has anyone ever converted from EDT to TPU/EVE as main editor ? >   B Yep. Until about 1989 I was frequently cursed with VT52 compatibleK terminals, which TPU/EVE would object to. I moved to TPU/EVE mainly for the K ability to have more than one window visible, so that I could read compiler C listings, program reports _and_ multiple source files from a single 	 terminal.   M > Sure.  (I started on RSX-11 and VMS at the same time; 1985.  I'd been using M > a line-mode editor on ModComp (which I actually reimplemented in Fortran so N > I could add more features; I particularly enjoyed doing the macro processor)$ > for about five years before that.) >  >> >>Was the conversion painful ?   > / > No, not really.  SET KEYPAD EDT helped a lot.  >   J It was painful before that command arrived. I used to carry a tape with myK section file on it. If I wasn't allowed to load my tape, I'd simply reserve C some time to create a section file with sensible keypad settings. I J particularly recall struggling to emulate KP 0 without getting an "alreadyJ at start of line" error message. Does anyone remember the first attempt atE emulating EDT withing EVE? EDT keypad, but no commands - no advantage  whatsoever, to me at least.    >>What prompted it ?   > N > DECterms with displays longer than 24 lines.  EDT couldn't do anything with  > 'em, EVE could.  >   H In my case, the 11/750 I was working on couldn't keep up with the work IK threw at it. I needed every trick I could think of to get my work done in a G timely manner. I ended up buying a pair of genuine VT220s out of my own D pocket for that project, and could still find time to read most of aA quality quality broadsheet newspaper during working hours :-) :-)    ------------------------------   Date: 28 Apr 02 13:12:28 +0200) From: p_sture@elias.decus.ch (Paul Sture)  Subject: Re: EDT or EVE ) Message-ID: <hxZbPnv4r53I@elias.decus.ch>   
 > My original ; > posting suggested the new user learn *both* (or TECO :-).  >   K Exactly. My own preference is to use EVE/TPU, resorting to EDT for things I J know it does better (and sometimes faster). Then TECO for other stuff suchI as editing files such as those output by the COBOL report writer and good  ole RUNOFF.    ------------------------------  # Date: Sun, 28 Apr 2002 16:26:36 GMT 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: EDT or EVE ' Message-ID: <3CCC259F.8877B139@fsi.net>    Paul Sture wrote:  >  > > My originalr= > > posting suggested the new user learn *both* (or TECO :-).  > >  > M > Exactly. My own preference is to use EVE/TPU, resorting to EDT for things IsL > know it does better (and sometimes faster). Then TECO for other stuff suchK > as editing files such as those output by the COBOL report writer and good 
 > ole RUNOFF.   D Oh yeah - you just reminded me: within limits, EDT will handle largeD files. So will TPU; however, TPU will read in the ENTIRE FILE before) returning control of the terminal to you.i  # Can be an inconvenience at times...w   -- A David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/e   ------------------------------  # Date: Sun, 28 Apr 2002 15:31:30 GMT ) From: DCantor@shore.net (David A. Cantor) % Subject: Fix for EDT emulation in EVEu: Message-ID: <mBUy8.39972$sI3.9279014@news02.optonline.net>  ; I published this in the encompaserve DEC-SOFTWARE notesfilepL a couple of years ago, but perhaps people reading this newsgroup might like  it.r   Enjoy.  I The EDT emulator in EVE (activated with SET KEYPAD EDT), doesn't quite do  the trick. e  J In particular, setting the keypad as above alone, you cannot cut an empty N string and then use it to replace a found string (e.g., say you want to delete& all occurences of the string 'fnord').  ; The following code, when placed into your TPU$COMMAND file,tG will modify it (in the privacy of your own process only, of course), sod3 that <SELECT><CUT> will clear the paste buffer and eE <GOLD>num<FIND NEXT><REPLACE> will properly replace occurences of the-$ sought string with the empty string.  C This works by making a local copy of the routine EVE$$STORE_REMOVE,oC modifying it to work "right", and compiling it.  Of course, you can8D save the modified routine to your private section file, if you wish.   Dave Cantoru ex-DECcie, still a Techied    - !------------- code begins here   -----------s     eve_set_keypad_edt;aN eve$x_select_remove_flag := FALSE;      !True if <SELECT><REMOVE> removes charJ                                         ! under cursor.  Set false for EDTL                                         ! behavior: don't remove, just clear7                                         ! paste buffer.h& ! Modify existing basic EVE proceduresB fixupbuf := create_buffer("orig_eve","sys$examples:eve$core.tpu"); set(no_write,fixupbuf,on);D set(system,fixupbuf);   !so we don't find it cycling through buffers position(fixupbuf);  position(beginning_of(C           search_quietly ("procedure eve$$store_remove",forward)));  position(line_begin);  move_horizontal(-1);- erase(create_range(buffer_begin,mark(none))); F position(beginning_of(search_quietly ("eve$clear_message;",forward))); position(line_end);3 split_line;g( copy_text("saved_mark := mark (NONE);"); split_line;<* copy_text("position (the_paste_buffer);"); split_line;8 copy_text("split_line;");i split_line;o$ copy_text("position (saved_mark);");: position(end_of(search_quietly ("endprocedure",forward))); move_horizontal(1);m+ erase(create_range(mark(none),buffer_end));  compile(fixupbuf);4 message("Using local version of EVE$$STORE_REMOVE");   ! -- end of code --t   ------------------------------  # Date: Sun, 28 Apr 2002 16:43:47 GMT 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>i) Subject: Re: Fix for EDT emulation in EVE ' Message-ID: <3CCC29A7.9416D4D7@fsi.net>h   "David A. Cantor" wrote: > = > I published this in the encompaserve DEC-SOFTWARE notesfilemM > a couple of years ago, but perhaps people reading this newsgroup might like- > it.2 >  > Enjoy. > [snip]
 > Dave Cantor  > ex-DECcie, still a Techieo  F By any chance, have you some tips on how to convert EDT's "DEFINE KEY"4 commands (line mode, as found in EDTINI.EDT) to TPU?   -- . David J. Dachterae dba DJE Systemss http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/i   ------------------------------  % Date: Sun, 28 Apr 2002 11:51:46 -0400c+ From: "Main, Kerry" <Kerry.Main@Compaq.com> 6 Subject: RE: I was right!  Alpha will live in Itanium!T Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF401AB1EB8@kaoexc01.americas.cpqcorp.net>   Bill,   H >>> I rather doubt that Oracle positions RAC as a way to improve overallF performance in general, because  1) it isn't and  2) it's considerablyD more complex to configure and tune than a single Oracle instance.<<<  H I suspect you are not that familiar with Oracle database technologies orB you would not be making this statement. However, here is an Oracle, whitepaper on 9i RAC for you to brush up on.  H "Oracle9i Real Application Clusters - Cache Fusion Delivers Scalability"A http://otn.oracle.com/products/oracle9i/pdf/appclusters_cache.pdft  G Bottom line is that most experienced Oracle DBA's will tell you that itoD is only marginally more difficult to manage a Oracle 9i RAC (8i OPS)D configuration vs a standalone database. Some of the advantages, like? shutting down one system/instance without impacting application B availability also adds additional value to the scalability factor.  G As with all application deployment decisions, there are pro's and con'stH that ones needs to consider, but Oracle RAC (OPS renamed) can definitely assist in scaling applications.t  @ >>> One of the main reasons TPC-C is falling out of favor is theD increasing use of the 'cluster cheat' - which of course isn't reallyG cheating, because the rules allow it, but is definitely misleading whenhF used to contend that one's system out-performs some competitor's which isn't using it.<<<  D As with all benchmarks, one should always look beneath the covers toG determine if that benchmark is applicable to their environment. That is F no different than any other benchmark. For those that have partitionedC applications similar to TPC, why would they not be intertested in at? configuration that provided both scaling and added availabilityi
 capabilities?.  C By the way, Sun is also looking very seriously at similar clusteredrA databases for the future as evidenced by their recent purchase ofj CLUSTRA.=20S  
 Reference: http://www.clustra.com/n   Regards,    
 Kerry Main Senior Consultantu Compaq Canada Corp.< Professional Serviceso Voice: 613-592-4660l Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----2 From: Bill Todd [mailto:billtodd@metrocast.net]=20 Sent: April 23, 2002 5:46 PM To: Info-VAX@Mvb.Saic.Comn5 Subject: Re: I was right! Alpha will live in Itanium!Y      6 "Main, Kerry" <Kerry.Main@Compaq.com> wrote in messageH news:BE56C50EA024184DAF48F0B9A47F5CF401AB1EB6@kaoexc01.americas.cpqcorp. net. .=F >>> The GS320 is third behind Fujitsu and HP and only 5% faster than a6 24 CPU IBM P680 despite having to use OPS in a box.<<<  D Yeah, yeah .. I know, vendors should not use parallel based softwareA like OPS (even though the cost is included in the benchmark) on a-+ partitioned system because real men do SMP.-   :-)-  B So go tell it to the TPC Council and convince them. Or go convince= Oracle to stop positioning 9i RAC as a way to improve overallrF performance. Don't continue to harp on this "SMP is the only way to do TPC" stuff here.   ***c  ? Kerry, it's never clear whether you really don't understand the G technical issues behind the spin you spout or simply hope your audienceuE doesn't.  If the former, you're an extremely slow learner, since this ? subject has been covered more than adequately in the past here.o  D I rather doubt that Oracle positions RAC as a way to improve overallF performance in general, because  1) it isn't and  2) it's considerablyG more complex to configure and tune than a single Oracle instance.  Whati@ RAC *is* is a way to improve *scalability*:  when you run out ofE headroom on your SMP system, it gives you somewhere to go to continuecG scaling (though at increased effort and usually - though not with TPC-Cs! - somewhat decreased efficiency).n  D The problem with using TPC-C to compare an N-processor box running aF single Oracle instance with an N-processor box running OPS/RAC is thatD while TPC-C just happens to partition very well (such that it scales; almost linearly with the number of processors in an OPS/RACaF configuration) most real-world applications don't (and even those thatF do require considerably more massaging to achieve that scaling than is> required with single-instance Oracle).  So for many real-worldG applications a Sun N-processor box running a single Oracle instance maybF well run them as fast as, or even faster than, a GS-series N-processorF box running OPS/RAC even if the latter achieves a higher TPC-C rating.  G When you compare single-instance Oracle TPC-C to single-instance OracleeE TPC-C, or RAC TPC-C to RAC TPC-C, you're at least comparing apples to F apples (even though, as always, the best benchmark is your application? itself) - and that's the only way to use this benchmark to ratenH competing systems effectively.  One of the main reasons TPC-C is fallingD out of favor is the increasing use of the 'cluster cheat' - which of@ course isn't really cheating, because the rules allow it, but is< definitely misleading when used to contend that one's system4 out-performs some competitor's which isn't using it.   - bill   ------------------------------  # Date: Sun, 28 Apr 2002 16:59:18 GMTm* From: "Bill Todd" <billtodd@metrocast.net>6 Subject: Re: I was right!  Alpha will live in Itanium!@ Message-ID: <GTVy8.49750$fg.4775308@bin2.nnrp.aus1.giganews.com>  6 "Main, Kerry" <Kerry.Main@Compaq.com> wrote in messageL news:BE56C50EA024184DAF48F0B9A47F5CF401AB1EB8@kaoexc01.americas.cpqcorp.net. .. Bill,   H >>> I rather doubt that Oracle positions RAC as a way to improve overallF performance in general, because  1) it isn't and  2) it's considerablyD more complex to configure and tune than a single Oracle instance.<<<  H I suspect you are not that familiar with Oracle database technologies orB you would not be making this statement. However, here is an Oracle, whitepaper on 9i RAC for you to brush up on.  H "Oracle9i Real Application Clusters - Cache Fusion Delivers Scalability"A http://otn.oracle.com/products/oracle9i/pdf/appclusters_cache.pdf    ***p  K Thanks, but not only do I already have (and have read) that white paper buttK I've also designed and implemented a very similar distributed cache myself,nF so I am somewhat more familiar with the technology than you 'suspect'.  J I meant (as usual) exactly what I said above.  RAC is not a way to improveL overall performance *in general*, but only after an SMP (or perhaps slightlyH NUMA) single-Oracle-instance solution has started to run out of gas:  asJ long as the SMP solution has not started to encounter locking or bandwidthK contention problems, it is a significantly more efficient way to handle therI database.  IOW, RAC enhances not performance per se (if you can get along D well without it, you'll perform better using the same resources) but scalability.  H Furthermore, while Oracle has significantly simplified the configurationL process, for RAC it (unlike the single-instance SMP approach) still requiresL characterization of the differences (CPU performance, memory capacity, etc.)I between the component nodes of the cluster (unless you want the system to K treat your DS10 as a more or less equal partner to your ES or GS box), plusoJ arbitrary amounts of tweaking depending on the degree of tuning one wishesL to perform.  Because (again, unlike the SMP approach) tuning (if you take itJ at all seriously) can be as complex as with a partitioned database:  whileL RAC makes it *possible* to share data between nodes, you're still better offH performance-wise if you arrange to minimize the need for this (as usual,H access localization wins).  Indeed, at least as of a couple of years agoL Oracle was working on configurations that allowed you to scale through threeE stages - single-instance SMP, through clusters of SMP nodes using theCE data-shipping/data-sharing model, to partitioned configurations usinglG function-shipping between separate clusters of SMP nodes:  they weren'tiI doing this for the fun of it, they were doing it because one size (RAC ing0 this case) definitely doesn't fit all uses well.   - bill   ------------------------------   Date: 28 Apr 02 13:51:33 +0200) From: p_sture@elias.decus.ch (Paul Sture) 8 Subject: Re: Itanium 2 hits ... Itanium 3 will be Alpha!) Message-ID: <OtvMyXDyY1HY@elias.decus.ch>l  } In article <3CC967E7.2080101@sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.remove_this.harrison@sun.com> writes:h >  >  > Doc.Cypher wrote:n > 8 >> NOTE: This message was sent thru a mail2news gateway.; >> No effort was made to verify the identity of the sender.P; >> --------------------------------------------------------- >> -9 >> On Fri, 26 Apr 2002, mustang@ucc.asn.au.invalid wrote:. >> ., >>>Bob Ceculski <bob@instantwhip.com> wrote:G >>>: I told you J.F. and Bill "the genius" Todd that Alpha will show uptK >>>: not to far down the road ... no doubt Itanium 3 will be the beginningsi >>>: of Alpha EV8!   >>>s >> rN >> Just where in your latest quoted article does it make any mention of Alpha? >> aJ >> BTW, since you have a surplus of punctuation marks why don't you donate: >> them to another newsgroup where people have a shortage? >>   >>   >> l! >>>Bob, kindly get over yourself.-D >>>If you want to poke fun at these two gentlemen, do it via e-mail. >>>r >> aG >> I for one certainly would not want to encourage that sort of action.U >> l >  >  > < > It seems unfair to single JF and bill out for this kind of9 > SPAM attack, nearly everyone else that has responded to = > Bobs posts on the Alpha IA64 subject seems to disagree with. > him as well. > 6 > The best place for Bob may well be Hyde Park Corner. >  LOL  I'll donate the soapboxi   ------------------------------   Date: 28 Apr 2002 17:13:02 GMT& From: peter@abbnm.com (Peter da Silva) Subject: Re: Itanium troublesT- Message-ID: <aahaiu$l7p@web.eng.baileynm.com>a  ' In article <3CCA2060.38880C49@fsi.net>,c0 David J. Dachtera <djesys.nospam@fsi.net> wrote:G > Even so, you still need to ask yourself: Would FreeBSD be where it is ; > today had it not made certain concessions to "usability"?d  C Why do you phrase that in terms that imply improving usability is a 
 "concession"?f   -- ;+  `-_-'   In hoc signo hack, Peter da Silva.sE   'U`    "A well-rounded geek should be able to geek about anything."nL                                                        -- nicolai@esperi.org          Disclaimer: WWFD?   ------------------------------  % Date: Sun, 28 Apr 2002 15:45:44 +0100 ; From: "Leigh G. Bowden" <LGBowden@bowdenfamily.fsnet.co.uk>e% Subject: MAIL on VMS and attachments. / Message-ID: <aah1ur$beu$1@newsg1.svr.pol.co.uk>m  3 Does VMS mail have the ability to send attachments?e  L I distinctly remember sending a "binary" with VMS one time many years ago or did I just imagine it?   ------------------------------  # Date: Sun, 28 Apr 2002 16:41:27 GMTe1 From: "David J. Dachtera" <djesys.nospam@fsi.net>2) Subject: Re: MAIL on VMS and attachments. ' Message-ID: <3CCC2918.4061CFB2@fsi.net>1   "Leigh G. Bowden" wrote: > 5 > Does VMS mail have the ability to send attachments?c > N > I distinctly remember sending a "binary" with VMS one time many years ago or > did I just imagine it?  E Well, not attachments as you probably think of them. To send non-text  files (within limits):   $ MAIL/FOREIGN  C ...maybe what you're thinking of. As of V7.2, the qualifier remainse) undocumented, AFAIK. Not sure about V7.3.l   -- u David J. Dachtera  dba DJE Systemso http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/r   ------------------------------    Date: 28 Apr 2002 12:43:03 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)w) Subject: Re: MAIL on VMS and attachments.n3 Message-ID: <jt3iJ8RdTQb8@eisner.encompasserve.org>t  m In article <aah1ur$beu$1@newsg1.svr.pol.co.uk>, "Leigh G. Bowden" <LGBowden@bowdenfamily.fsnet.co.uk> writes: 5 > Does VMS mail have the ability to send attachments?y > N > I distinctly remember sending a "binary" with VMS one time many years ago or > did I just imagine it?  ? MAIL/FOREIGN is only documented in the X.25/X.29 documentation,h? for reasons lost in the mists of internal DEC politics.  But ite works wherever VMSmail works.t  @ Of course this is not the same format as the "attachment" format@ used by SMTP, so you should have a VMS system at both ends.  ButA anyone with taste in operating systems will already have made thet! right decision in that regard :-)l   ------------------------------   Date: 28 Apr 02 12:43:10 +0200) From: p_sture@elias.decus.ch (Paul Sture)t! Subject: Re: Netscape and Mozillav) Message-ID: <0srQa5Kj6TUK@elias.decus.ch>t  W In article <25APR200221531183@gerg.tamu.edu>, carl@gerg.tamu.edu (Carl Perkins) writes: 7 > "David J. Dachtera" <djesys.nospam@fsi.net> writes...- > }John Reagan wrote:2 > }> Bart Zorn wrote:  > }> > John Reagan wrote:D > }> >I > }> >> No Marvel here, but as others have mentioned, I have a reasonablesJ > }> >> machine, with reasonable memory, with additional tuning for "large& > }> >> memory footprint applications" > }> tJ > }> > I certainly can't afford such a configuration. AND I still have theO > }> > feeling that the balance between functionality and required resources ise > }> > completely lost.g > }> > > }> > Bart Zorn > }> > > }> t > }> No argument here. > } G > }Hhmmm... Sounds like that old "affordability" thing rearing its uglyc > }head again... > }  > }--  > }David J. Dachtera > ? > In this particular case, it sounds more like a poorly writteno > piece of software. > C > If you need to let it use >200MB of RAM to work well, my question @ > is what the heck is is actually doing with all that memory? DoB > you know how much that is? It is a LOT. You can probably fit theE > entire managers section (and maybe the general section, too) of theiG > full VMS documentation set using plain uncompressed ASCII, along withoE > the diagrams at screen resolution, into 200MB. (And I think many ofi" > us know just how big that is...) > F > Basically, I expect it is a piece of junk. Unfortunately, so is mostE > of the other software out there. Probably the most common thing you C > can hear me say about software is things along the lines of "what " > idiot wrote this piece of junk"? >   H I tend to agree with the "piece of junk" label. Here on my 512MB system, I only have ~15% memory free.    ------------------------------  % Date: Sun, 28 Apr 2002 17:52:44 +0200e From: Dirk Munk <munk@home.nl>! Subject: Re: Netscape and Mozillas& Message-ID: <3CCC1ACC.2010407@home.nl>   Tom Linden wrote:) >  >>  -----Original Message-----) >>  From: Dirk Munk [mailto:munk@home.nl]o+ >>  Sent: Wednesday, April 24, 2002 1:24 PMe >>  To: Info-VAX@Mvb.Saic.Comr% >>  Subject: Re: Netscape and Mozillai >> >> >>  Tom Linden wrote:( >>  >e" >>  >>  -----Original Message-----A >>  >>  From: Hoff Hoffman [mailto:hoffman@xdelta.zko.dec.nospam] - >>  >>  Sent: Tuesday, April 23, 2002 6:18 PMf! >>  >>  To: Info-VAX@Mvb.Saic.Coms) >>  >>  Subject: Re: Netscape and Mozillas >>  >> >>  >>F >>  >>  In article <CIEJLCMNHNNDLLOOGNJIKEPHEMAA.tom@kednos.com>, "Tom( >>  >>  Linden" <tom@kednos.com> writes:J >>  >>  :Installed Netscape 3.03 on 7.1-2 AXP and Mozilla 0.9.9 on 7.3 AXP	 >>  >>  :ML >>  >>  :Two thumbs down for each.  Not ready for prime time and so far from? >>  >>  :it, that it probable isn't salvageable.  Certainly notp >>  useable.  Upon5 >>  >>  :launching Netscape it gets Javacript errors!2 >>  >>C >>  >>    Netscape 3.03 is so ancient, it nicely avoids many of the 
 >>  obnoxiouseK >>  >>    new problems, um, features of the newer web browsers.  That said, @ >>  >>    there are a number of holes in the Javascript, and the >>  HTML standardi >>  >>    is ancient.y >>  >>* >>  >>  :Mozilla not even worth reviewing. >>  >>K >>  >>    Mozilla works nicely if you have sufficient memory and processor. K >>  >>    (Given that it is largely based on Java, this is not a surprise.)lK >>  >>    If you don't have circa 256 MB (or more) and an EV56 (or better),-I >>  >>    then Mozilla and CWSB are unfortunately rather poor choices forKI >>  >>    daily use -- Mozilla and CWSB (and Java) are significantly moreuE >>  >>    resource-intensive than were previous generations of tools.i >>  >>B >>  >>  :I would certainly like to see Compaq entice Opera to port >>  >>  there browser. >>  >>= >>  >>    Empirical evidence indicates the Opera folks do notb >>  answer their mail.C >>  >>    (I also don't have buckets of money to throw into this or  >>  >>  other portingO >>  >>    projects, however.)  >>  >l >>  >eC >>  > I know we have discussed this before, but I think that IE6 isR >>  an excellentC >>  > browser, and Opera is even better.  Now Opera has been portedr >>  to systemsA >>  > that have far less presence than VMS, like BeOS, Mac, OS/2,  >>  QNX, Symbian< >>  > Linux, Solaris and of course Windows.  So why not VMS? >>  >g@ >>  > On another note, I think the ubiquitous use of explorer in >>  Windows is aH >>  > big forward leap,  not going to totally replace DCL or Bash, but a= >>  > marvelous tool.  I am sure you and your colleagues havet >>  looked into this,a	 >>  > buto> >>  > probably dismissed cause all you had to use was Mozilla. >>  Can't be thatt@ >>  > difficult or costly to write one from scratch if you can't >>  buy one, liken? >>  > Opera.  My point is that a browser is more than a desktopo >>  nicety for theL >>  > internet explorer (no pun intended)  it is an interface to a computing >>  > environment,A >>  > and as heretical as it may sound to this group, in my view,o >>  Microsoft hasfA >>  > definitely done something right, and I think the have set ac >>  new standard,a
 >>  > like >>  > it or not. >>  >e >>  >f >>  >e >>  >  >>  >i8 >>  >>   ---------------------------- #include <rtfaq.h>% >>  >>  -----------------------------p; >>  >>        For additional, please see the OpenVMS FAQ --e >>  >e >>  > www.openvms.compaq.com0 >>  >  --------------------------- pure personal) >>  > opinion ---------------------------k5 >>  >    Hoff (Stephen) Hoffman   OpenVMS Engineering  >>  hoffman#xdelta.zko.dec.com >>  >  >>4 >>  I think we can come to the following conclusion: >>I >>  If your alpha is not to ancient, if you have a fairly decent graphics G >>  adapter, if you have at least 256 MB memory, if you have tuned your:K >>  system and DecWindows a bit, THEN Mozilla is a very usable browser (andt? >>  the only modern browser there is on VMS). A few more helpern >>  applications >>  would be nice though.t >  > K > OK, but you don't have to do any of that when you are running on W2K. Them9 > machine I installed on was a stock PWS 433au with 512MBc  F Of course you don't have to do these things with W2k. W2k is a single H user system, it does not know quotas etc. Every application can take as I much resources as it likes. VMS is multiuser, so you will have to adjust o9 system and user settings for your particular environment.eF As I wrote before, i'm using a system very similar to yours. The only H difference is processor speed (500 Mhz with me) and maybe graphics card  (4D20 with me).i  F On my system Mozilla runs fine, so if it doesn't run all that well on / your system you may have to check the settings.h       >  > K >>  You like IE6 and Opera more, and that's ok. I personnaly dislike any M$ I >>  stuff, if only for the brand name. Ever since a DIR command on MS-Dosf? >>  gave me the free space on the disk, and not the size of thec >>  directory, IK >>  could not believe that a company like that could produce anything good.o= >>  I have not been disappointed in that view over the years.a >  >  > No argument there  >  > K >>  And the good things they do have were mostly stolen, bought, or copied.i >>K >>  Opera has a good reputation. It has been recommended to test web pages.oH >>    If it works well with Opera, it will work with other browsers too. >>K >>  But I am really interested in what you are missing in Mozilla, and whatf= >>  you do find in IE6 and Opera. Perhaps you can give us fewe >>  examples, thatF >>  would make things a bit more clear for us. So if it is not to much >>  trouble, tell us !!t >  > K > I would love to spend the time to do a comparison but I am sure there ares > peopleK > better at it than I.  To me its just a tool, and if I have a better one Ih > will use.eF > it.  I think the whole layout is ugly, its too busy, I would go more > hierarchical.cM > Installing plugins is a bit more of a challenge.  In some ways, its like ana	 > editor.c; > I use emacs and dislike vi others would say the opposite.aH > I will try to give more constructive comments as the occassion arises. >  >  >>  Regards, >> >>  Dirk >> >> >  >    ------------------------------  % Date: Sun, 28 Apr 2002 12:15:02 -0400C+ From: "Main, Kerry" <Kerry.Main@Compaq.com> $ Subject: RE: Openvm 7.3 and DECeventT Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4023D8F0B@kaoexc01.americas.cpqcorp.net>   Hello -n  
 Re: DECevent-m  8 The following url might be of interest:(note doc's link): http://www.support.compaq.com/svctools/decevent/index.html   regardsr  
 Kerry Main Senior Consultant  Compaq Canada Corp.u Professional Servicesd Voice: 613-592-4660e Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----5 From: G. W. Hairston [mailto:ghairst@mcvh-vcu.edu]=20  Sent: April 24, 2002 3:02 PM To: Info-VAX@Mvb.Saic.Com.  Subject: Openvm 7.3 and DECevent    G Please give me insight on installing Decevent under openvms 7.3. We are1! new at using openvms.  Thank you.r   ------------------------------    Date: 28 Apr 2002 07:07:38 -0000= From: Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> ( Subject: Re: Purify or equiv. on OpenVMS6 Message-ID: <20020428070738.18803.qmail@gacracker.org>  C On Sun, 28 Apr 2002, "John E. Malmberg" <wb8tyw@qsl.network> wrote:n >Doc.Cypher wrote: >>  3 >>      /standard=ansi89/prefix=all/warn=enable=all  >>  L >>    and you'll save yourself a lot of trouble later on.  It will also make >>    your code more portable. o >i  >For newer releases of Compaq C: >l >/warn=enable=(level4) >r >And I prefer: >l  >/warn=enable=(level4,questcode)   Thanks!d  F Should the /standard and /prefix lines remain or should it just be the /warn?    G On a related note I've got a couple of people doing a little work on aneK updated version of PGP for VMS. Its based on 2.6.3 and supports a number ofrH new algorithms plus key lengths up to 8K bits. Only problem is that thisI has uncovered a problem in it, and the currently available 2.6.3 version,aE it doesn't check signatures properly. Its been years since I used thetI debugger on VMS with C and I'm not seeing anything obvious plus the otherhI people working on it have no experience of this, if anyone else can spares5 some time to have a look at it I would appreciate it.a     Doc. --  6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                             https://vmsbox.cjb.netl   ------------------------------  # Date: Sun, 28 Apr 2002 17:49:40 GMTu- From: "John E. Malmberg" <wb8tyw@qsl.network>i( Subject: Re: Purify or equiv. on OpenVMS* Message-ID: <3CCC5218.4070001@qsl.network>   Doc.Cypher wrote:,F  > On Sun, 28 Apr 2002, "John E. Malmberg" <wb8tyw@qsl.network> wrote:  >>m#  >> For newer releases of Compaq C:s  >>a  >> /warn=enable=(level4)b  >>p  >> And I prefer:u  >>f#  >> /warn=enable=(level4,questcode)t  >  >
  > Thanks!  >E  > Should the /standard and /prefix lines remain or should it just be 
  > the /warn?y  B I generally let the /standard default to /standard=RELAXED_ANSI89.+ Stricter compliance should not hurt though.u  H The /prefix=all is required by a lot of code.  It is used for resolving D routines in the Compaq C RTL that are not specified by the compiler 
 standards.   -Johno wb8tyw@qsl.network Personal Opinion Only    ------------------------------  # Date: Sun, 28 Apr 2002 08:23:11 GMT * From: "Bill Todd" <billtodd@metrocast.net>Y Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For  New Chip a@ Message-ID: <PjOy8.45516$fg.4409541@bin2.nnrp.aus1.giganews.com>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3CCBACAE.B9BC9BEA@videotron.ca... > Bill Todd wrote:G > > So unless Hammer tanks, Intel will have to field something similar.@ >aI > The wildcard here is compilers. Intel has Digital's compiler engineers.  DoesE > AMD have experienced compiler people to develop compilers that will8 maximize! > the Hammer chip's performance ?n  H Probably not.  But Microsoft will produce the required Windows compilersL (which should be as good as their IA32 counterparts, since the technology isH similar) and the open-source people have already started producing a GNUL compiler (which will likely be as bad as the existing IA32 GNU compiler, but0 that's never seemed to discourage its use much).  J x86-64 is far from the radical departure from existing compiler technologyK that EPIC is.  And if Intel *does* actually adopt AMD's instruction set (asaK Microsoft appears to be pressuring it to), then Intel's (and anyone else's)rE Yamhill compilers can be used for Hammer as well:  while there may beaG second-order optimizations due to differences in internal architecture,a? identical instruction sets should result in similar first-order> optimizations.   - bill   ------------------------------    Date: 28 Apr 2002 06:56:40 -0700( From: bob@instantwhip.com (Bob Ceculski)Y Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For  New Chip t= Message-ID: <d7791aa1.0204280556.25cc8d2d@posting.google.com>M  a JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message news:<3CCBACAE.B9BC9BEA@videotron.ca>...  > Bill Todd wrote:G > > So unless Hammer tanks, Intel will have to field something similar.  > N > The wildcard here is compilers. Intel has Digital's compiler engineers. DoesN > AMD have experienced compiler people to develop compilers that will maximize! > the Hammer chip's performance ?e  H not even dec's compiler engineers will be able to save itanic w/epic ...C I read a retiring dec engineers white paper on the Q site before itaG was pulled a few years back and he stated that compilers were epic weakiH point and why it failed in the early 80's ... they had estimated becauseG of that risc would always outperform epic on a compiler basis, and morenE importantly felt parallel (EV8) was the way to go ... eventually theytI will come to the same conclusions and have to throw that boat anchor off!    ------------------------------    Date: 28 Apr 2002 06:58:25 -0700( From: bob@instantwhip.com (Bob Ceculski)Y Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For  New Chip M= Message-ID: <d7791aa1.0204280558.6723042d@posting.google.com>s  ` "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3CCB6613.4809B716@fsi.net>... > Bob Ceculski wrote:  > >  > > Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> wrote in message news:<20020427065618.10491.qmail@gacracker.org>...v? > > > On 26 Apr 2002, bob@instantwhip.com (Bob Ceculski) wrote:S > > >t > > > <snip> > > >D > > > >disclaimer:P > > > >The opinions expressed by myself on this board do not necessarily reflectN > > > >those of other self appointed vms experts on this board.  I am the onlyN > > > >poster on this board who is totally clueless ... all other posters knowL > > > >all the future holds for vms, and should not be confused with myself. > > > N > > > Look, it's not a board. It is a NEWSGROUP, or for some people, a mailing. > > > list. Please get your terminology right. > > >tO > > > As to your comments about "other self appointed vms experts", well I findoP > > > what they post far more plausible. Stating that Alpha will live in ItaniumQ > > > suggests that Itanium will either change to using the Alpha instruction setjQ > > > or the EPIC core will be replaced with the RISC core from Alpha. Neither ofrM > > > these seems likely to happen. What does seem likely is that some of theLJ > > > non-specific features from the Alpha processor will be grafted on toJ > > > Itanium. This will not make the damn thing an Alpha processor. Okay? > > > 
 > > > Doc. > > J > > well I say you are wrong ... I say for Intel to stay competitive, theyI > > will have to drop the epic boat anchor ... are you saying there is noiL > > chance at all of this happening?  And the last time I checked, the firstJ > > ammendment was still in force ... or has communism already got to thisK > > board?  Should I refer to you as heir doc and Bill Todd as heir genius?  > 
 > Um, Bob? > ' > Do you know anyone who speaks German?- > I > I don't speak it myself, but the little I know includes this: "heir" iseD > an Angloid extraction, generally meaning one who stands to inheritI > (whatever). The word you probably want is "herr", roughly equivalent todF > "mister", "sir", or perhaps "brother", depending upon the supporting > context... > F > ...to the best of my extremely limited knowledge. Perhaps one of theJ > German posters can offer a better explanantion, if they are so inclined.   I'm part german ...l   ------------------------------  # Date: Sun, 28 Apr 2002 07:45:58 GMTl* From: "Bill Todd" <billtodd@metrocast.net>Y Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip F>B Message-ID: <WMNy8.246613$K5.19440897@bin5.nnrp.aus1.giganews.com>  J "Doc.Cypher" <Use-Author-Supplied-Address-Header@[127.1]> wrote in message/ news:20020427215315.5056.qmail@gacracker.org...k   ...o  K > As to Intel dropping EPIC, that would involve a serious loss of face, notD/ > to mention a serious write-off of investment.i  J Indeed - as well as serious anger from those they have convinced to devoteI their own efforts to supporting it.  At this point, given that it appearsiK that the basic McKinley design will be used for Madison/Deerfield and their B follow-on Montecito, the cost of continuing at least through thoseL generations (i.e., until at least 2005) is relatively small, so Itanic wouldI likely have to fail spectacularly and incontrovertibly before Intel wouldtK actually drop it rather than just consider pursuing other possibilities fori the 2005 and later time-frame.    I for one do not think thatF > is particularly likely if Itanic can be given reasonable performance > improvements.a  I The main problem seems to be about the next 3 years, since likely nothingeI much can be done before 2005.  By some time in 2005, Intel should be ableyK both to ship chips with the same kind on-chip memory glue that EV7, Hammer,-A and (though I wasn't aware of it) USIII have and the same kind of@H multi-processor glue that EV7, Hammer, and POWER4 have, and also by thatJ time continued process advances may have made the McKinley core small (and1 low-power) enough to place two of them on a chip.n  I That won't create a world-beater, but it may allow the product to surviveRA against US and POWER, which may be the only significant remaining>G competition by then.  If whoever Alpha's owner is keeps up with processmJ shrinks EV7 *might* still be performance-competitive in 2005, but with theK prospect that it will only fall increasingly behind beyond that time it maysE not be very attractive (one could imagine a time when shrinks allowedkJ multiple entire EV7 images, not just cores, to be placed on a single chip,K but it's not clear that Alpha's owner will still have even that engineering> capability left).f  J But Hammer is still the wild card in all this.  Intel will *have* to matchE Hammer if Hammer succeeds in covering the market right up through theuI mid-range (i.e., everything that all but the largest current server boxes I do:  even an 8-processor Hammer system should, after the 2003H2 shrink to G 0.09 micron process technology and consequent speedup, be able to offer J about the same performance as any 32-processor box you can purchase today,B and Fred Weber's October MPF presentation slides included diagramsI indicating that dual-core-on-a-chip Hammers were planned for, which wouldyF presumably then support 16-processor Hammer systems which would be theC equivalent of today's 64-processor servers).  Hammer is 'disruptivemI technology':  while Itanic is the last of the boutique (at least comparednJ with 'commodity'), high-end 64-bit micros, Hammer is the first of the realK commodity 64-bit micros, and will force all the boutique players (includingu& Itanic) farther and farther up-market.  J So unless Hammer tanks, Intel will have to field something similar.  GivenK that this product (call it Yamhill) will extend Intel's reach to the bottombH of the high end, it's not at all clear that Intel will feel the need forH *any* yet-higher-end product - which is why the outlook for a redesignedG Itanic with Alpha internals is not good, even if without it Itanic willnL continue to be a dog.  Rather than throw a lot more money at a significantlyI smaller high-end niche, Intel it may just allow Itanic to wallow along in J the background until all its passengers have jumped ship, while it focusesJ its real efforts on Yamhill:  it has already gotten badly burned trying toI shoot the moon once in the high-end space, back when the potential returnpG for succeeding was considerably greater than it seems likely to be now.    - bill   ------------------------------    Date: 28 Apr 2002 06:04:38 -0700( From: bob@instantwhip.com (Bob Ceculski)Y Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip F < Message-ID: <d7791aa1.0204280504.72ae0e2@posting.google.com>  z Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> wrote in message news:<20020427215315.5056.qmail@gacracker.org>...; > On 27 Apr 2002, bob@instantwhip.com (Bob Ceculski) wrote:yK > >Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> wrote in message 5 > >news:<20020427065618.10491.qmail@gacracker.org>...t> > >> On 26 Apr 2002, bob@instantwhip.com (Bob Ceculski) wrote: > >>   > >>  K > As to Intel dropping EPIC, that would involve a serious loss of face, not-K > to mention a serious write-off of investment. I for one do not think that-F > is particularly likely if Itanic can be given reasonable performance > improvements.l >  > Doc.  C Lose face?  I think they are past that phase ... didn't you read mynF other post from the inquirer about itanium 2?  That is because itaniumB 1 like Mike suggested sunk!  If they don't come up w/chip to beat B hammer and perform better than power in the high end, they will be) losing more than face ... try their jobs!    ------------------------------  % Date: Sun, 28 Apr 2002 04:02:57 -0400s- From: JF Mezei <jfmezei.spamnot@videotron.ca> Y Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip F , Message-ID: <3CCBACAE.B9BC9BEA@videotron.ca>   Bill Todd wrote:E > So unless Hammer tanks, Intel will have to field something similar.   L The wildcard here is compilers. Intel has Digital's compiler engineers. DoesL AMD have experienced compiler people to develop compilers that will maximize the Hammer chip's performance ?r   ------------------------------  # Date: Sun, 28 Apr 2002 16:28:53 GMTl1 From: "David J. Dachtera" <djesys.nospam@fsi.net>sY Subject: Re: put up or shut up (was: RE: AMD To Announce Microsoft Support For New Chip Fa' Message-ID: <3CCC2628.C2CDC51E@fsi.net>n   Bob Ceculski wrote:  > b > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3CCB6613.4809B716@fsi.net>... > > Bob Ceculski wrote:d > > >m > > > Doc.Cypher <Use-Author-Supplied-Address-Header@[127.1]> wrote in message news:<20020427065618.10491.qmail@gacracker.org>...aA > > > > On 26 Apr 2002, bob@instantwhip.com (Bob Ceculski) wrote:e > > > >  > > > > <snip> > > > >t > > > > >disclaimer:R > > > > >The opinions expressed by myself on this board do not necessarily reflectP > > > > >those of other self appointed vms experts on this board.  I am the onlyP > > > > >poster on this board who is totally clueless ... all other posters knowN > > > > >all the future holds for vms, and should not be confused with myself. > > > >rP > > > > Look, it's not a board. It is a NEWSGROUP, or for some people, a mailing0 > > > > list. Please get your terminology right. > > > >aQ > > > > As to your comments about "other self appointed vms experts", well I findeR > > > > what they post far more plausible. Stating that Alpha will live in ItaniumS > > > > suggests that Itanium will either change to using the Alpha instruction setrS > > > > or the EPIC core will be replaced with the RISC core from Alpha. Neither of O > > > > these seems likely to happen. What does seem likely is that some of thecL > > > > non-specific features from the Alpha processor will be grafted on toL > > > > Itanium. This will not make the damn thing an Alpha processor. Okay? > > > >i > > > > Doc. > > >oL > > > well I say you are wrong ... I say for Intel to stay competitive, theyK > > > will have to drop the epic boat anchor ... are you saying there is no2N > > > chance at all of this happening?  And the last time I checked, the firstL > > > ammendment was still in force ... or has communism already got to thisM > > > board?  Should I refer to you as heir doc and Bill Todd as heir genius?  > >  > > Um, Bob? > >e) > > Do you know anyone who speaks German?/ > >oK > > I don't speak it myself, but the little I know includes this: "heir" isCF > > an Angloid extraction, generally meaning one who stands to inheritK > > (whatever). The word you probably want is "herr", roughly equivalent to H > > "mister", "sir", or perhaps "brother", depending upon the supporting > > context... > >bH > > ...to the best of my extremely limited knowledge. Perhaps one of theL > > German posters can offer a better explanantion, if they are so inclined. >  > I'm part german ...-  < I'm all Polish, but I know extremely little of the language.   -- - David J. Dachterar dba DJE Systems] http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 28 Apr 02 12:27:57 +0200) From: p_sture@elias.decus.ch (Paul Sture)d! Subject: Re: Security question...l) Message-ID: <s1bqHny2sXsl@elias.decus.ch>l  c In article <Z1CAGx7rDsW7@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:fk > In article <8b51ed8.0204251254.1d2efcd9@posting.google.com>, vmendham@altavista.com (Vic Mendham) writes: ? >> We're having a discussion at work about log file security... G >> Can a user (system or root) manipulate a logfile without anyone ever ! >> knowing about it. If so how???  >   H Perhaps the following text is appropriate. It is included in the regular* security reports my Linux system sends me.   "DISCLAIMER   I Please note that these security checks are neither complete nor reliable.eF Any attacker with proper experience and root access to your system can deceive *any* security check!"  I > A root user cannot do that, if by root user you mean someone who does > F not know enough about VMS to know that the term is "privileged" user.  >  p >>E I don't think so cause the revised date on the file would be changed  F >> and it should indicate the user who last accessed the file. If the I >> file is open, you can't edit it or copy it as this is a file access ( 1 >> sharing ) violation.  >  lH > Privileges beat all.  Physical access beats privileges.  You might as H > well combine those and not allow any privileged access to the machine H > outside the locked computer room.  Besides auditing access within the F > computer room, video recording should be used to watch for abnormal F > physcial activity.  And to consider the human factor make sure that D > nobody is ever in the computer room unaccompanied.  That means at I > least three people on each shift (to allow for breaks).  Best practice oK > would be to rotate the schedule so the same team is not working together h > all the time.   J Your security guy is not the only one concerned. I once witnessed an auditB inspection carried out by an insurance company at a customer I wasK visiting. Having noticed that I was signed into the computer room they also J checked out who I was and whether I was supervised. That audit was useful,I as they discovered kitchens with deep fat friers above the computer room,S and a petrol line too.   ------------------------------   Date: 28 Apr 02 12:35:12 +0200) From: p_sture@elias.decus.ch (Paul Sture)r) Subject: Re: Suggestion: SET DEVICE/RESETo) Message-ID: <fIpUAWxIoV49@elias.decus.ch>s  [ In article <3CC8B8B3.D323E800@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:e  > mckinneyj@cpva.saic.com wrote: >> e* >> In article <3CC77402.E91255B5@fsi.net>,6 >> "David J. Dachtera" <djesys.nospam@fsi.net> writes:# >> > mckinneyj@cpva.saic.com wrote:b >> >>  >> [...snip...]l >> >> > M >> >> > I recently discovered - these drives and their younger cousins (TZ88,aL >> >> > etc.) do this to the HSJs, as well. Ya gotta delete them and add 'em; >> >> > back before youy can use 'em again after a failure.a >> >> >e >> >> > :-(G >> >> >eL >> >> It's been a while so I forget all the details, but, in the instance ofJ >> >> a tape drive hung off an HSJ you can frequently recover use of it byL >> >> issueing a SET SECURITY command targeting the device. Try changing theL >> >> the protection mask to something else and back again. A side effect ofL >> >> this is that a bit in some status mask gets cleared. I don't recall ifD >> >> the bit is within one of the UCB fields or the ORB. Anyway, myI >> >> recollection is that the bit represented "damaged" hardware. Sorry,-I >> >> I can't remember all the details (it's been a while); don't know if@) >> >> this would help with JF's TK drive.: >> >K >> > My problem was that the HSJ reported the device as "misconfigured - nosL >> > device at PTL" or something. VMS simply reported it as off-line. So, no7 >> > matter what you did, the HSJ still saw it as AWOL.9 >> > >> o= >> Yep, the HSJ can't serve it if it doesn't know what it is.n > H > But, *THAT* is the problem! Up until it went south, HSJ knew perfectlyE > well what it was! Is HSJ making brash assumptions because the drived? > "dropped ready" temporarily? ...or asserted its fault signal?a >  [snip]  D Simple to reproduce. Try writing to a cleaning tape and the HSJ will@ then report that the device is misconfigured. The simple way outB is to delete it (and any references to it) at the HSJ prompt, then/ RUN CONFIG, then add back the references to it.i   ------------------------------  % Date: Sun, 28 Apr 2002 17:40:09 +0200w From: Dirk Munk <munk@home.nl>
 Subject: testr& Message-ID: <3CCC17D9.8010001@home.nl>   testing my news settings   ------------------------------   Date: 28 Apr 02 13:50:19 +0200) From: p_sture@elias.decus.ch (Paul Sture)-0 Subject: Re: VMS on http://www.windows-sucks.org) Message-ID: <8quwlP8SjiAl@elias.decus.ch>R  d In article <aabpqk$8ifs$1@lead.zk3.dec.com>, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> writes:N > Frankly, with a site name of "windows-sucks" it's not likely that Compaq (orH > many companies, except perhaps Sun) would directly contribute to it...' > interested users on the other hand...h > 8 The proxy servers at work won't let me look at that url.  > But it is probably the sort of site which has the potential of? introducing VMS to new users, who could after all be tomorrow's A purchasers. Even bringing the existence of VMS to their attention  is something worthwhile IMHO.y   ------------------------------    Date: 28 Apr 2002 11:37:21 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) L Subject: [OT] Fixed Wireless connections to the Internet (was: download VMS)3 Message-ID: <fgqq8GQ9EZtx@eisner.encompasserve.org>   [ In article <3CCB7167.1F842306@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:t   > The problem here is two-fold:. > G > 1. The cable infrastructure was not installed as two-way capable. The H > current vendor/owner (AT&T who did not install it) has no incentive to
 > upgrade. > C > 2. Likewise the telco service. There is currently no incentive toh% > provide remote terminal and DSLAMs.o > I > Sprint Broadband Direct is not accepting new customers, AFAIK, which ismF > just as well. This far from Sears tower, coverage is very spotty, at > best.c > G > Based on demographic info. available from local municipal websites, I F > guesstimate there are some 40,000 households waiting for a broadbandE > option other than satellite in the area surrounding my home (5 milee
 > radius). > G > So, unless someone can negotiate with the local municipalities to get D > fixed wirless towers up and running, satellite is the only option.  C http://towerstream.com describes an outfit providing fixed wirelessgB in New England.  It is not certain if they will stay around longerA than others who have tried, but I thought it interesting that thes/ approach is still being worked by some vendors.t   ------------------------------   End of INFO-VAX 2002.234 ************************