1 INFO-VAX	Thu, 16 Nov 2000	Volume 2000 : Issue 641       Contents:8 %RMS-W-RTB, 3373 byte record too large for user's buffer< Re: %RMS-W-RTB, 3373 byte record too large for user's buffer< Re: %RMS-W-RTB, 3373 byte record too large for user's buffer" Alpha 8KB vs 64KB Memory Page Size& Re: Alpha 8KB vs 64KB Memory Page Size& Re: Alpha 8KB vs 64KB Memory Page Size( Alphaserver 4100 + KGPSA (Fiber Channel), Re: Alphaserver 4100 + KGPSA (Fiber Channel), Re: Alphaserver 4100 + KGPSA (Fiber Channel), RE: Alphaserver 4100 + KGPSA (Fiber Channel)E Re: AS 255/233, which graphics card do I have without opening the box < Booting standalone backup from disk (MicroVAX cluster, DSSI) DEC 5305 question  Re: DEC 5305 question  Re: f$search improvement ? file transfer VMS to unix 9 Re: Forcasting the Tropical Year - the Laskar expression? 9 Re: Forcasting the Tropical Year - the Laskar expression? 5 Re: FREE OpenVMS Windows NT intergration book *NOT*!! ) FW: [Fwd: RE: Technology of US elections] - Re: FW: [Fwd: RE: Technology of US elections] - Re: FW: [Fwd: RE: Technology of US elections] - Re: FW: [Fwd: RE: Technology of US elections] 9 Gartner: Where =?iso-8859-1?q?=B4s_OpenVMS_=3F=3F=3F=3F?= = Re: Gartner: Where =?iso-8859-1?q?=B4s_OpenVMS_=3F=3F=3F=3F?=  Idle process Killer  Re: Idle process Killer  Re: Idle process Killer  LDAP Re: LDAP) Re: Looking for RJ45-to-ThinWire adapter.  Looking for ROM bootstrap info. # Re: Looking for ROM bootstrap info.  Mime attachment  Re: Mime attachment + Re: open VMS backdrop for motif or xwindows + Re: Pascal calling convention from C++ code + Re: Pascal calling convention from C++ code + Re: Pascal calling convention from C++ code D PC programs in OpenVMS (was: Re: How can I watch RealPlayer movies?)0 Removing trailing spaces from a 1700 byte record4 RE: Removing trailing spaces from a 1700 byte record4 Re: Removing trailing spaces from a 1700 byte record4 Re: Removing trailing spaces from a 1700 byte record4 Re: Removing trailing spaces from a 1700 byte record4 Re: Removing trailing spaces from a 1700 byte record; Re: RMS way to get long sequential records in multiple $GET . Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed. Re: Severity of eBay's ecache problem detailed! Re: SHOW PROC (Peak virtual size) $ Re: Still looking for JNET software. strange SMTP problem Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: TCP/IP and load balancing  Re: Technology of US Elections Re: Technology of US Elections RE: Technology of US elections RE: Technology of US Elections re: Technology of US Elections RE: Technology of US Elections RE: Technology of US Elections Re: Technology of US Elections Re: Technology of US elections Re: Technology of US elections Re: Technology of US elections Re: Technology of US Elections Re: Technology of US elections Re: Technology of US Elections Re: Technology of US Elections Re: Technology of US Elections Re: Technology of US Elections RE: Technology of US elections Re: Technology of US Elections Re: Technology of US elections Re: Technology of US elections Re: Technology of US elections Re: Technology of US Elections Re: Technology of US elections Re: Technology of US Elections Re: Using PIPE for UCX RE: Using PIPE for UCX  What is ASYNCWRTER system crash?& Where are Compaq's Technical Journals?* Re: Where are Compaq's Technical Journals?) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections] ) Re: [Fwd: RE: Technology of US elections]  [Q] MadGoat FTP   F ----------------------------------------------------------------------  % Date: Thu, 16 Nov 2000 09:35:04 -0600 ) From: "The Newbie" <billames@accunet.net> A Subject: %RMS-W-RTB, 3373 byte record too large for user's buffer / Message-ID: <t17v51hfpn4r93@corp.supernews.com>   F Im assuming I have to set my user buffer to a greater number.  I tried looking for it in AUTHORIZE 5 but I'm not for sure which setting I should increase.    Another Dumb Question,  
 The Newbie   Here are the specifics:     # SYSDEVICE:[BAMES.COMPILE] @FIND_DIR 8 %RMS-W-RTB, 3373 byte record too large for user's buffer  4  RTB,  'nnn' byte record too large for user's buffer  7   Facility:     RMS, OpenVMS Record Management Services   L   Explanation:  A record returned from a $GET operation is too large for theI                 user buffer provided. The status value (STV) field of the J                 RAB contains the size of the record that is too large; theL                 returned record is truncated to the size of the user buffer.  A   User Action:  Correct the program by providing a larger buffer.      Here are my user stats:   A Username: BAMES                            Owner:  SYSTEM MANAGER B Account:  SYSTEM                           UIC:    [40,0] ([40,0])2 CLI:      DCL                              Tables: Default:  SYS$SYSDEVICE:[BAMES]  LGICMD:  Flags:  DisPwdDic DisPwdHis # Primary days:   Mon Tue Wed Thu Fri + Secondary days:                     Sat Sun  No access restrictionsD Expiration:            (none)    Pwdminimum:  4   Login Fails:     0> Pwdlifetime:           (none)    Pwdchange:   7-AUG-2000 08:22> Last Login: 16-NOV-2000 08:35 (interactive), 16-NOV-2000 08:47 (non-interactive) 9 Maxjobs:         0  Fillm:       200  Bytlm:       100000 9 Maxacctjobs:     0  Shrfillm:      0  Pbytlm:           0 9 Maxdetach:       0  BIOlm:       150  JTquota:       8192 9 Prclm:          10  DIOlm:      4096  WSdef:         1024 9 Prio:            4  ASTlm:      4096  WSquo:         2048 9 Queprio:         0  TQElm:        20  WSextent:      4096 9 CPU:        (none)  Enqlm:      2000  Pgflquo:     200000  Authorized Privileges:I   ACNT         ALLSPOOL     ALTPRI       AUDIT        BUGCHK       BYPASS H   CMEXEC       CMKRNL       DIAGNOSE     DOWNGRADE    EXQUOTA      GROUPH   GRPNAM       GRPPRV       IMPERSONATE  IMPORT       LOG_IO       MOUNTI   NETMBX       OPER         PFNMAP       PHY_IO       PRMCEB       PRMGBL H   PRMMBX       PSWAPM       READALL      SECURITY     SETPRV       SHMEMJ   SYSGBL       SYSLCK       SYSNAM       SYSPRV       TMPMBX       UPGRADE   VOLPRO       WORLD Default Privileges: I   ACNT         ALLSPOOL     ALTPRI       AUDIT        BUGCHK       BYPASS H   CMEXEC       CMKRNL       DIAGNOSE     DOWNGRADE    EXQUOTA      GROUPH   GRPNAM       GRPPRV       IMPERSONATE  IMPORT       LOG_IO       MOUNTI   NETMBX       OPER         PFNMAP       PHY_IO       PRMCEB       PRMGBL H   PRMMBX       PSWAPM       READALL      SECURITY     SETPRV       SHMEMJ   SYSGBL       SYSLCK       SYSNAM       SYSPRV       TMPMBX       UPGRADE   VOLPRO       WORLD   ------------------------------  % Date: Thu, 16 Nov 2000 15:35:35 -0000 * From: "Richard Brodie" <R.Brodie@rl.ac.uk>E Subject: Re: %RMS-W-RTB, 3373 byte record too large for user's buffer + Message-ID: <8v0us8$sae@newton.cc.rl.ac.uk>   ^ "The Newbie" <billames@accunet.net> wrote in message news:t17v51hfpn4r93@corp.supernews.com...? > Im assuming I have to set my user buffer to a greater number.   % > SYSDEVICE:[BAMES.COMPILE] @FIND_DIR : > %RMS-W-RTB, 3373 byte record too large for user's buffer  D Probably not: you often hit this from exceeding the DCL command lineB maximum length. I assume you are typing @FIND_DIR and have changedI your prompt. Maybe you have a logical pointing at something else, or have 8 mangled the file somehow. DIR/FULL  is your friend here.   ------------------------------   Date: 16 Nov 2000 16:13:00 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)E Subject: Re: %RMS-W-RTB, 3373 byte record too large for user's buffer 6 Message-ID: <8v112c$q1e$1@mailint03.im.hou.compaq.com>  [ In article <t17v51hfpn4r93@corp.supernews.com>, "The Newbie" <billames@accunet.net> writes: G :Im assuming I have to set my user buffer to a greater number.  I tried  :looking for it in AUTHORIZE6 :but I'm not for sure which setting I should increase.  F   The RTB error indicates an application buffer has been overrun, not C   a system parameter nor process quota.  This means that the buffer C   is too small for the application and the application buffer needs F   to be expanded to read in the record, or it means that an incorrect D   or corrupted input file is involved and the application is getting
   bogus data.   B :  User Action:  Correct the program by providing a larger buffer.  H   I'll pass along a request that the RTB message be updated to indicate G   that the RTB error can also be caused by a bogus input file, and that G   this message is an application or an application file problem and not .   a system parameter or process quota problem.    $ :SYSDEVICE:[BAMES.COMPILE] @FIND_DIR9 :%RMS-W-RTB, 3373 byte record too large for user's buffer   D   Use SET VERIFY or similar to figure out which DCL command "blew", ;   then look at the application and at the associated files.   N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------   Date: 16 Nov 2000 15:53:05 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)+ Subject: Alpha 8KB vs 64KB Memory Page Size 6 Message-ID: <8v0vt1$psi$1@mailint03.im.hou.compaq.com>  G   All currently-available OpenVMS Alpha systems have an eight kilobyte  E   (8KB) memory page size.  (For comparision, the VAX memory page size E   is 512 bytes.)  The memory page size is the minimum unit of memory  I   allocation, and the value also obviously determines the page alignment  H   for global sections, lib$get_vm_page, COMMONs, and device driver page    allocations.  C   The Alpha Architecture and the OpenVMS documentation have always  F   indicated that the system memory page size can range from 8KB up to B   an architectural limit of sixty-four kilobytes (64KB), and that C   application code should NOT assume that the OpenVMS Alpha memory  &   page size is and will always be 8KB.  E   Why?  By having larger memory pages, more virtual and more physical F   memory can be addressed.  With 8KB memory pages, 43 bits of virtual E   address and 45 bits of physical address space are available.  With  >   64KB memory pages, these values are increased to 55 and 48, A   respectively.  Put another way, this increase in the page size  A   increases the virtual addressing from eight terabytes (8TB) to  D   thirty-two petabytes (32PB), and with a three-bit increase in the E   Alpha system's theoretical physical memory addressing limits: from  ?   the current theoretical 32TB physical addressing up to 256TB.   C   If you are responsible for OpenVMS Alpha application code, please G   check your code to ensure that it does not require nor assume an 8KB  I   page alignment, nor an 8KB page size.  Linker options, global sections, J   COMMONs, anything that tries to pack page-based data structures tightly F   together, assumptions of the amount of data that will fit on a page,I   and potentially kernel-mode routines that require pages be locked into     physical memory.  Etc.  I   Use of the existing sys$getsyi/f$getsyi PAGE_SIZE itemcode is strongly  E   recommended, as this allows your code to dynamically determine the  I   memory page size.  The hard-coding of the memory page size is strongly  H   discouraged.  If you must assume a memory page size, please assume at     least a 64KB memory page size.  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Thu, 16 Nov 2000 11:53:37 -0500 2 From: rdeininger@mindspring.com (Robert Deininger)/ Subject: Re: Alpha 8KB vs 64KB Memory Page Size L Message-ID: <rdeininger-1611001153370001@user-2ivea3d.dialup.mindspring.com>  [ In article <8v0vt1$psi$1@mailint03.im.hou.compaq.com>, hoffman@xdelta.zko.dec.nospam wrote:   I >   All currently-available OpenVMS Alpha systems have an eight kilobyte   >   (8KB) memory page size.    I believe Hoff is telling us that there is hardware in the pipeline with a page size bigger than 8 KB.  Probably, that hardware is seeing light at the end of the tunnel.   W Or maybe Hoff has been listening to little birds singing, and just felt inspired... :-)    --   Robert Deininger rdeininger@mindspring.com    ------------------------------  % Date: Thu, 16 Nov 2000 12:33:12 -0500 5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> / Subject: Re: Alpha 8KB vs 64KB Memory Page Size , Message-ID: <8v161s$752j$1@lead.zk3.dec.com>  L Give that man a cookie.  Note that there is plenty of time, and at least theH initial machines that will provide it will allow the size to be selected, between the current, and a larger page size.  C As an aid, the simplest way for a user-mode application to do it is  something like:    #include "lib$routines.h"  #include "syidef.h"   	 int size; # lib$getsyi(&SYI$_PAGE_SIZE, &size);    for the size in bytes.  3 Drivers and inner mode code use... mmg$gl_page_size       % Robert Deininger wrote in message ... 7 >In article <8v0vt1$psi$1@mailint03.im.hou.compaq.com>, $ hoffman@xdelta.zko.dec.nospam wrote: > I >>   All currently-available OpenVMS Alpha systems have an eight kilobyte  >>   (8KB) memory page size. > K >I believe Hoff is telling us that there is hardware in the pipeline with a K page size bigger than 8 KB.  Probably, that hardware is seeing light at the  end of the tunnel. > H >Or maybe Hoff has been listening to little birds singing, and just felt inspired... :-)  >  >--  >Robert Deininger  >rdeininger@mindspring.com   ------------------------------  % Date: Thu, 16 Nov 2000 08:58:29 -0300 ) From: fabio_compaq@ep-bc.petrobras.com.br 1 Subject: Alphaserver 4100 + KGPSA (Fiber Channel) L Message-ID: <OF5CF46016.20554E9B-ON03256999.0041428D@ep-bc.petrobras.com.br>  K I am planning to migrate my storage from StorageWorks to EMC in a few weeks  - Ok you dont agree   but it was a corporate decision.  C I am running OpenVMS  7.1-1H2 and 7.2 in two Alphaserver 4100 5/600  servers. (non clustered)  H I am reading in the http://www.openvms.compaq.com/openvms/fibre that the KGPSA is compatible K only with OpenVMS 7.2-1. Is this true ? I must upgrade my operating systems  to use a Fiber Channel	 adapter ?   E My other option is use the KZPBA adapter, this I know will work fine.    Regards   
 Fabio Cardoso    ------------------------------  # Date: Thu, 16 Nov 2000 11:34:05 GMT % From: Uwe Zessin <zessin@my-deja.com> 5 Subject: Re: Alphaserver 4100 + KGPSA (Fiber Channel) ) Message-ID: <8v0gnd$3ln$1@nnrp1.deja.com>   7 In article <OF5CF46016.20554E9B-ON03256999.0041428D@ep-  bc.petrobras.com.br>, ,   fabio_compaq@ep-bc.petrobras.com.br wrote: > C > I am planning to migrate my storage from StorageWorks to EMC in a @ > few weeks - Ok you dont agree but it was a corporate decision. > E > I am running OpenVMS  7.1-1H2 and 7.2 in two Alphaserver 4100 5/600  > servers. (non clustered) > F > I am reading in the http://www.openvms.compaq.com/openvms/fibre thatA > the KGPSA is compatible only with OpenVMS 7.2-1. Is this true ?     Which one? KGPSA-BC or KGPSA-CA?B It is likely that you will be using the 64-bit adapter (KGPSA-CA).C Then, yes, you need V7.2-1 plus ECOs to be able to use this adapter  in a supported configuration.   < > I must upgrade my operating systems to use a Fiber Channel > adapter ?   > Yes. V7.1-x anyway. V7.2 is no longer in 'standard support' or: how this called and I'm sure it (even after installing all. available ECOs) does not support the KGPSA-CA.  G > My other option is use the KZPBA adapter, this I know will work fine.g  D That is Ultra SCSI, not "Fibre Channel" (note the spelling - "Fiber" is just the physical cable).  D You need to work _with_ EMC to make sure this works - it sounds like" you are planning _without_ them ?!   --
 Uwe Zessin    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Thu, 16 Nov 2000 09:55:22 -0300 ) From: fabio_compaq@ep-bc.petrobras.com.br 5 Subject: Re: Alphaserver 4100 + KGPSA (Fiber Channel) L Message-ID: <OFE0D0F6DE.EEC5D6EE-ON03256999.0046CF5F@ep-bc.petrobras.com.br>   Hi  J I know KBPSA is a SCSI  :-) and I am planning it with EMC too. I am trying to collect much more0 information to this upgrade in case of problems.   Regardso   FC                  6 Uwe Zessin <zessin@my-deja.com> em 16/11/2000 09:34:05L                                                                             L                                                                             L                                                                                 >                                                               >                                                               >                                                               >  Para:    Info-VAX@Mvb.Saic.Com                               >                                                               >  cc:      (bcc: Fabio dos Santos Cardoso/E-P-BC/Contratada)   >                                                               >                                                               >                                                               >  Assunto: Re: Alphaserver 4100 + KGPSA (Fiber Channel)        >                                                                             7 In article <OF5CF46016.20554E9B-ON03256999.0041428D@ep-r bc.petrobras.com.br>, ,   fabio_compaq@ep-bc.petrobras.com.br wrote: > C > I am planning to migrate my storage from StorageWorks to EMC in as@ > few weeks - Ok you dont agree but it was a corporate decision. >lE > I am running OpenVMS  7.1-1H2 and 7.2 in two Alphaserver 4100 5/600y > servers. (non clustered) > F > I am reading in the http://www.openvms.compaq.com/openvms/fibre thatA > the KGPSA is compatible only with OpenVMS 7.2-1. Is this true ?     Which one? KGPSA-BC or KGPSA-CA?B It is likely that you will be using the 64-bit adapter (KGPSA-CA).C Then, yes, you need V7.2-1 plus ECOs to be able to use this adapteri in a supported configuration.T  < > I must upgrade my operating systems to use a Fiber Channel > adapter ?t  > Yes. V7.1-x anyway. V7.2 is no longer in 'standard support' or: how this called and I'm sure it (even after installing all. available ECOs) does not support the KGPSA-CA.  G > My other option is use the KZPBA adapter, this I know will work fine.C  D That is Ultra SCSI, not "Fibre Channel" (note the spelling - "Fiber" is just the physical cable).  D You need to work _with_ EMC to make sure this works - it sounds like" you are planning _without_ them ?!   --
 Uwe Zessin    & Sent via Deja.com http://www.deja.com/ Before you buy.h   ------------------------------  % Date: Thu, 16 Nov 2000 12:00:41 -0500S: From: "Koska, John C. (LNG-MBC)" <John.C.Koska@bender.com>5 Subject: RE: Alphaserver 4100 + KGPSA (Fiber Channel)hK Message-ID: <3D35AD137AAAD411A6BA0008C7B1B12D05BAF7@MBCALBEXC03.BENDER.COM>E   > -----Original Message-----. > From: Uwe Zessin [mailto:zessin@my-deja.com]+ > Sent: Thursday, November 16, 2000 6:34 AMS > To: Info-VAX@mvb.saic.com 7 > Subject: Re: Alphaserver 4100 + KGPSA (Fiber Channel)o > 9 > In article <OF5CF46016.20554E9B-ON03256999.0041428D@ep-I > bc.petrobras.com.br>,E. >   fabio_compaq@ep-bc.petrobras.com.br wrote: > > E > > I am planning to migrate my storage from StorageWorks to EMC in a?B > > few weeks - Ok you dont agree but it was a corporate decision. > > G > > I am running OpenVMS  7.1-1H2 and 7.2 in two Alphaserver 4100 5/600  > > servers. (non clustered) > >FH > > I am reading in the http://www.openvms.compaq.com/openvms/fibre thatC > > the KGPSA is compatible only with OpenVMS 7.2-1. Is this true ?h > " > Which one? KGPSA-BC or KGPSA-CA?D > It is likely that you will be using the 64-bit adapter (KGPSA-CA).E > Then, yes, you need V7.2-1 plus ECOs to be able to use this adapter  > in a supported configuration.N  iG You should, if using the 64-bit adapter, make sure the firmware on the g adapter is correct.  See,   H http://www.compaq.com/products/storageworks/adapters/64bitpcifibre.html   E Without the correct firmware in addition to the ECOs that Uwe Zessin uD alludes to, there can be problems; even in a Compaq supported fibre  channel configuration.    
 Good Luck, :) jck     ------------------------------  % Date: Thu, 16 Nov 2000 18:19:02 +0000o- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>eN Subject: Re: AS 255/233, which graphics card do I have without opening the box( Message-ID: <3A142516.73B17AE@bbc.co.uk>   Fred Kleinsorge wrote:  H > My guess is that you have a Mach-64 ISA graphics card (the most common > ISA card out there for VMS). > A > Check the IO prefixes (SYSMAN IO SHO PREFIX) and make sure that B > DECW is one of them.  Do a SYSMAN IO AUTO/FULL to see what ICBMs? > are being invoked - and make sure DECW$ICBM is being invoked.e >u  @ I checked the ICBM thing (memories from a past life). Its there.   $ mcr sysman io sh prefix   2 %SYSMAN-I-OUTPUT, command execution on node xxxxxxO %SYSMAN-I-IOPREFIX, the current prefix list is: DECW$,DECW$GRAPHICS_,SYS$,DECW$l   >Y9 > If so, check that DECW$AUTOCONFIG.DAT is in SYS$MANAGER  >r  
 Its is indeed    >e1 > If so, edit the file and add a line: DEBUG = -1   ) I did that, I guess I have to reboot now.   M As I mentioned, it was VMS 6.2-1H3 that I had DECWIndows working under. MaybeeO I need to use ISACFG to make DECWindows know about the graphics device? Is that1, a program on a floppy, or a console command?   Regards    >: >  > And try the SYSMAN command.a >cA > Tim Llewellyn wrote in message <3A117B4B.54384BFA@bbc.co.uk>...  > >  > >: > >"Mark D. Jilson" wrote: > > L > >> That appears to be the sound card,  here is an excerpt from the DSNlinkL > >> article - [Motif, MME] Troubleshooting the AUA0: Device on AlphaStation > >> Systems > >> > >P0 > >OK, sounds plausible, here is the full output > >  > >Adapter Configuration:  > >----------------------LM > >TR Adapter     ADP      Hose Bus   BusArrayEntry  Node Device Name / HW-IdON > >-- ----------- -------- ---- -------------------- ---- -------------------- > ----- . > > 1 KA0D02      80DB7840    0 BUSLESS_SYSTEM# > > 2 PCI         80DB7A40    0 PCI I > >                                   80DB7D68  PKA:    6 NCR 53C810 SCSI @ > >                                   80DB7DA0          7 SATURNI > >                                   80DB7EB8  PKB:   12 NCR 53C810 SCSI D > >                                   80DB7F28  EWA:   14 NI (Tulip)# > > 3 ISA         80DB8240    0 ISADK > >                                   80DB8418          0 EISA_SYSTEM_BOARDRB > >                                   80DB8450          1 ISA--VGA? > >                                   80DB8488          2 PCXBJE$ > > 4 XBUS        80DB87C0    0 XBUS> > >                                   80DB8998          0 MOUS= > >                                   80DB89D0          1 KBD-E > >                                   80DB8A08  TTA:    2 Serial PortiM > >                                   80DB8A40  TTB:    3 NS16450 Serial Port F > >                                   80DB8A78  LRA:    4 Line Printer > (parallel po > >rt)@ > >                                   80DB8AB0  DVA:    5 Floppy > >SDA>n > >a% > >The graphics is ISA-VGA, I guess ?u > >  > >>M > >>    'The AlphaStation 200 series systems come with a small version of thefK > >>    Microsoft Sound card plugged in to a single ISA slot off the systemoM > >>    module. This ISA slot is separate from the bus I/O Card that contains  > >> theL > >>    PCI and EISA/ISA Bus slots. This card shows up under the SRM Console	 > >> SHOWwM > >>    CONFIG table as the following device.  The option number is PBXJA-AA.N > >>G > >>    ISA Slot  Device   Name     Type       Enabled   BaseAddr   IRQc > >> DMAH > >>    --------  ------   -------  -----      -------   --------   ---- > >> ---F > >>    2        0        PCXBJ    Singleport   Yes      530         9 > >> 0'e > >> > >> Tim Llewellyn wrote:r > >> > > >> > Fred Kleinsorge wrote:s > >> >N > >> > > You need to find out the what the device is.  Is it a PCI device?  If > it is,L > >> > > then a CLUE CONFIG in SDA will tell you the device ID in HEX - from	 > that weoL > >> > > can tell you what it is - or at least if it is supported.  If it is > an ISAI > >> > > device, then you'll need to open up the box and figure it out byx > looking atL > >> > > the board... although for most known ISA graphics, you just need to > do the. > >> > > ISACFG for a device ID of "ISA--VGA". > >> > >N > >> > > The console "should" tell you something if you do a SHOW CONFIG -- at > leasth' > >> > > something like "VGA Graphics".  > >> > > > >> >G > >> > Thanks Fred and others, I'd forgotten about SHOW CONFIG from thei
 > console, > >> > I was using SHOW DEV. > >> >: > >> > I'll take the system down later on, for now I have: > >> >( > >> >   3 ISA         80DB8240    0 ISA= > >> >                                    80DB8418          0I > EISA_SYSTEM_BOARDrF > >> >                                    80DB8450          1 ISA--VGAC > >> >                                    80DB8488          2 PCXBJo > >> >: > >> > so it looks like an ISA card of some type (PCXBJ?). > >> >
 > >> >  --= > >> > Tim Llewellyn, OpenVMS Infrastructure, Remarcs Projecta7 > >> > MedAS at the BBC, Whiteladies Road, Bristol, UK.cH > >> > Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk > >> >H > >> > I speak for myself only and my views in no way represent those of > >> > MedAS or the BBC. > >> > >> --eK > >> Jilly   - Working from Home in the Chemung River Valley - Lockwood, NY K > >>         - jilly@clarityconnect.com                      - Brett Bodinea > fannL > >>         - Mark.Jilson@Compaq.com                        - since 1975 or > so8 > >>         - http://www.jilly.baka.com               - > >. > >--h9 > >Tim Llewellyn, OpenVMS Infrastructure, Remarcs Projecta3 > >MedAS at the BBC, Whiteladies Road, Bristol, UK.nD > >Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk > >4D > >I speak for myself only and my views in no way represent those of > >MedAS or the BBC. > >  > >d   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukn  A I speak for myself only and my views in no way represent those ofm MedAS or the BBC.l   ------------------------------  % Date: Thu, 16 Nov 2000 09:51:15 -0800c% From: Doug Kiner <dkiner@pacbell.net>hE Subject: Booting standalone backup from disk (MicroVAX cluster, DSSI)b+ Message-ID: <3A141E93.8C8025E1@pacbell.net>   D I've created a standalone backup kit on a non-system DSSI disk in my? cluster (DIA5) and want to boot a MicroVAX 3400 from that kit..6  E Confirming that I need only enter  ">>> BOOT DIA5:"  (No registers orsF other special qualifiers required) to boot from the kit on that drive?  H Also, will I need to a >>>SET BOOT to change the boot device back to theH (original) system disk once I am done with my backup/restore operations?   TIA    ------------------------------   Date: 16 Nov 2000 13:21:53 GMT2 From: "Paul McCormack" <paul.mccormack@compaq.com> Subject: DEC 5305 question7 Message-ID: <01c04fd0$2c60b0e0$d570ac10@meomccormackp1>s  K What is the magic to enable a "Digital Server 5305" to run Tru64 or OpenVMSr ? E This is one of the old DEC "Windows NT Only" systems that were sold ao, couple of years ago. Does such magic exist ?   Paul McCormack   ------------------------------  + Date: Thu, 16 Nov 2000 14:18:41 +0000 (UTC) ' From: Osmo Kujala <kujala@tukki.jyu.fi>n Subject: Re: DEC 5305 question, Message-ID: <8v0qc1$er3$1@mordred.cc.jyu.fi>  @ In comp.os.vms Paul McCormack <paul.mccormack@compaq.com> wrote:M > What is the magic to enable a "Digital Server 5305" to run Tru64 or OpenVMS  > ? G > This is one of the old DEC "Windows NT Only" systems that were sold aD. > couple of years ago. Does such magic exist ?   I think yes. From dejanews:e  N ------------------------------------------------------------------------------ Subject:1       RE: Can OVMS be installed on Digital server        3300?e  Date:       03/15/2000  Author:'       Ebinger . Eric <EEbinger@drc.com>s      4  The "magic" was to set the system to reset on boot,%  edit the nvram file and add the lineh     10 set srm_boot ONl   D  then boot the system and it will reset and not "poison" the HWRPB.    I  Fred Kleinsorge made a comment about the SRM_BOOT setting in a posting a-  while back.   <  Again, this is not guarenteed to work and is NOT supported.     Eric Ebinger ...oN ------------------------------------------------------------------------------   regards1          Osmo Kujala   ------------------------------  % Date: Thu, 16 Nov 2000 10:40:55 -0500l2 From: rdeininger@mindspring.com (Robert Deininger)# Subject: Re: f$search improvement ?eL Message-ID: <rdeininger-1611001040560001@user-2ivea3d.dialup.mindspring.com>  H In article <3A100FFA.53C9038A@bbc.co.uk>, tim.llewellyn@bbc.co.uk wrote:     K > Yes, I have had to insert a check for a wildcard in the search string andoN > have two code paths, one with a loop (if "*" found in search string) and one8 > without (no "*" found) in some procedures I have made.   It's worse than that.  You also need to check for "%" and "...".  And if the file spec starts with a logical name that is a search list, that counts as a wildcard also.  g It seems better to compare the f$search result to the previous result, as other posters have suggested.s   -- g Robert Deininger rdeininger@mindspring.comi   ------------------------------  % Date: Thu, 16 Nov 2000 09:30:41 -0800   From: Tobias Funk <funk@lbl.gov>" Subject: file transfer VMS to unix' Message-ID: <3A1419C1.CE081387@lbl.gov>c   Hi,-E I want to transfer a file from a VMS-machine to a unix machine from a B fortran program. My idea was using ftp. This works fine to another' VMS-machine but not to an unix-machine.O the command I used wasD reply=LIB$spawn('multinet ftp address  /user=#####  /pass=$$$$$  put test.for test.for')oG Unix seems to run ftp only interactively, so it doesn't except user andd password specifications.3 Has anyone a suggestion, how to solve this problem?  Thanks,. Tobias     --4 ____________________________________________________   Tobias Funkh LBNL 1 Cyclotron Rd , MS 6-2100 Berkeley CA 94720. USAh   phone : 1 (510) 495 2889 fax   : 1 (510) 486 5664 e-mail: funk@lbl.gov   ------------------------------   Date: 16 Nov 2000 12:06:44 GMT# From: himself@esands.com (him self)dB Subject: Re: Forcasting the Tropical Year - the Laskar expression?4 Message-ID: <slrn917jei.rt.himself@snark.esands.com>  I On Thu, 16 Nov 2000 03:28:40 GMT, Jerome Fine <jhfine@idirect.com> wrote:$   > D > of the leap year situation in the future?  My initial calculations= > using the Laskar expression suggest that both 4000 C.E. andb? > 6000 C.E. should not be leap years, but thereafter, I can seeo  . Do you expect your pdp-11 to still be running?   > < > And finally, the actual goal of the exercise is to be able6 > to write some code which can be used for a long time6 > into the future, as opposed to calendars which still  , I don't think there can be an answer to your) question because we don't know when or ifs, leap years will be invoked or revoked in the distant future.   + All you can do is assume the current rules .* will still be in use and use and base your+ computation on that. I believe leap secondsn, are invoked on an "as requires" basis. Maybe) people will do that with leap years too -  but who is to know?y   ------------------------------  # Date: Thu, 16 Nov 2000 13:15:52 GMTe  From: djim55@datasync.com (D.J.)B Subject: Re: Forcasting the Tropical Year - the Laskar expression?0 Message-ID: <3a13ddbc.2316018@news.datasync.com>  ' Jerome Fine <jhfine@idirect.com> wrote:l@ []Considering the case sensitivity posts over the last few days,9 []these news groups seems like a reasonable place to posed@ []a question about the use of the Laskar (1986) expression which- []http://astro.nmsu.edu/~lhuber/leaphist.htmly8 []http://www.treasure-troves.com/astro/TropicalYear.html  E You might try looking at two places: the British Royal Observatory atu9 Greenwich, England and the US Navy Observatory web sites.g   JimP.E --6 djim55 at tyhe datasync dot com. Disclaimer: Standard.' My Web pages Updated: November 6, 2000: . http://www.crosswinds.net/~djim51/updated.html Registered Linux user#185746   ------------------------------  % Date: Thu, 16 Nov 2000 19:14:52 +0010w% From: paddy.o'brien@zzz.tg.nsw.gov.au2> Subject: Re: FREE OpenVMS Windows NT intergration book *NOT*!!5 Message-ID: <01JWMAHETSG2006O58@tgmail.tg.nsw.gov.au>    Jay,  K >I had troubles with the online order form the first two times I tried it.  J >Maybe that's his problem.  Course I could always send him mine, it wasn't >what I was expecting.    O I'm losing track as to which, but many of these ordering things that require a  O state do not accept NSW.  It's not a state in US.  I get stuck at that point.   P Is that the same thing that happens to folk in the old dart, where there are no  states -- or the continent?    Regards, Paddy   Paddy O'Brien, Transmission Development, 
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australia    Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people, ; but that little whizz-bang apostrophe gives me little spam.    ------------------------------  % Date: Thu, 16 Nov 2000 08:47:05 -0500r; From: "Everhart, Glenn (FUSA)" <GlennEverhart@FirstUSA.com>b2 Subject: FW: [Fwd: RE: Technology of US elections]N Message-ID: <4B279CF3578CD211B945009027178017046D0173@swilnts809.wil.fusa.com>  = Right; but in this case there have been reports of comparablet9 numbers of dropped ballots in other areas of Florida, and = reports that the numbers dropped in Palm Beach are comparabler8 with the numbers dropped last election. (The differences< can be explained by larger turnout...14000 in 1996, 19000 inC 2000.) That kind of thing can be checked for and corrected. However,; I see NO evidence that Daley & Co. want to have an unbiasedn: election. They speak of a full count but limit it to their@ heavily favored counties. Speaking of how to achieve an unbiasedA count in a scientifically defensible way is pretty well not going  on.u  : You could correct for differences in machine tendencies to8 drop ballots by taking a set of ballots and running them8 thru all machines of a type, where the machines were the> same. For different machine types, all you can do is try based8 on failure rates normalized by different voter precinct < profiles (education, party, etc.) to correct for differences9 in machines. I believe this can be done with less risk ofj: election BIAS than manual recounts. If manual recount must= be used, though, you need to do at least a couple recounts in ; all areas with teams of different human bias and attempt tor? correct for human bias. Doing so would also mean first deciding0< on a set of criteria to be used, and it would take a helluva< lot of time. If you split which ballots were looked at first; by which group, systematic differences between halves mightmE tell you something about the rate of deterioration/mangling/tampering  that is going on.F  G I had understood that the vote-a-matic machines have voter instructionsa@ associated that tell voters to remove the chads also. These sameB machines have been in place for decades, as should have been thoseA same instructions. Arguably a first time voter might have ignoredmA the instructions, but it is harder to excuse voters who have been  there for multiple elections.F  ? At any rate I hear nothing much from Gore about unbiased count,r? and only a little from Bush about it that suggests he's gropingh< around the edges of the issue but doesn't really get it veryA clearly. (This is however just what an experimental physicist has @ to cope with all the time, which is how I happen to approach the> issue.) Bush realizes the partial recounts are biased, figures? that double counting takes care of mechanical gross malfunction ? issues, and doesn't believe in manual recount as less unbiased.bA Perhaps he does know how you correct for bias but doesn't want to E explain its model-dependence. At any rate, with no criteria given form@ how a ballot is to be determined to have been "intended" for oneA candidate or the other it gets pretty easy to tamper. Especially :? accepting "pregnant" holes...a finger might do that. This is ofjB course why Florida has such rigorous legislation limiting time for@ recounts; they've had a helluva lot of that kind of fraud in the past.:  F It would be refreshing to hear a complete discussion of how this oughtF to be handled, but I don't think either Gore or Bush is able really toD understand it or willing to approach the issue as a sampling problemG and to want to get a correction that is as nearly unbiased as possible.WD A full discussion might yield some decent measures and would I thinkD show that correcting the mechanical count can be done better than byC letting naked apes romp over all the ballots ad nauseam. But again,rC what they should be trying for is not a FULL count, but an UNBIASEDcC count. I wish I knew email addresses for either candidate; I'd have A written them something like the foregoing, only with more detail, @ days ago. Probably won't do much good at this point. The time toE evaluate bias was right after the election. That time was wasted withCB agitation. (It is also a sad commentary on gaps in the training ofA lawyers; how few of them must have ever tried to do statistics ornC extract a signal from noise. I wonder if either campaign has anyone1$ with scientific training available?)   Glenn Everhart     -----Original Message-----' From: Kilgallen@eisner.decus.org.nospamx* [mailto:Kilgallen@eisner.decus.org.nospam]) Sent: Thursday, November 16, 2000 7:56 AM  To: Info-VAX@Mvb.Saic.Com 2 Subject: Re: [Fwd: RE: Technology of US elections]    ? In article <3A13281D.E4675529@gce.com>, Glenn and Mary EverhartL <Everhart@gce.com> writes:  @ > There is no reason to complain about machine errors unless youB > believe the machine is biased toward one or the other candidate.> > (These machines are punched card readers...tough to build in > a bias there.)    @ If the machine fairly skips a certain percentage of the ballots,> and is located in an area of the state where one candidate has> more support than the other, it hurts that candidate more than
 the other.  5 That bias is not in the machine but in its placement.    ------------------------------    Date: 16 Nov 2000 15:40:36 +0100G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>R6 Subject: Re: FW: [Fwd: RE: Technology of US elections]H Message-ID: <y4em0cvv2z.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  = "Everhart, Glenn (FUSA)" <GlennEverhart@FirstUSA.com> writes:   . > I believe this can be done with less risk of< > election BIAS than manual recounts. If manual recount must? > be used, though, you need to do at least a couple recounts ini= > all areas with teams of different human bias and attempt to6A > correct for human bias. Doing so would also mean first decidingR> > on a set of criteria to be used, and it would take a helluva > lot of time.  G Funnily enough, all elections in Germany are purely manual, deliver the E results after about six hours (the semi-certified results are usuallymE published at about 1 a.m.), with predictions accurate to about 0.5-1% G (based on early counts and previous results) usually available after anpC hour or two. I have never heard anybody claim bias in the method.     F My background is in image processing and scene understanding. I would H always trust the human counter (if unhurried and under supervision) moreD than any machine. The machine will never find an unexpected problem.   	Jan   ------------------------------    Date: 16 Nov 2000 10:20:34 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)C6 Subject: Re: FW: [Fwd: RE: Technology of US elections]+ Message-ID: <56pKadEARCiE@eisner.decus.org>>  < Your original statement was theoretical, as was my response.< A machine that errs uniformly can certainly be used in a way8 to make things unfair.  That is not a judgement on which county in Florida did what.    In article <4B279CF3578CD211B945009027178017046D0173@swilnts809.wil.fusa.com>, "Everhart, Glenn (FUSA)" <GlennEverhart@FirstUSA.com> writes: >  > ? > Right; but in this case there have been reports of comparable  <snip>        A > In article <3A13281D.E4675529@gce.com>, Glenn and Mary Everhart" > <Everhart@gce.com> writes: > A >> There is no reason to complain about machine errors unless you C >> believe the machine is biased toward one or the other candidate. ? >> (These machines are punched card readers...tough to build ino >> a bias there.)  > B > If the machine fairly skips a certain percentage of the ballots,@ > and is located in an area of the state where one candidate has@ > more support than the other, it hurts that candidate more than > the other. > 7 > That bias is not in the machine but in its placement.B >    ------------------------------  % Date: Thu, 16 Nov 2000 17:52:53 +0100 6 From: "Martin Knoblauch" <martin.knoblauch@compaq.com>6 Subject: Re: FW: [Fwd: RE: Technology of US elections]6 Message-ID: <8v13dc$qgq$1@mailint03.im.hou.compaq.com>  L "Jan Vorbrueggen" <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> wrote inJ message news:y4em0cvv2z.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de...? > "Everhart, Glenn (FUSA)" <GlennEverhart@FirstUSA.com> writes:M >S > I > Funnily enough, all elections in Germany are purely manual, deliver theiG > results after about six hours (the semi-certified results are usually G > published at about 1 a.m.), with predictions accurate to about 0.5-1%.I > (based on early counts and previous results) usually available after an.C > hour or two. I have never heard anybody claim bias in the method.M >n  K  Not entirely true. Depending on the "state" or the type of election, there J are some mechanical counting machines, which give immediate numbers at theI end of the poll. But yes, most of our elections on all levels are counted 	 manually.   :   Some of the reasons that we have little problems may be:  J - we have a lot of relatively small counting precincts. That makes countig. the voting sheets fast (divide and conquer :-)L - counting is done in teams that always have members of at least the two big parties.I - counting is at least done twice for sanity checking. it continues until0 the number is stable.oL - even if there is error, there is no "the winner takes it all" on the stateL and federal level. this makes random errors less "dissastrous". on the localI level we have direct voting and those ballots are sometimes disputed veryD hard.d  G > My background is in image processing and scene understanding. I wouldlJ > always trust the human counter (if unhurried and under supervision) moreF > than any machine. The machine will never find an unexpected problem. >O  H  I tend to agree. Definitely a process where you first must mechanicallyJ punch a hole and then mechanically count the holes is bound for dissaster.   Martin   ------------------------------  % Date: Thu, 16 Nov 2000 12:34:34 -0300.) From: fabio_compaq@ep-bc.petrobras.com.brDB Subject: Gartner: Where =?iso-8859-1?q?=B4s_OpenVMS_=3F=3F=3F=3F?=L Message-ID: <OFC50D1982.28474048-ON03256999.005567D5@ep-bc.petrobras.com.br>  D Gardner predictions about the future of  OSes.  Where=B4s OpenVMS ??  H http://enterprise.cnet.com/enterprise/0-9566-714-1631259.html?tag=3Dst.=' it.9566.insnap.1641505-9566-714-1631259    Regards    FC=    ------------------------------   Date: 16 Nov 2000 16:57:39 GMT* From: helbig@astro.rug.nl (Phillip Helbig)F Subject: Re: Gartner: Where =?iso-8859-1?q?=B4s_OpenVMS_=3F=3F=3F=3F?=. Message-ID: <8v13m3$p7f$5@info.service.rug.nl>  
 In articleA <OFC50D1982.28474048-ON03256999.005567D5@ep-bc.petrobras.com.br>,", fabio_compaq@ep-bc.petrobras.com.br writes:   F > Gardner predictions about the future of  OSes.  Where=B4s OpenVMS ??  D Being used in the real-world and not being touted by self-appointed H pundits in the pay of someone they are supposed to judge objectively in I some publication which only those folks read who can't make up their own s minds.   ------------------------------  % Date: Thu, 16 Nov 2000 15:05:32 +0800a! From: Wayne Yung <iswayne@ust.hk>o Subject: Idle process Killer& Message-ID: <3A13873B.96167300@ust.hk>  F I am looking for some freeware or shareware programs which can monitorF the VMS processes and kill the processes which are idled for a certain period.    ------------------------------   Date: 16 Nov 2000 07:51:04 GMT) From: leslie@clio.rice.edu (Jerry Leslie)-  Subject: Re: Idle process Killer' Message-ID: <8v03l8$87s$1@joe.rice.edu>   " Wayne Yung (iswayne@ust.hk) wrote: : H : I am looking for some freeware or shareware programs which can monitorH : the VMS processes and kill the processes which are idled for a certain	 : period.n   From ftp.wku.edu:<  	   WATCHER0$      Version:      V3.1, 28-APR-1999.      Description:  Idle process monitor/killer;      Author:       Matthew D. Madison <madison@MadGoat.com>       Architecture: VAX,AXP      # of parts:   7      Language:     BLISS    4 --Jerry Leslie     (my opinions are strictly my own)   ------------------------------   Date: 16 Nov 2000 15:54:38 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)  Subject: Re: Idle process Killer6 Message-ID: <8v0vvu$psi$2@mailint03.im.hou.compaq.com>  J In article <3A13873B.96167300@ust.hk>, Wayne Yung <iswayne@ust.hk> writes:  G :I am looking for some freeware or shareware programs which can monitor G :the VMS processes and kill the processes which are idled for a certainc :period.     Please see the OpenVMS FAQ.3  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Thu, 16 Nov 2000 10:50:54 +0200d3 From: "Phillip du Plooy" <itbpjdp@puknet.puk.ac.za>#
 Subject: LDAP - Message-ID: <974364692.117191@news.puk.ac.za>t   Hi,i  K Is there anyone who syncs his or her OpenVMS accounts username and passwordoI with Novell, is this possible at all?  I am thinking in the line of LDAP._ What do I need to do this?  
 Thank you, Phillip du Plooy Potchefstroom University SouthAfrica    ------------------------------   Date: 16 Nov 2000 11:31:18 GMT1 From: "Kari Keronen" <kari.keronen@radiolinja.fi>- Subject: Re: LDAPa2 Message-ID: <01c04fc1$13302a60$5609a8c0@rlhkikker>  F Netscape FastTrack Server for VMS includes LDAP support. Also next VMSJ version (7.3) includes LDAP client. That should be out Q1 2001. Field test version of 7.3 is out now.   -Kari-  < Phillip du Plooy <itbpjdp@puknet.puk.ac.za> wrote in article$ <974364692.117191@news.puk.ac.za>... > Hi,n > D > Is there anyone who syncs his or her OpenVMS accounts username and passwordK > with Novell, is this possible at all?  I am thinking in the line of LDAP.b > What do I need to do this? >  > Thank you, > Phillip du Plooy > Potchefstroom University
 > SouthAfricab >  >  >  >    ------------------------------  % Date: Thu, 16 Nov 2000 10:57:11 -0500h# From: Jim Agnew <agnew@hsc.vcu.edu>u2 Subject: Re: Looking for RJ45-to-ThinWire adapter.+ Message-ID: <3A1403D7.F91C43DA@hsc.vcu.edu>u  ) oh, yeah, you're exactly right...  duh...t   Q'uaPla!   Javier Henderson wrote:a > ' > Jim Agnew <agnew@hsc.vcu.edu> writes:l > ; > > the cabletron one does not, it runs off the aui power..e > H >         The original poster was looking for RJ45 to ThinWire, however,E > so any solution to that problem will need an external power supply.t >  > -jav   ------------------------------  # Date: Thu, 16 Nov 2000 16:10:06 GMT 0 From: Timothy Stark <sword7@grace.speakeasy.org>( Subject: Looking for ROM bootstrap info.7 Message-ID: <yFTQ5.16369$wW2.570779@news1.giganews.com>f   Hello folks:  G I am preparing for my VAX emulator developement soon.  I am looking foreF ROM bootstrap information to load boot loader or image into its memoryC from any storage medium, etc. to start to execute operating system.a< I have VAX arch handbook (1st edition, not 2nd edition yet).  F To to load and run OpenVMS operating system or others, I have to writeD mu own ROM routine because I learned that from my VAX arch handbook.; That's why I am looking for ROM bootstrap information, etc.o  
 Thank you!   -- Tim Stark   -- i, Timothy Stark	<><	Inet: sword7@speakeasy.orgJ --------------------------------------------------------------------------F "For God so loved the world, that he gave his only begotten Son, that H whosoever believeth in him should not perish, but have everlasting life.. Amen." -- John 3:16 (King James Version Bible)   ------------------------------   Date: 16 Nov 2000 16:34:42 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman), Subject: Re: Looking for ROM bootstrap info.6 Message-ID: <8v12b2$q7j$1@mailint03.im.hou.compaq.com>  j In article <yFTQ5.16369$wW2.570779@news1.giganews.com>, Timothy Stark <sword7@grace.speakeasy.org> writes:H :I am preparing for my VAX emulator developement soon.  I am looking forG :ROM bootstrap information to load boot loader or image into its memoryrD :from any storage medium, etc. to start to execute operating system.  G   Um, on what platform are you looking to implement the emulator -- theoF   question you ask is largely a question that is specific to the host C   environment.  (In the case of a VAX or Alpha system, the console  D   program will provide for the initial load of the bootstrap image.)  = :I have VAX arch handbook (1st edition, not 2nd edition yet).U :cG :To to load and run OpenVMS operating system or others, I have to write2E :mu own ROM routine because I learned that from my VAX arch handbook.;< :That's why I am looking for ROM bootstrap information, etc.  J   Huh?  Most VAX and the Alpha system consoles include enough capabilitiesG   to decode the contents of the bootstrap block and load the referencedlG   image (VMB or APB) into memory, and transfer control to it.  Various 9F   VAX systems include and use VMB on the console ROM or on the console   storage media.  H   The VAX architecture handbook is entirely the wrong reference materialF   for this purpose -- you will want the internals and data structures D   manual, and the processor-specific technical manuals.  The former G   describes the general case with general information from many of the  H   systems, and the latter contains detailed information specific to the    particular system.  F   In the case of the technical manual, EK-KA630-UG -- if you can stillA   get it -- might be a good choice.  This is the MicroVAX II CPU dB   technical manual, and it discusses the implementation of and the-   various options of the KA630 CPU bootstrap.   D   The other question is "what system" -- if you are running your VAXC   emulator on another operating system, you will likely use the I/O F   constructs of that system to load and transfer control to VMB or (inI   the case of an Alpha system) to APB.  VMB and APB will typically expectsF   some assistance from the primitive I/O routines that are part of the   console program.  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Thu, 16 Nov 2000 14:16:45 +0000v9 From: "Miller, Daniel" <Daniel.Miller@nightfreight.co.uk>h Subject: Mime attachmentA Message-ID: <5E8A0E4920B0D411B1E900508BFCB240299415@NF-HOUSE-NT1>    Hi everyone,  L We are trying to setup a system on our alpha that will automatically processH files received by email.  We are using the DELIVER program to handle theJ incoming mail and using munpack(thanks to everyone who suggested MPACK, weJ had a problem in that we dont have DEC C, but i have managed to compile it with GCC)to extract the files.  B The problem is we are expecting text files with one line about 600D characters long, but when we receive the file everything after 255th character is beingH placed on to separate lines.  We have tracked this down to only when theH files are from one outside source and the mime files have the following:  ; Content-Type: application/octet-stream; name=COLLECTION.TXTr  K The files where originally .ps files, does anyone have any ideas about whatwH sort of file this could be?  The interesting thing is when the very sameF file is sent to another email address (on MS exchange, rather than theK alpha), and then forwarded to the alpha, it comes through as one line.  YouoK can see the problem is there in mail, before extracting it, so i think thatn it is not munpack.  ' We are running vms 7.2-1 and tcpip 5.0at   thanks,u  
 Daniel Millert Nightfreight plc   ------------------------------   Date: 16 Nov 2000 16:56:21 GMT* From: helbig@astro.rug.nl (Phillip Helbig) Subject: Re: Mime attachment. Message-ID: <8v13jl$p7f$4@info.service.rug.nl>  A In article <5E8A0E4920B0D411B1E900508BFCB240299415@NF-HOUSE-NT1>,s< "Miller, Daniel" <Daniel.Miller@nightfreight.co.uk> writes:   N > We are trying to setup a system on our alpha that will automatically processJ > files received by email.  We are using the DELIVER program to handle theL > incoming mail and using munpack(thanks to everyone who suggested MPACK, weL > had a problem in that we dont have DEC C, but i have managed to compile it  > with GCC)to extract the files.   Have you tried $  MC MIME?  D WHY DOES THIS CLEAR THE SCREEN!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?!?  D > The problem is we are expecting text files with one line about 600F > characters long, but when we receive the file everything after 255th > character is beingJ > placed on to separate lines.  We have tracked this down to only when theJ > files are from one outside source and the mime files have the following: > = > Content-Type: application/octet-stream; name=COLLECTION.TXTi > M > The files where originally .ps files, does anyone have any ideas about whatb > sort of file this could be?  :  C BAD PostScript files.  PostScript is supposed to be 7-bit printable>F US-ASCII.  They are text files.  There is no reason they need to have   lines longer than 80 characters.   ------------------------------  # Date: Thu, 16 Nov 2000 11:09:25 GMT = From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)t4 Subject: Re: open VMS backdrop for motif or xwindows0 Message-ID: <009F331C.3C122CE2@SendSpamHere.ORG>  k In article <8uulsk$d5p$3@mailint03.im.hou.compaq.com>, hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) writes:e >rm >In article <8umefo$1ub$1@aristotelis.cytanet.com.cy>, "Constandinos Stavrou" <sdinos@cytanet.com.cy> writes: M >:Can anybody cprovide any interesting backdrops for modif or xwidnows?  I amf1 >:stuck with the ones that came with openVMS 6.2.: >:H >  Well, since no backdrops came with OpenVMS...  :-)   Seriously, what J >  version of DECwindows is in use, and are you using OpenVMS VAX V6.2 or  >  OpenVMS Alpha V6.2? >eK >  With tools such as those in decw$utils:, you can set the root window to sJ >  be most anything you want; decw$utils:xsetroot or similar.  Tools such M >  as xv and imagemagick can also be used to display backgrounds (backdrops). L >  With recent versions of DECwindows CDE, you can add your own backdrops... > M >  There are a pile of X toys on the OpenVMS Freeware, including xphoon, etc.u  ( You can add your own.  It's quite easy.   * Here's one I have on one of my workspaces:    http://www.tmesis.com/backdrops/  J Restore the .ZIP file to CDE$SYSTEM_DEFAULTS:[BACKDROPS] and then use the ( Style Manager "backdrop" to select it.     --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              O city, n., 1. a place where trees are cut down and streets are named after them.s   ------------------------------  % Date: Thu, 16 Nov 2000 05:01:47 +0100y2 From: martin@radiogaga.harz.de (Martin Vorlaender)4 Subject: Re: Pascal calling convention from C++ code; Message-ID: <3a135c2b.524144494f47414741@radiogaga.harz.de>s  : Paul Johnston (pwjohnston@no_spam.dingoblue.net.au) wrote:< : Can anyone help me ? I need to know the calling convention< : to call a Pascal routine from a C++ module on Vax-OpenVMS.J : is it something like this ?, or do I need to specify far or near or some# : other tricky calling convention ?u  , "far"? "near"? No Intel segment crap here...  ( : extern "C" void MyPascalFunc(int * a);  8 What's more interesting would be the Pascal declaration. From the above I would deductw  ,   PROCEDURE MyPascalFunc( VAR a : INTEGER );  % Have a look at the DEC Pascal docs at B http://www.openvms.digital.com/commercial/pascal/pascal_index.html   cu,    Martin --J One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer7 One OS to find them           | work: mv@pdv-systeme.derN One OS to bring them all      |       http://www.pdv-systeme.de/users/martinv/> And in the Darkness bind them.| home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 16 Nov 2000 12:07:47 +0100a- From: Jouk Jansen <joukj@hrem.stm.tudelft.nl> 4 Subject: Re: Pascal calling convention from C++ code3 Message-ID: <3A13CE13.5D3261D5@hrem.stm.tudelft.nl>s   Martin Vorlaender wrote: > < > Paul Johnston (pwjohnston@no_spam.dingoblue.net.au) wrote:> > : Can anyone help me ? I need to know the calling convention> > : to call a Pascal routine from a C++ module on Vax-OpenVMS.L > : is it something like this ?, or do I need to specify far or near or some% > : other tricky calling convention ?n > . > "far"? "near"? No Intel segment crap here... > * > : extern "C" void MyPascalFunc(int * a); > : > What's more interesting would be the Pascal declaration. > From the above I would deducte > . >   PROCEDURE MyPascalFunc( VAR a : INTEGER ); > @ It also depends how you compiled the C++ routine. if you specify@ /name=as_is or /name=lower the pascal function will not be foundE since the current Pascal compiler for VMS upcases all external-names.r  G Question for Compaq : is the /NAME qualifier planned to be added in thee!   next version of PASCAL for VMS?t                              Jouk    -- e  > Ceterum censeo tertium millennium post Christum natum anno MMI incepturum esse.  P >------------------------------------------------------------------------------<  
   Jouk Jansenw 		 e   joukj@hrem.stm.tudelft.nl   E   Technische Universiteit Delft        tttttttttt  uu     uu  ddddddd F   Nationaal centrum voor HREM          tttttttttt  uu     uu  dd    ddG   Rotterdamseweg 137                       tt      uu     uu  dd     ddgG   2628 AL Delft                            tt      uu     uu  dd     ddbF   Nederland                                tt      uu     uu  dd    ddE   tel. 31-15-2781536                       tt       uuuuuuu   ddddddd"  P >------------------------------------------------------------------------------<   ------------------------------  % Date: Thu, 16 Nov 2000 09:18:41 -0500e- From: John Reagan <reagan@hiyall.zko.dec.com>h4 Subject: Re: Pascal calling convention from C++ code2 Message-ID: <3A13A671.7892AEDA@hiyall.zko.dec.com>   Jouk Jansen wrote: > I > Question for Compaq : is the /NAME qualifier planned to be added in the # >   next version of PASCAL for VMS?u >   F No, we have no plans for adding a /NAME qualifier to Pascal.  However,B if you want to specify a non-uppercase external name, we have that ability.  F [external('AMixedCaseName')] procedure AMixedCaseName( var i : integer ); external;  * will get you the mixed case external name.   -- t John Reagan> Compaq Pascal Project Leader   ------------------------------  + Date: Thu, 16 Nov 2000 07:58:53 +0000 (UTC)s' From: Osmo Kujala <kujala@tukki.jyu.fi>9M Subject: PC programs in OpenVMS (was: Re: How can I watch RealPlayer movies?)-, Message-ID: <8v043t$5lm$1@mordred.cc.jyu.fi>  G Bart Z. Lederman <lederman@star.enet.dec.disable-junk-email.com> wrote:s  0 >>Do I have say to no thanx to RealPlayer stuff?  ! > I would, but for other reasons.   J Good reasons. (Proprietary format and spying reputation) I'll try to avoid RealPlayer too.-  H Lets forget RealPlayer. Picked it as an example and it was kind of miss.  J But there are plenty of common "PC-programs" that we need to run at times.H There have been discussions about this before but I'm trying to check ifA someone has something new to say. About FX!32 maybe? Softwindows?w SoftLinux! ?...t   regards            Osmo Kujala    ------------------------------  % Date: Thu, 16 Nov 2000 09:10:38 +0000g. From: 'o-Dzin Tridral <TridralO@Cardiff.ac.uk>9 Subject: Removing trailing spaces from a 1700 byte recordr- Message-ID: <3A13A49B.5926.550D713@localhost>m  K A colleague needs to remove trailing spaced from a fixed-length record fil=  e, dK where the records are all 1700 bytes.  I usually attack these things with =  a bit H of DCL to read and write the file, but of course 1700 bytes is too much.  D Is there a simple way to remove the training spaces from this file?    File organization:  SequentialF File attributes:    Allocation: 342, Extend: 0, Global buffer count: 0$                     No version limit2 Record format:      Fixed length 1700 byte records Record attributes:  None RMS attributes:     None Journaling enabled: None  Total of 1 file, 333/342 blocks.   Thank you for any assistance,m  	 '=F6-Dzini -- 'o-Dzin Tridral E Senior Computer Officer, UIS, Cardiff University, PO Box 78, CF10 3XLoK T +44 29 2087 6160  E TridralO@cf.ac.uk  F +44 29 2087 4531  W http://www.=v cf.ac.uk   ------------------------------  % Date: Thu, 16 Nov 2000 10:26:28 +0000I- From: "POWERS, John" <John.POWERS@sema.co.uk>a= Subject: RE: Removing trailing spaces from a 1700 byte record.= Message-ID: <D30A62ABC710D211AEE100A0C9D615EE01528F51@REAES2>o  5 From: 'o-Dzin Tridral [mailto:TridralO@Cardiff.ac.uk]e   > H > A colleague needs to remove trailing spaced from a fixed-length record file, K > where the records are all 1700 bytes.  I usually attack these things withn a bit J > of DCL to read and write the file, but of course 1700 bytes is too much. > F > Is there a simple way to remove the training spaces from this file?  >   > File organization:  SequentialH > File attributes:    Allocation: 342, Extend: 0, Global buffer count: 0& >                     No version limit4 > Record format:      Fixed length 1700 byte records > Record attributes:  None > RMS attributes:     None > Journaling enabled: None" > Total of 1 file, 333/342 blocks. >  > Thank you for any assistance,y  F Well this sounds like the ideal question to reactivate the TPU vs. EDT jihad again!  B I have a prewritten command procedure in my utils directory calledH TRIMFILE.COM, which does this nice and quickly. Unfortunately (for some)+ it uses TPU and is therefore 'untouchable'.s  G Well this is fast and easy to use. All those qualifiers like NODISPLAY,lF NOJOURNAL and NOINIT are all there to speed it up. If you want to makeC it go slower and thereby help the EDT cause, then chop them out !-)2   $!TRIMFILE.COMG $ edit /nodisp/noini/nojou/command=sys$input 'p1' /out='f$parse(p2,,p1)u  eve$trim_buffer(current_Buffer);	 eve$exit;  $ exit  I Note also that if you have a symbol in your login.com that says somethingiG like EDIT :== EDIT/EDT, then this will fail spectactularly! I leave theo6 workaround in this case as an exercise for the reader.    -- Cheers, John  F  - Note  This message represents my opinions and nothing else, not theI   opinion of SEMA, my family, or the cricket club - though my dog Meg didhE   nod in agreement whilst I was typing. If you have any problems then D   please complain to her (or me, but not SEMA, my family or the CC).    K ___________________________________________________________________________ B This email is confidential and intended solely for the use of the H individual to whom it is addressed. Any views or opinions presented are E solely those of the author and do not necessarily represent those of 0 Sema Group. M If you are not the intended recipient, be advised that you have received thispI email in error and that any use, dissemination, forwarding, printing, or o- copying of this email is strictly prohibited.l  E If you have received this email in error please notify the Sema Groupl. Helpdesk by telephone on +44 (0) 121 627 5600.K ___________________________________________________________________________k   ------------------------------    Date: 16 Nov 2000 09:07:05 -0500, From: koehler@eisner.decus.org (Bob Koehler)= Subject: Re: Removing trailing spaces from a 1700 byte record.+ Message-ID: <ImYB$WlShr4H@eisner.decus.org>n  ^ In article <3A13A49B.5926.550D713@localhost>, 'o-Dzin Tridral <TridralO@Cardiff.ac.uk> writes:M > A colleague needs to remove trailing spaced from a fixed-length record fil=  > e, dM > where the records are all 1700 bytes.  I usually attack these things with =v > a bit J > of DCL to read and write the file, but of course 1700 bytes is too much. > F > Is there a simple way to remove the training spaces from this file?  >   ? Here's one of the first little things I wrote back when we were:H transfering 80 column Fortran source card images from an IBM mainframe. H It assumes 80 column input and ignore everything after column 72, but itH can be teaked easily, or rewritten if you don't have a Fortran compiler.D On average only 40 columns were actually used in our source, so this( saved us just about half our disk space.   c R. Koehler c get rid of wasted space 0 	character*255 name,image*133,card*133,prompt*9,' 	1 command*132,token*132,delim*1,clen*2l 	integer point# 	data prompt/'Squeeze>'/,delim/'/'/g* 	istat=lib$get_foreign(command,prompt,len). 	call get_token(command,token,point,len,delim) 	name=tokenc
 	delim=' '. 	call get_token(command,token,point,len,delim) 	clen=token(2:3) 	decode (2,1000,clen)lenro 1000	format(i2)o 	IF (LENR.EQ.0) LENR=133" 	  open (2,file=name,status='old')" 	  open (3,file=name,status='new', 	1 carriagecontrol='list') 	  do while (.true.) 	    read(2,200,end=400)image  200	    format(1a133)M# 	    call str$trim(card,image,ilen)?" 	    if (lenr .lt. ilen) ilen=lenr 	    write (3,500) image(:ilen)  500	    format(a)t	 	  end do  400	  close (2)n 	  close (3)
 999	call exitg 	end' 	SUBROUTINE GET_CHAR(COMMAND,CHR,POINT)l# 	CHARACTER*132 COMMAND,CHR*1,CHAR*1l 	INTEGER POINT 	CHR=COMMAND(POINT:POINT)e/ 	IF(ICHAR(CHR).GE.97.AND.ICHAR(CHR).LE.122)THENu 		CHR=CHAR(ICHAR(CHR)-32)  	END IFe 	RETURNe 	END4 	SUBROUTINE GET_TOKEN(COMMAND,TOKEN,POINT,LEN,DELIM)/ 	CHARACTER*132 COMMAND,CHAR*1,TOKEN*132,DELIM*1c 	INTEGER POINT 	TOKEN='                    'e 	IF(POINT.GT.LEN)RETURNe" 	CALL GET_CHAR(COMMAND,CHAR,POINT) 	DO WHILE(CHAR.EQ.DELIM) 	    POINT=POINT+1 	    IF(POINT.GT.LEN)RETURN & 	    CALL GET_CHAR(COMMAND,CHAR,POINT) 	END DOn 	LAST=1a 	TOKEN(1:1)=CHAR 	POINT=POINT+1" 	CALL GET_CHAR(COMMAND,CHAR,POINT) 	DO WHILE(CHAR.NE.DELIM)( 	    TOKEN(1:LAST+1)=TOKEN(1:LAST)//CHAR 	    LAST=LAST+1 	    POINT=POINT+1 	    IF(POINT.GT.LEN)RETURNU& 	    CALL GET_CHAR(COMMAND,CHAR,POINT) 	END DOt 	RETURN  	END  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporationc= NASA GSFC Flight Software       | Federal Sector, Civil GroupME                                 | please remove ".aspm" when replyingI   ------------------------------    Date: 16 Nov 2000 10:27:57 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)-= Subject: Re: Removing trailing spaces from a 1700 byte record#+ Message-ID: <IF3pPX6YtNh7@eisner.decus.org>-  ^ In article <3A13A49B.5926.550D713@localhost>, 'o-Dzin Tridral <TridralO@Cardiff.ac.uk> writes:M > A colleague needs to remove trailing spaced from a fixed-length record fil=B > e, 1M > where the records are all 1700 bytes.  I usually attack these things with =a > a bit J > of DCL to read and write the file, but of course 1700 bytes is too much. > F > Is there a simple way to remove the training spaces from this file?    	$ EDIT/TECO <filespec>  	<fn^Esn 	$ 	$;>ex$$  D You might have to reset the file characteristics on the new version.   ------------------------------  # Date: Thu, 16 Nov 2000 15:35:13 GMT0 From: abirkett@my-deja.com= Subject: Re: Removing trailing spaces from a 1700 byte record ) Message-ID: <8v0urh$f3v$1@nnrp1.deja.com>e   Hi,d  E I may be mistaken but can't you write symbols upto 2048 characters ifs$ you use the $write/symbol technique?   $ read/end=eof infile inrecC $ inrec=f$edit(inrec,"trim") $ write/symbol outfile inrec   ???    Adey --D Intelligence is being able to listen to the William Tell overture...7                  ...without thinking of the Lone Ranger     & Sent via Deja.com http://www.deja.com/ Before you buy.t   ------------------------------    Date: 16 Nov 2000 12:12:38 -0500 From: briggs@eisner.decus.orgt= Subject: Re: Removing trailing spaces from a 1700 byte recordv+ Message-ID: <7XZYYgPIXpi4@eisner.decus.org>l  F In article <8v0urh$f3v$1@nnrp1.deja.com>, abirkett@my-deja.com writes:G > I may be mistaken but can't you write symbols upto 2048 characters if:& > you use the $write/symbol technique? >  > $ read/end=eof infile inreca > $ inrec=f$edit(inrec,"trim") > $ write/symbol outfile inrec  - Yup.  You can read 'em and you can write 'em.k   But you can't f$edit 'em.    	John Briggs   ------------------------------  % Date: Thu, 16 Nov 2000 02:29:34 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> D Subject: Re: RMS way to get long sequential records in multiple $GET, Message-ID: <3A137ECC.8291FD2E@videotron.ca>   hein@eps.zko.dec.c*m wrote:pP >>         And is that delimitor an RMS record terminator? Or does it correspondJ >         with the end of an RMS record? I guess I'm a litte dense tonite.  H Those files are provided to me as variable length record files, but someL records are quite long. I have elected to use your suggestion and allocate aL 32k buffer and just read every line as one large chunk and then break it up.M My poor Microvax II won't like me for allocating such a HUGE amount of memoryo though :-) :-) :-) :-) :-) :-)  N >         a large enough buffer. Heck, the characer cell solution might simply- >         have too small an OUTPUT buffer no?i  J No, the character cell solution simply displays the first 80 bytes of each1 line, so I won't see the rest of the "paragraph".    > >ok. "PRETTY PLEASE" ! > H > no, pretty please is spelled SYS$MODIFY as per example in other reply.  : I guess I need to call SYS$THANKS("Pretty thanks"); then ?   ------------------------------  % Date: Thu, 16 Nov 2000 10:42:05 +0000o0 From: andrew harrison <andrew.nospam@uk.sun.com>7 Subject: Re: Severity of eBay's ecache problem detailedo* Message-ID: <3A13B9FD.4666D0AB@uk.sun.com>   Jordan Henderson wrote:I > , > In article <3A12B4DA.74BCA7CB@uk.sun.com>,5 >   andrew harrison <andrew.nospam@uk.sun.com> wrote:s > > Fred Kleinsorge wrote: > > >r > > > Calling the kettle black?d > > >I > >  > > This is Arnes Rule > >mH > > > > Arnes rule to distinguish between technical people and marekting	 > people:iG > > > > Do they consider information about systems not shipping yet forM > fact ?* > > > > Yes => marketing. No => technical. > > > >a > >s> > > I havn't made any pronouncments of "fact" about any future2 > > systems. So I fail Arnes test for a Marketeer. > >s, > > Rob on the other hand has pronounced on. > >d: > > Spiralog (when it was a lab project and a white paper) > > Galaxies (ditto) > > WildFire (ditto) > > 21264  (ditto) > >aH > > In the case of Spiralog for example all of Robs pronouncements aboutJ > > the performance gains to be had using Spiralog turned out to be untrueI > > and his extrapolation to its effects on TPC-C performance also turnedS  > > out to be wildly optimistic. > >hE > > Now if Rob had prefaced all these pronouncments with a disclaimer G > > then by Arnes definition he could still claim to be technical as it D > > is he didn't and so he cannot (at least using Arnes definition). > D > What an amazing thing you've been able to do, Andrew.  You've beenH > able to uncover someone in comp.os.vms who is enthusiastic and perhaps7 > overly optimistic about certain OpenVMS technologies.D > G > Oh, we should be ashamed for fostering an atmosphere where discussionoB > of future OpenVMS technologies are put in such a positive light. > @ Oh I have no problem with people being optimistic about OpenVMS A if they have anything to back that optimism up and if they don't o2 try to use that "optimism" to bash other vendors.   ? But Rob's Spiralog and Galaxies are going to whip Sun/HP/IBM's  ? arses postings, which even included suggestions about what sort ; of TPC-C performance you might expect to get fall into the c= category of "optimism" with nothing to back it up being used o to bash other vendors.  = And Kerrys "everything in the OpenVMS garden is rosy" type ofe= posts also fall into this category because they don't reflecta. reality and are simply whistling in the wind.     G > We should go back to topics of discussion that are actually pertinent F > for comp.os.vms, the US Election and fabio's request for information8 > about a file system product available for Sun systems. >   ? I think you will find I answered Fabios question to best of my o ability.     Regardst Andrew Harrisonu Enterprise IT Architect    ------------------------------  % Date: Thu, 16 Nov 2000 11:14:24 +0100l= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>@7 Subject: Re: Severity of eBay's ecache problem detailed?) Message-ID: <3A13B37F.2FE6DB34@gtech.com>    Larry Kilgallen wrote:j > In article <3A10EC91.15E9592@gtech.com>, Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> writes: > > This is a technical forum !e >  > Not in this thread.r  1 So maybe this thread do not belong in this forum.v  B > > As long as Ebay does not publicise a statement explaining what > > hardware/software ; > > was responsible, then Robs guess is as good as anyones.c > B > It is not reasonable for eBay to discuss reliability of computer? > systems, since it detracts from their main marketing message,n? > which is the bidding experience.  This has nothing to do witha@ > what machines they use and how often they fail.  When they areA > _forced_ into discussing downtime, it pushes them into a losingS > position.    ????  D I can not see why Ebay making a public statement somewhere about the< cause of the now very famaous incident should distract their customers !-  H The incident ar every well-know, lots of speculation etc.. One statementC and the discussion would be if not finished, then at least based ont factsuH instead of guess-work (when SUN are known for requiring NDA's in case ofD technical problems, then assurance from a SUN guy about it not being* SUN's fault is not particular convincing).   Arne   ------------------------------  # Date: Thu, 16 Nov 2000 13:52:02 GMT   From: jlsue <jlsuexxxz@home.com>7 Subject: Re: Severity of eBay's ecache problem detailed 8 Message-ID: <f6p71t45rg2b32ob0roiokcfeg4jblqpjj@4ax.com>  3 On Mon, 13 Nov 2000 15:29:51 +0000, andrew harrison-! <andrew.nospam@uk.sun.com> wrote:h  H >ECC would help but we are getting single and Multi-bit errors and many G >people would not consider multi-bit ECC for the whole cache as being a1 >reasonable proposition. >.B >Not everyone agrees on the necessity of having ECC for the whole @ >of the cache either provide you can reload the offending block. >rA >Despite Robs FUD the fact is that people who have installed the PA >cache scrubber and either changed to the current SRAM or changeda@ >the server environment or both have seen very big reductions inB >failure rates. In a couple of customers I know of in the UK this  >has resulted in 0 failures.  F Fine, Andrew, but this still doesn't explain why NO other vendor seemsD to be having the same problems.  Only Sun.  Is Sun the only computer< maker using that supplier?  And how does the scrubber affect performance?  1 Not speaking for anyone, certainly not DEC/Compaq$- (get rid of the xxxx in my address to e-mail)e   ------------------------------  % Date: Thu, 16 Nov 2000 15:18:06 +0000n0 From: andrew harrison <andrew.nospam@uk.sun.com>7 Subject: Re: Severity of eBay's ecache problem detailed * Message-ID: <3A13FAAE.5E33639A@uk.sun.com>   Rob Young wrote: > _ > In article <3A0FFC6D.3DE5CA24@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:" > > Rob Young wrote: > >>b > >> In article <3A0C0865.497FBF17@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:- > >> > "Brian Schenkenberger, VAXman-" wrote:w > >> >> ` > >> >> In article <vXePVDXmGW4S@eisner.decus.org>, young_r@eisner.decus.org (Rob Young) writes: > >> >> >{...snip...} , > >> >> >       Come spin with us Andrew!!!!	 > >> >> >u+ > >> >> >                               Roby	 > >> >> >  > >> >>DR > >> >> Rob, really!  Andrew's contribution here is to tout Sun and FUD Compaq andQ > >> >> VMS technologies.  You continue to invite him.  I know that you and KerryuP > >> >> and several others are trying to dispell his FUD and show what a FUDsterP > >> >> he is himself but the truth is that I believe we all know what a FUDster/ > >> >> he is.  Can we try to stop baiting him?u > >> >>i > >> > > >> > >> >K > >> > Just occasionally someone pops out of the woodwork and says no thatsn1 > >> > B****H**T but only very very occasionally.r > >> > > >> > No wonder I stayed. > >> >L > >> > You have mostly Rob to blame for this so you are posting to the right > >> > person. > >> >H > >> > Incedentally despite labeling me as a FUDSTER and a spin merchantD > >> > the fact is that my objections to mainly Robs hype on Galaxy,O > >> > Spiralog, ISV interest in OpenVMS, the performance of the WildFire boxes- > >> > etc have been correct.  > >> > > >>I > >>         No ... I made a few mistakes.  I bought into Spiralog and itnK > >>         is/was great technology.  I stand fast on the other things you0R > >>         sling at me.  I am ahead of the curve on Wildfire and Son-of-WildfireO > >>         but faster 21264 and 21364 are a long time coming (and yes, othersdP > >>         aren't standing still but Compaq is doing VERY well in the high-endJ > >>         as they and IBM are suddently the platform of choice in HPC). > >>L > > No Rob Spiralog wasn't great technology, it was a great white paper. Had > > it beeni* > > great technology it might have worked. > G >         It was great technology and it does work.  Folks are using it 1 >         today, it doesn't have a future though.e >   C Weasel words Rob it does work if you call a functioning filesystem oC as working, it does not deliver the performance benefits which were D its main design goals and which you trumpeted to the adoring masses.  C Nor did it have the impact you said it would on TPM performance on t	 OpenVMS.    C It doesn't have a future because it did not deliver the performancei< it was designed to deliver and which you predicted it would.  # > It also illustrates perfectly thesD > > problem you seem to have, all the things you refer to, Spiralog, > > WildFireE > > etc have been great white paper material its the translation intoe > > products > > thats been the problem.- > >  > * >         No... Wildfire is a product too. > % Much later than you said it would be.v! Slower than you said it would be.$- With fewer CPU's than you said it would have.s/ Not whiping anyones arses as you said it would.e  I > > You were probably ahead of the curve on WildFire for the same reason,u > > yousK > > read too many white papers published about a product that didn't exist.s > >e >  >         Both products exist. > B But not with the characteristics you said they would, this is just( a repeat of your Spiralog weasel words.     J > > You and Jordan also appear to share the same rather cavalier attitiude > > towardsmI > > time lines, you I think pre-announced that WildFire was going to whip_K > > everyones asses nearly 2 years before it actually shipped. Ahead of ther) > > Curve or Totally Jumped the Gun ?????a > >s > H >         Ahead of the curve.  Jumping the gun is when Sun presumed theyC >         would win the would win the Lawrence Livermore ASCI Q bidi >         and they didn't. > = What has that got to do with a discussion about your ability s9 to be a credible source of Compaq related information. Wei9 arn't talking about what Sun predicted but about what youi
 predicted.  = > http://www5.compaq.com/hpc/tsn/iss017/hptc_iss017_fa.html#11 > K > > And ahead of the curve assumes that you think that WildFire is going to_J > > get fixed ? What changes are Compaq going to make to WildFire's memoryP > > subsystems to reduce local memory latency and reduce the ratio between local > > and remote memory latency ?s > >D= > > What predictions are you making for Son of WildFire ?????. > >1 > ? >         Higher bandwidth, lower latency.  Much lower latency:o > 8 I know you even posted numbers didn't you, how very like you.  . > http://www.alphapowered.com/ev7-and-ev8.html > I > >>         But yes, you are a FUDSTER and a spin merchant as you rarelyeI > >>         if ever address the contents of a post and typically lie lowaH > >>         until the contents have shifted ever so slightly to warrantT > >>         a total change of subject.  You sure you didn't co-invent the Internet? > >>K > >>         By the way, regarding the subject line, how about this anyhow:g > >>9 > >> http://public.wsj.com/sn/y/SB973636942369499535.htmlo > >>M > >> At online auctioneer eBay Inc., a major Sun customer, the memory problemrR > >> initially resulted in one to two big-system crashes every month, says MaynardR > >> Webb, president of eBay Technologies. The memory problem wasn't behind eBay'sQ > >> most serious outage, a 22-hour failure in 1999, but has caused other serviceiQ > >> outages, Mr. Webb says, although he adds that with Sun's fixes in place eBaye2 > >> can now go "months" without seeing a problem. > >oK > > How remarkable you actually included the punchline in one of your postsa > > instead of cutting it out. > >oL > > "he adds that with Sun's fixes in place eBay can now go "months" without > > seeing a problem." > > # > > Remarkable and a first for you.g > >0 > G >         Andrew, that's not something to be proud of.  Boxes shouldn't.= >         crash every few months because of CPU cache issues.m  ? Rob you are on record as saying that it has nothing to do with  - the environment that machines are running in.   Q > >>         Do something and shock us.  Admit that you have attempted to totally(K > >>         mislead us for the past year on this issue.  Perhaps you thinkDP > >>         it will simply go away or will not longer make headlines?  In lightJ > >>         of the above, how do those postings in December 1999 now makeI > >>         you look?  No different.  "We" essentially knew all along it)N > >>         wasn't a matter of datacenter temperature or environmentals.  ButH > >>         you wouldn't give it up.  Now you are forced to give up theI > >>         datacenter spin but it is all out there for us to see and toIJ > >>         trot out as we wish.  But you are oh so sure of yourself whenJ > >>         you write some things and the DANGER is kids that are lurkingM > >>         wouldn't have a clue that what you pen is often VERY misleading.- > >>K > >>         Your reputation took a hit.  Reputation?  Well.... never mind.o > >>G > > Despite your claim that I havn't answered your questions Rob I have A > > in fact explained this to you a number of times and each time-A > > you either cannot or will not understand what the issues are.  > >l > L >         Sure I understand the issue.  You dodge every question put to you.C >         Ask Jordan.  Ask Kerry.... ask anyone else in this forum.o > < Since when has Kerry and in particular Jordan put a question to me that I have dodged.   = Kerrys "questions" tend to be non questions intersperced withxB marketing URLS and Jordan is obsessed with a "carefull" definitionD of FUD which would be fine except that he doesn't use the "carefull" definition he keeps quoting.  @ And anyway why would I want to ask the choir when I can ask the  choirmaster.   > >r@ > > 1.    There is a problem with commodity SRAM supplied by one> > >       vendor to Sun. This problem causes bits to flip when9 > >       they should not. These can be multi-bit errors.  > >0H >         So... if as John Shoemaker relates Sun hadn't made the blunderK >         of using ECC L2 cache in their servers this would be a non-issue.eL >         IBM, HP and Compaq choose to use ECC L2 in their high-end servers. >   A Sure, having ECC would have helped but do Compaq or IBM implimentC	 multi-bit$E ECC for their caches. It would not have eradicated the problem at then> error rate that we were getting with the origional SRAM parts.  D Nor is not having ECC necessarely the issue you think it is, you can re-loado$ the offending page from main memory.  I > > 2.    The bits don't just flip by themselves they require an externalm, > >       force, static, gamma radation etc. > >rB >         Trying to sneak static in there on us?  How about gamma. > H Rob, you claim you understand what is going on and then sneak in a pointD that proves you don't. Gama radiation and static can both cause the 
 bits to flip.:  = Why do you think that reducing the level of static in peoplesf datacenters E has reduced their failure rate, just a little hint reducing static iso	 unlikely '2 to have any impact on the level of gama radiation.  8 I though you said that you understood what was going on.  . > > 3.    The fixes for the above problem are. > >fK > > A.    Reduce the incidence of flipping be reducing the incidence of theoM > >       external force that causes the bits to flip, e.g make sure that thezL > >       systems are not running in an environment where there is excessive > >       heat/static. > >  > & >         Won't eliminate the problem.  H Well that depends Rob, some people have had 0 ecache failures since they fixed their environments.  > % > > B.    Install the cache scrubber.  > >s > & >         Won't eliminate the problem.  H Well that depends Rob, some people have had 0 ecache failures since they have added the cache scrubber. s > @ > > C.    Change the cache to ecache made by a different vendor. > >  > K >         Or do as your top 3 competitors have done.  Design ECC correctione5 >         in from the outset as Sun did for the UIII.e > G Sure but again if we had the failure rate in the UIII ECC cache that wetD had with the SRAM components initially used for the UII caches then 5 ECC would have helped but not eliminated the problem.e   > >.F > > For most customers actions A and B have resulted in either a major
 > > reduction-- > > in or the elimination of ecache failures.0 > >r > I >         Reduction yes.  Elmination, no.  Hence as John Shoemaker states ? >         mirror cache will soon be available for the UII CPUs.o >   ? Again Rob some customers will add mirrored caches some will nots> because some customers have zero error rates with the current A caches. In fact some customers have had zero error rates with thek older SRAM as well.   K > > Some customers cannot do much about A and so their best option is B ande > > C. > >k1 > > All this information is in my previous posts.i > >oI > > I don't know which of the three eBay applied but as they said and you F > > so kindly reproduced the course of action recommended by Sun fixed > > their problem. > >yF > > You of course have tried to rubbish all the remedial actions which= > > is I assume why you claim I havn't answered the question.  > >cK > > The problem origionally surfaced with Sun's 400 Mhz 4 MB ecache modules(J > > and at the time the SRAM supplier claimed to have rectified the issue. > >aI > > Sun replaced 400 Mhz 4 MB cache modules with new 400 Mhz 8 MB modulesnD > > only to discover that the problem that was supposed to have been
 > > rectified E > > was still there in the 8 MB modules. We are now using a differents > > supplier > > of SRAM. > >  > I >         Too bad.  Testing might have revealed that.  But the real issue F >         is ECC L2.  As you pointed out in another thread, be carefulJ >         what we say as Compaq was using the same SRAM.  But Compaq isn'tA >         having problems with it are they?  No.  ECC masks that.   F Again not necessarely ECC only protects you against a specified error B level, so if you have single bit ECC then it will not protect you  from multi-bit errors. a   I assume you knew this.i  F > > For some time Sun and our first SRAM supplier did not know why theJ > > SRAM was suceptible to having the bits flip, we have however known for > > some> > > time what the external forces are that make the bits flip. > >  > I >         Sure.  Gamma effects have been around for a while.  Hence, yournG >         wiser competitors have been using ECC L2 for quite some time.o > E > > You of course don't want to contemplate the fact that it was yourb > > FUDDINGdL > > Sun over eBay and as you appear to have admitted your ludicrous boostingH > > of Compaq future products which got me involved in this group in the	 > > firstt
 > > place. > >  > H >         No... this isn't the case.  As you pointed out in another postJ >         it was eBay issues that got you involved.  And of course, JordanK >         points out you have been posting to comp.os.vms for quite a whilelE >         before that.  So the story changes ever so slightly to say:tF >         "[Rob's] boasting got me involved."  Once again we catch youI >         changing your story in realtime mode.  You didn't co-invent the? >         Internet , did you?c >@E Of course I have Rob but only on an infrequent basis and I didn't say  onlyG eBay, I also said you ludicrous Spiralog, Galaxy, ISV support etc posts1 becausegF if you rememeber you posted many of these before you posted your eBay  FUD.  H > > Now I am here you still don't seem to have learnt your lesson as the > > Marvel/ L > > Son of WildFire posts show, at one point recently you actually published > > the C > > expected Marvel memory latencies (from a white paper no doubt).  > >o > M >         Son-of-Wildfire they are in the URL above and yes they will be thatt >         good or better.r > O >         Regarding Wildfire, what were those great latencies I published?  YousG >         see of course that most of us are willing to cut and paste toWD >         substantiate what we say.  You dredge something up with noK >         supporting evidence at all.  Recalling posts from the top of youraG >         head.  But you are right , I did post latencies and there was - >         nothing wrong with those latencies.  >   B I suggest you post that claim to Comp Arch, I don't think that youE will find anyone outside Compaq who would agree that there is nothingfA wrong with WildFire latencies unless people have had a recent andh dramatic change of heart.n  C But in case you hadn't noticed I wasn't suggesting that you posted vB current WildFire latencies, you did however post Marvel latencies.     regards2 Andrew Harrisonl Enterprise IT Architect    ------------------------------  % Date: Thu, 16 Nov 2000 16:18:06 +000010 From: andrew harrison <andrew.nospam@uk.sun.com>7 Subject: Re: Severity of eBay's ecache problem detailedp* Message-ID: <3A1408BE.FB88BE59@uk.sun.com>   jlsue wrote: > 5 > On Mon, 13 Nov 2000 15:29:51 +0000, andrew harrisono# > <andrew.nospam@uk.sun.com> wrote:c > I > >ECC would help but we are getting single and Multi-bit errors and manysI > >people would not consider multi-bit ECC for the whole cache as being aa > >reasonable proposition. > >iC > >Not everyone agrees on the necessity of having ECC for the wholesB > >of the cache either provide you can reload the offending block. > >nB > >Despite Robs FUD the fact is that people who have installed theC > >cache scrubber and either changed to the current SRAM or changed0B > >the server environment or both have seen very big reductions inC > >failure rates. In a couple of customers I know of in the UK thist > >has resulted in 0 failures. > H > Fine, Andrew, but this still doesn't explain why NO other vendor seemsF > to be having the same problems.  Only Sun.  Is Sun the only computer> > maker using that supplier?  And how does the scrubber affect > performance? >   > I don't know who uses our first SRAM supplier though I have no5 doubt that they use their own SRAM for their systems.-  A As to the question about other vendors failure rates or otherwisecB I could not comment, it is however interesting that CERN are doing5 research into exactly the problem that Sun is having.   ? The scrubber has no discernable effect on performance, we have p? run a number of benchmarks including TPC-C with and without theo. scrubber and the effect is hardly measurable.    Regards  Andrew Harrisoni Enterprise IT Architect)   ------------------------------   Date: 16 Nov 2000 17:05:38 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb) 7 Subject: Re: Severity of eBay's ecache problem detailed 0 Message-ID: <8v1452$elk$1@aquila.news.mdx.ac.uk>  ] In article <3A13FAAE.5E33639A@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:  >Rob Young wrote:  >> e >>  I >>         No... this isn't the case.  As you pointed out in another post K >>         it was eBay issues that got you involved.  And of course, JordaneL >>         points out you have been posting to comp.os.vms for quite a whileF >>         before that.  So the story changes ever so slightly to say:G >>         "[Rob's] boasting got me involved."  Once again we catch youdJ >>         changing your story in realtime mode.  You didn't co-invent the >>         Internet , did you? >>F >Of course I have Rob but only on an infrequent basis and I didn't say >onlycH >eBay, I also said you ludicrous Spiralog, Galaxy, ISV support etc posts >becauseG >if you rememeber you posted many of these before you posted your eBay   >FUD.e >h   Andrew,r  F Just because dejanews no longer allows us to easily prove how long andO frequently you have been posting doesn't mean that all our memory's have becomee faulty.i  % You have been posting here for years. K (The SIMS issue was late 1998 early 1999 and you had been posting your FUD 1, to this newsgroup a long while before that).    rM And as for being an infrequent poster - You have certainly posted a lot more h than I have.    
 David Webb VMS and unix team leader CCSS Middlesex University   ------------------------------    Date: 16 Nov 2000 12:14:22 -0500/ From: jordan@lisa.gemair.com (Jordan Henderson) 7 Subject: Re: Severity of eBay's ecache problem detailed * Message-ID: <8v14le$ini$1@lisa.gemair.com>  * In article <3A13FAAE.5E33639A@uk.sun.com>,2 andrew harrison  <andrew.nospam@uk.sun.com> wrote: >>[big snip] >> tM >>         Sure I understand the issue.  You dodge every question put to you. D >>         Ask Jordan.  Ask Kerry.... ask anyone else in this forum. >>  = >Since when has Kerry and in particular Jordan put a questionp >to me that I have dodged.   >u> >Kerrys "questions" tend to be non questions intersperced withC >marketing URLS and Jordan is obsessed with a "carefull" definition E >of FUD which would be fine except that he doesn't use the "carefull"t >definition he keeps quoting.  >   > I could quote dozens upon dozens of questions of mine that you? have dodged, but since I would like to try and keep comp.os.vms.7 on-topic and focused, I'll just point out that you haved9 studiously dodged my challenges as to where I've used thew term FUD inconsistently. o  > I challenged you to come up with any support whatsoever that I: had ever used the term FUD inconsistently in this message:  2   http://x63.deja.com/threadmsg_ct.xp?AN=690881271   and in this message:  2   http://x57.deja.com/threadmsg_ct.xp?AN=690541508  : Now, you just insist that you shown it to be the case that: I use the term inconsistently (and not even in a reply to 8 one of _my_ messages so you won't get challenged on your9 lie).  This is much like you insist that it's been proven 9 beyond all doubt that Sun is not involved in the infamous < eBay 22 hour outage, etc.  You just keep insisting it hoping< it'll be accepted, but when challenged to provide a shred of credible proof, you dodge.  7 Now, I could go on and on showing where you have dodged 7 questions that I've asked in the past, but I'd like to i9 avoid extended wrangling with you.  It's so unproductive i; because even after you are pretty much proven to be wrong,  D you just go on spouting your lies as thought they were indisputable A facts.  Besides, this isn't comp.os.vms.show.Andrew.up.as.a.liar,c> no I wouldn't hang out there, it's old news and there's really no disputing it now anyway.t  B Hosever, I won't allow you to lie about me personally as you have B above (about how I don't use the term FUD consistently) without a 
 challenge.  A >And anyway why would I want to ask the choir when I can ask the D
 >choirmaster.i >o  = I don't sing in Rob's choir and neither does Kerry.  You like = to lump us all together so that when you can show someone hasn> made some exaggeration or error that we all get blamed for it.  ? We sing pretty different songs, actually.  Rob seems to believey; you that the eBay 22 hour outage had nothing to do with Sun>= HW/SW, while I'm skeptical since you really can't show me anyi9 evidence to believe otherwise.  Rob likes to make lots ofo0 market prognostications, while I don't really.    ; Kerry pretty much sings his own song using customer success : stories and marketing materials combined with his personal( insights from his own field experiences.  < My singing is generally humorous parody.  I take what you'veA sung in the past (or perhaps things sung by your own Choirmasters>> in Sun management), show how it's discordant with what you are: singing today and join them together in a comical montage.   >> > [big snip]a >t >regards >Andrew Harrison >Enterprise IT Architect   -Jordan Hendersons jordan@greenapple.com    ------------------------------  % Date: Thu, 16 Nov 2000 17:10:40 +0000e0 From: andrew harrison <andrew.nospam@uk.sun.com>7 Subject: Re: Severity of eBay's ecache problem detailedt* Message-ID: <3A141510.45B33BCB@uk.sun.com>   Jordan Henderson wrote:o > , > In article <3A0C0865.497FBF17@uk.sun.com>,4 > andrew harrison  <andrew.nospam@uk.sun.com> wrote:) > >"Brian Schenkenberger, VAXman-" wrote:  > >>] > >> In article <vXePVDXmGW4S@eisner.decus.org>, young_r@eisner.decus.org (Rob Young) writes:v > >> >{...snip...}) > >> >       Come spin with us Andrew!!!!r > >> >( > >> >                               Rob > >> > > >>O > >> Rob, really!  Andrew's contribution here is to tout Sun and FUD Compaq andtN > >> VMS technologies.  You continue to invite him.  I know that you and KerryM > >> and several others are trying to dispell his FUD and show what a FUDster M > >> he is himself but the truth is that I believe we all know what a FUDstero, > >> he is.  Can we try to stop baiting him? > >> > > G > >How ironic, you see it was Rob's origional FUDDING of Sun over eBays F > >outage that really drew me into this group. What I found was really > >interestingE > >a whole group of people where someone could publish either totallyn > >unfoundedH > >and as it turned out totally untrue allegations about another vendorsF > >systems and no one would even question the veracity of the posting. > >e > F > Talk about untruths!  Fortunately, deja still goes way back past theF > 18 months necessary to show that you were quite active in this group@ > long before the subject of eBay's outages were discussed here. > ; Oh I had been posting before that but mainly in response to $ Robs absurd boosting of OpenVMS over? Spiralog and all the subsequent unsubstantiated and undeliverede hype.w  ' The volume simply increased after eBay.u  I > In fact, I don't think Rob has the least interest in Sun or eBay excepts# > to discredit one Andrew Harrison.l >   = Could well be true. I wonder why, could it have been because  1 he was found out over all the previous boosting ?s  7 Where does that put your support for him ? interesting a: moral dilemma, you appear to be trying to support someone ; who in your own words appears to be conducting a personnal h: crusade because he got caught out. Or does Rob have higher motives.     Regardse Andrew Harrison  Enterprise IT Architect    ------------------------------  % Date: Thu, 16 Nov 2000 17:39:29 +0000 0 From: andrew harrison <andrew.nospam@uk.sun.com>7 Subject: Re: Severity of eBay's ecache problem detailedi) Message-ID: <3A141BD1.55C99A3@uk.sun.com>n   "D.Webb" wrote:S > _ > In article <3A13FAAE.5E33639A@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:e > >Rob Young wrote:  > >> > >>K > >>         No... this isn't the case.  As you pointed out in another posthM > >>         it was eBay issues that got you involved.  And of course, JordantN > >>         points out you have been posting to comp.os.vms for quite a whileH > >>         before that.  So the story changes ever so slightly to say:I > >>         "[Rob's] boasting got me involved."  Once again we catch you.L > >>         changing your story in realtime mode.  You didn't co-invent the  > >>         Internet , did you? > >>H > >Of course I have Rob but only on an infrequent basis and I didn't say > >only J > >eBay, I also said you ludicrous Spiralog, Galaxy, ISV support etc posts
 > >becauseH > >if you rememeber you posted many of these before you posted your eBay > >FUD.y > >i > 	 > Andrew,g > H > Just because dejanews no longer allows us to easily prove how long andQ > frequently you have been posting doesn't mean that all our memory's have becomee	 > faulty.a > ' > You have been posting here for years.tL > (The SIMS issue was late 1998 early 1999 and you had been posting your FUD. > to this newsgroup a long while before that). >  >   9 When did Rob post the Spiralog rubbish to this group ????t  7 lets see could it have been about 1997-1998 seems about  right. u   regards  Andrew Harrison  Enterprise IT Architect    ------------------------------  % Date: Thu, 16 Nov 2000 17:54:43 +0000t- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>f7 Subject: Re: Severity of eBay's ecache problem detailed.( Message-ID: <3A141F63.6DB9181@bbc.co.uk>   andrew harrison wrote:   >  >bC > As to the question about other vendors failure rates or otherwise D > I could not comment, it is however interesting that CERN are doing7 > research into exactly the problem that Sun is having.   I  Are you suggesting that the sort of commercial companies that experience J these problems have particle accelerators, radiation sources and detectorsK made from exotic materials utilitizing hight voltages, and large numbers ofp channelsG of very high speed analogue and digital electronics, near their machinen rooms?E A particle physics computing environment is NOT the same as a bankingd environment,J unless they are located on top of whatever the latest Fermilab accelerator
 is called.  J Please look up a recent thread (approx 8-10 weeks back) in comp.os.vms and especially Bob Kohler's replys.l   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukc  A I speak for myself only and my views in no way represent those of  MedAS or the BBC.c   ------------------------------  % Date: Thu, 16 Nov 2000 13:09:36 -0500 5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> 7 Subject: Re: Severity of eBay's ecache problem detailedy, Message-ID: <8v17ro$76cp$1@lead.zk3.dec.com>    What a poor excuse for a "rule".  F My experience with your threads generally lead me to believe that yourG knowledge is a block wide and an inch deep.  You are certainly buzzwordeJ compatable, and know where to find lots of benchmark data dejur.  But whenL pinned down on specific technical issues you either clam up, or redirect the% discussion to previous "discussions".   L I don't claim to be a database expert, nor am I pretentious enough to appendJ "Architect" to my signature.  However, I've designed and written plenty ofD operating system code, designed parts of the underpinnings of GalaxyL (complete with a number of patents pending), written more device driver codeF that I would like, worked with Alpha CPU and IO chip designers for newF systems, and  have 22+ years experience in computers, programming, andI engineering. My title (Consulting Software Engineer, or I guess as we nowrJ call it at Compaq "Senior Member of Technical Staff") required a corporateH board approval of my credentials and experience - so I "think" I can say that I am technical.  F I think Sun is a fine company.  I have friends (former co-workers) whoK actually do O/S work there.  I think they compete well with what they have,lL and in many ways I envy them because they have resources to do things that IK *know* we could do better if we had just a few more people.  They have beenGI able to make silk purses out of pigs ears for years, and sell them in theo/ face of better systems from many other vendors.i  L The cache design for a high-reliable large scale system was just plain poor.L There isn't much excuse for it, and there isn't much spin to put on it otherH than to say "We have a fix, here it is" and "We'll make sure we don't toK this again".  A background memory "scrubber" isn't a fix, it just increasestI marginally the MTBF.  You really need to hack up a mirrored cache or sometI other kludge for the current HW.  But hey, UNIX is new to the business of K 24x376 and so this isn't really a huge sin, you're just learning.  Tru64 iseF lucky in many respects, since they get the benefit of the 25+ years of= system engineering experience that gave VMS it's reliability.u   _Fredn  B andrew harrison wrote in message <3A12B4DA.74BCA7CB@uk.sun.com>... >Fred Kleinsorge wrote:e >> >> Calling the kettle black? >> >I >This is Arnes RuleB > E >> > Arnes rule to distinguish between technical people and marektinga people:yK >> > Do they consider information about systems not shipping yet for fact ? ' >> > Yes => marketing. No => technical.a >> > > ; >I havn't made any pronouncments of "fact" about any futureh/ >systems. So I fail Arnes test for a Marketeer.  > ) >Rob on the other hand has pronounced on.( >m7 >Spiralog (when it was a lab project and a white paper)l >Galaxies (ditto)h >WildFire (ditto)  >21264 (ditto) > E >In the case of Spiralog for example all of Robs pronouncements about G >the performance gains to be had using Spiralog turned out to be untruesF >and his extrapolation to its effects on TPC-C performance also turned >out to be wildly optimistic.w > B >Now if Rob had prefaced all these pronouncments with a disclaimerD >then by Arnes definition he could still claim to be technical as itA >is he didn't and so he cannot (at least using Arnes definition).  >  >u >Regards >Andrew Harrison >Enterprise IT Architect   ------------------------------    Date: 16 Nov 2000 13:48:09 -0500/ From: jordan@lisa.gemair.com (Jordan Henderson)p7 Subject: Re: Severity of eBay's ecache problem detailed * Message-ID: <8v1a59$oh9$1@lisa.gemair.com>  * In article <3A141510.45B33BCB@uk.sun.com>,2 andrew harrison  <andrew.nospam@uk.sun.com> wrote: >Jordan Henderson wrote: >> n- >> In article <3A0C0865.497FBF17@uk.sun.com>,-5 >> andrew harrison  <andrew.nospam@uk.sun.com> wrote:m* >> >"Brian Schenkenberger, VAXman-" wrote: >> >>c^ >> >> In article <vXePVDXmGW4S@eisner.decus.org>, young_r@eisner.decus.org (Rob Young) writes: >> >> >{...snip...} * >> >> >       Come spin with us Andrew!!!! >> >> > ) >> >> >                               Rob  >> >> >d >> >>uP >> >> Rob, really!  Andrew's contribution here is to tout Sun and FUD Compaq andO >> >> VMS technologies.  You continue to invite him.  I know that you and KerryiN >> >> and several others are trying to dispell his FUD and show what a FUDsterN >> >> he is himself but the truth is that I believe we all know what a FUDster- >> >> he is.  Can we try to stop baiting him?d >> >>  >> >H >> >How ironic, you see it was Rob's origional FUDDING of Sun over eBaysG >> >outage that really drew me into this group. What I found was really  >> >interestingsF >> >a whole group of people where someone could publish either totally
 >> >unfoundeduI >> >and as it turned out totally untrue allegations about another vendorssG >> >systems and no one would even question the veracity of the posting.p >> > >>  G >> Talk about untruths!  Fortunately, deja still goes way back past the G >> 18 months necessary to show that you were quite active in this groupRA >> long before the subject of eBay's outages were discussed here.i >> r< >Oh I had been posting before that but mainly in response to% >Robs absurd boosting of OpenVMS over @ >Spiralog and all the subsequent unsubstantiated and undelivered >hype. >f  B I had many extensive debates with you going back to late 1998 and C possibly before, so I know that you were not drawn into this group b by the eBay debacle.  ( >The volume simply increased after eBay.   Possibly true.   >-J >> In fact, I don't think Rob has the least interest in Sun or eBay except$ >> to discredit one Andrew Harrison. >> e >o> >Could well be true. I wonder why, could it have been because 2 >he was found out over all the previous boosting ? >a8 >Where does that put your support for him ? interesting ; >moral dilemma, you appear to be trying to support someone o< >who in your own words appears to be conducting a personnal ; >crusade because he got caught out. Or does Rob have higheri	 >motives.  >e  1 "your support for him"... spin, spin, spin, spin    < I haven't really been supporting him, my posts to this point: have been pretty Rob neutral.  I guess I have pointed out 8 that holding you up to ridicule and humiliation for your= ernest campaign to FUD Compaq and OpenVMS here is appropriate? for this newsgroup.v   >  >Regards >Andrew Harrison >Enterprise IT Architect   -Jordan Hendersond jordan@greenapple.com,   ------------------------------   Date: 16 Nov 00 05:13:08 PST From: mckinneyj@cpva.saic.como* Subject: Re: SHOW PROC (Peak virtual size)( Message-ID: <hUVIPmcqntLg@cpva.saic.com>  3 In article <3A12F992.3F686BE2@clarityconnect.com>, p3 "Mark D. Jilson" <jilly@clarityconnect.com> writes:eG > Remember that all of your virtual address space is not mapped against I > the system pagefiles.  Some of it will be mapped against image sections G > and some against private backing store for global sections.  VMS does 8 > not track peak page file usage on a per process basis.  L Thanks Mark, this makes sense... after posting I realized that I should haveM also noted that the PQL value for WSEXTENT was less than the authorized valuerM as this could further muddy the waters (as Keith Parris correctly pointed oute in another followup).$   >   > mckinneyj@cpva.saic.com wrote: >> E) >> Environment is VMS 7.1-1H1 on Alpha...n >> n' >> Would someone enlighten me as to why> >> >* >>   $ SHOW PROCESS/ACCOUNTING/ID=xxxxxxxx >> e >> might yeild >> yF >>    Direct I/O count:      196938  Peak virtual size:         382192 >> >( >> the peak virtual page size seen above >> dF >> when the UAF record for the owner process defines the PGFLQUOTA and >> WSEXTENT as seen below? >> p> >>   Queprio:         4  TQElm:        50  WSextent:     32768> >>   CPU:        (none)  Enqlm:      2000  Pgflquo:     100000 >> uJ >> Shouldn't the peak size be limited to the sum of the pagefile usage andG >> the working set use (which are limited by the UAF quotas)? Thanks...u >> k >> --y >> - Jim >  > -- .F > Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY2 > 	- jilly@clarityconnect.com			- Brett Bodine fan0 > 	- Mark.Jilson@Compaq.com			- since 1975 or so. > 	- http://www.jilly.baka.com               - --   - Jimm --A mailto:James.F.McKinney@cpmx.SAIC.com | SAIC PSC | (858) 826-2075_   ------------------------------   Date: 16 Nov 2000 14:57:15 GMT= From: jlw@psulias.psu.edu (j.lance wilkinson, (814) 865-1818)h- Subject: Re: Still looking for JNET software.n, Message-ID: <8v0skb$1cqg@r02n01.cac.psu.edu>  k >In article <KZFQ5.14934$wW2.490149@news1.giganews.com>, Timothy Stark <sword7@grace.speakeasy.org> writes:  >>Hello folks: >>J >>I still am looking for Jnet software for my Hobbyist OpenVMS 7.2 system I >>that allow send individual messages and/or files between users locally. K >>Many years ago, I used Jnet software to send individual messages or fileshL >>through BITNET at Gallaudet University.  However, I was looking for it but/ >>was told that Jnet software was discontinued.r >> >>Thank you! >> >>-- Tim Stark  C 	Wingra Technologies bought JNET years ago.  See the BOTTOM of thiss
 	web page:  , 		http://www.wingra.com/Products/index2.html  M +-"Never Underestimate the bandwidth of a station wagon full of mag tapes"--+nN | J.Lance Wilkinson ("Lance")            InterNet:  Lance.Wilkinson@psu.edu | M | Systems Design Specialist - Lead       AT&T:      (814) 865-1818          |rM | Library Computing Services             FAX:       (814) 863-3560          |sM | E3 Paterno Library                     "I'd rather be dancing..."         |hM | Penn State University         A host is a host from coast to coast,       |uM | University Park, PA 16802     And no one will talk to a host that's close |fM | <postmaster@psulias.psu.edu>  Unless the host that isn't close            | M | VMS GopherMeister             Is busy, hung or dead.                      |tM +------"He's dead, Jim. I'll get his tricorder. You take his wallet."-------+i9                  [apologies to DeForest Kelly, 1920-1999]t3 <A Href="http://perdita.lcs.psu.edu">home page</a> iJ <a Href="http://perdita.lcs.psu.edu/junkdec.htm">junk mail declaration</a>   ------------------------------   Date: 16 Nov 2000 17:31:12 GMT* From: helbig@astro.rug.nl (Phillip Helbig) Subject: strange SMTP problemi. Message-ID: <8v15l0$q2p$1@info.service.rug.nl>  G $  SHOW QUEUE *SMTP* shows a lot of messages holding.  Usually, things tG are not in this queue more than a few seconds.  This has been going on dG for about 6 hours.  One can see them as MAILDIR:*.*_TEXT.  The strange nD thing is, only about 20% of outgoing mails are affected, apparently  independent of destination.w  G Also, OPCOM signals incoming SMTP requests from time to time, but this eE never gets delivered to an account.  (This occasionally happens when n@ spammers attempt to use my machine as a relay or something, not H realising it won't do what they want, but has been happening a lot more D recently).  Thus, I think some INCOMING mail might be getting stuck I somewhere as well.  However, if the SMTP request is reaching me, then it .I is stuck somewhere on my box, as perhaps the outgoing mail as well.  Any   ideas?   ------------------------------  # Date: Thu, 16 Nov 2000 12:44:14 GMT   From: jlsue <jlsuexxxz@home.com>' Subject: Re: Sun "uptime" belly laugher 8 Message-ID: <05l71tg9gv5hgduuh8dcdrvficj1473924@4ax.com>  3 On Tue, 14 Nov 2000 16:04:17 +0000, andrew harrisons! <andrew.nospam@uk.sun.com> wrote:u  F >> Andrew, you should have used a VMS cluster. Then you could turn offD >> the node and kept the replication running on the other node. In aG >> properly configured cluster with a decent database like RDB you telltA >> RDB on one node to shut itself down after the last transactionp> >> conpletes and not to accept any new transactions. Once thatG >> transaction completes you do wahtever work you need to do and rebootn >> the node. >> dB >That would be fine if the whole of the replication server process@ >ran on UNIX (or VMS) it however does not, part of it runs on NT& >making it too complicated to migrate. >uC >This will change as Seibel have ported the whole of Seibel to UNIX D >but my customer does not have that version which was not available ! >when they rolled out the system.r  E Then WHAT THE HELL was your point in bringing it up?!!!  The point ofeE the conversation you were responding to (from me, anyway) was whether B the capabilities of VMSclusters were really what Compaq folks wereE saying.  In response, I put forth proof that it can be done, as Kerrye> has been trying to get through your thick skull, and gave some4 information as to the environments (e.g., telnet).    F Somehow, in countering my discussion you bring up something that isn'tB even relevant?!!!  What a crock.  This is the best that Sun has to throw at us?  1 Not speaking for anyone, certainly not DEC/Compaqh- (get rid of the xxxx in my address to e-mail)i   ------------------------------  # Date: Thu, 16 Nov 2000 12:49:46 GMT   From: jlsue <jlsuexxxz@home.com>' Subject: Re: Sun "uptime" belly laugheru8 Message-ID: <qdl71tkvf9o6ug4u0nr5lj6nql94h11nqt@4ax.com>  E On Mon, 13 Nov 2000 11:21:30 -0500, rdeininger@mindspring.com (Robert  Deininger) wrote:3  X>Any work model that requires very long, uninterrupted "sessions" cannot survive the shutdown of the VMS node >where the sessions are running.  You prefer to think in terms of "sessions" that are very short, and situations where >new sessions can be routed to a different node on the fly.  Fine.  None of your arguements apply to long sessions.  E Okay, fine.  But this is such a convoluted, and fairly rare (compareda? to the rest of the universe) that is is *hardly* an issue worthwB bringing up.  The fact is, only extreme users (programmers, systemD admins) would be affected.  I know, I did this for years.  However IC had a workstation on my desk with many windows open to other nodes. E When one went down, just the xterms to that system went away.  What'st= the problem?  If I can't handle that situation I need a rest.tA Besides, I wasn't *running* anything there, I was just logged-in. ? Losing the session lost little, if anything.  Stuff I had to do F repetitively I placed in command procedures so I didn't lose anything,@ and, of course, I could always save my command recall list to be restored the next day.  1 Not speaking for anyone, certainly not DEC/Compaqt- (get rid of the xxxx in my address to e-mail)b   ------------------------------  # Date: Thu, 16 Nov 2000 12:40:07 GMTo  From: jlsue <jlsuexxxz@home.com>' Subject: Re: Sun "uptime" belly laugher.8 Message-ID: <phk71t8mc4jrhingicin5155d4efe7o8g0@4ax.com>  3 On Fri, 10 Nov 2000 15:53:53 +0000, andrew harrison ! <andrew.nospam@uk.sun.com> wrote:w  
 >jlsue wrote:l >> c   >> fG >> Typical Andrew, you misrepresent what I said, AGAIN.  What I said is3I >> that the individual users don't stay logged-on 24x7.  And, as has been G >> explained to you many, many, many times, the users will "migrate" tooG >> the other systems over time via disabling the logons; thus they willc- >> "move" transparently to the other systems.u >> eA >Its not the users directly that cause the problem, but the users 
 >worldwideI >update the master Seibel database, all the databases worldwide are kept m@ >syncronised using a replication mechnism provided by the Seibel >application.  >eA >So a seibel user in Tokyo updating a database is also updating a 	 >databaseyH >in London, Frankfurt etc while the users that access these systems are  >all logged out and asleep.o  C Okay, so what?!!  In a VMScluster application, the database updatesp@ would to to the remaining node(s) if/when new connections to the? desired node were told to stop.  The disks are shared... you dot understand that, right?    > F >Turning off a node  even if it has no users attached to it also stops >the  I >replication from running. While this isn't an unrecoverable situation ito >is F >downtime for the replication server and can cause quite a big impact.  E If it has to be explained to you again, you *really* don't understandnC the concept of a shared-everything cluster.  This is really gettinglA boring, and you really can't blame Kerry for telling you the samea story over, and over.e     >y >p >> >L >> >This forces the customer to do some remedial work, particularly where it >> >is >> >urgent during the week.r >> >K >> >Now in this case these are actual connected users but the same customer L >> >also has systems with no connected users but lots of connected processesK >> >which also run 24 hours a day. Some of the processes are extremely longaH >> >running making the cost of killing the processes higher further into >> >the process run. >> nA >> Any processing that runs over, say 10 hours, that doesn't havetI >> built-in restart capabilities, deserves all the problems that E*bay isiE >> seeing today. (I know, that's not a good example, I'm just jerkin'a >> your chain.)e >> r >> >F >> >Same for their treasury applications same for the upcoming general& >> >ledger implimentation etc etc etc. >> g$ >> These jobs run for over 10 hours? >> m > D >I thought you said you were familiar with this kind of environment. >oC >10 hours is quite short for some jobs, a general re-pricing can ineC >the customer I work for run for up to 20 hours. Now of course you -A >can restart the re-pricing, you don't lose data but you do wastel- >processing time and you may run out of time.   E DUH!  I didn't say that a 10 hour job is uncommon, did I?  Re-read mybE first paragraph.  I just put a stake in the ground at 10 hours, ymmv.s@ All I  said was that any job that runs over 10 hours should haveA re-start capabilities.  Hey, I've done programming, I *KNOW* thisr? isn't that difficult.  It is *especially* important in databasesE applications (e.g., financial institutions) where you may not want tohE commit too often (for speed), but don't want to commit the entire job B either.  In that case, if you don't write re-start capabilities to< re-do everything that hadn't been committed (but to skip theC operations that had been), then you really are in the wrong line ofI work.w     >> uC >> Sure, but they are single, persistent connections for those 24x7eG >> windows.  Also, most "good" applications will detect connection loss G >> and re-connect and re-transmit (if necessary).  This is a real worldi/ >> scenario, and networks go down all the time.a >> A >4 >7F >Ahh you used the word "good". A lot of the apps that my customer runsI >at least at the back end were written long before anyone had ever heard hI >of the WEB or at least before the first edition of Bill Gate's book :):)n >m >o  : Er... so what?  What does the WEB have to do with it?  TheB applications I had in mind were written back in the 70's and 80's.D What's your point?  They still should detect dropped connections andE reconnect.  Again, any programmer who programs in a box so small that > they don't look at the real world issues where they'll be usedE (networks have problems too, ya know), then they shouldn't be in thatr lline of work.  1 Not speaking for anyone, certainly not DEC/Compaq - (get rid of the xxxx in my address to e-mail)    ------------------------------  % Date: Thu, 16 Nov 2000 13:53:36 +0000t0 From: andrew harrison <andrew.nospam@uk.sun.com>' Subject: Re: Sun "uptime" belly laugherc* Message-ID: <3A13E6E0.61C879A4@uk.sun.com>   jlsue wrote: > G > On Mon, 13 Nov 2000 11:21:30 -0500, rdeininger@mindspring.com (Robertt > Deininger) wrote:r > Z> >Any work model that requires very long, uninterrupted "sessions" cannot survive the shutdown of the VMS node >where the sessions are running.  You prefer to think in terms of "sessions" that are very short, and situations where >new sessions can be routed to a different node on the fly.  Fine.  None of your arguements apply to long sessions. > G > Okay, fine.  But this is such a convoluted, and fairly rare (comparedFA > to the rest of the universe) that is is *hardly* an issue worthoD > bringing up.  The fact is, only extreme users (programmers, systemF > admins) would be affected.  I know, I did this for years.  However IE > had a workstation on my desk with many windows open to other nodes.wG > When one went down, just the xterms to that system went away.  What's>? > the problem?  If I can't handle that situation I need a rest. C > Besides, I wasn't *running* anything there, I was just logged-in.iA > Losing the session lost little, if anything.  Stuff I had to do H > repetitively I placed in command procedures so I didn't lose anything,B > and, of course, I could always save my command recall list to be > restored the next day. >  No it isn't rare.o  A Take something like SAP, no users connect to the backend database>< server at all except for administration purposes. The users ; connect to apps servers and the apps server connect (for a  7 very long time e.g as long as the apps server stays up)>3 to the database server. This is a common model and  4 one reason why people like Oracle are now trying to 6 charge on a CPU basis rather than a user basis because. there are no users connected to the databases.   Regards  Andrew Harrisont Enterprise IT Architect    ------------------------------   Date: 16 Nov 2000 14:57:36 GMT' From: david20@alpha2.mdx.ac.uk (D.Webb)r' Subject: Re: Sun "uptime" belly laugherm0 Message-ID: <8v0sl0$c6c$1@aquila.news.mdx.ac.uk>  ] In article <3A13E6E0.61C879A4@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes: 
 >jlsue wrote:d >> sH >> On Mon, 13 Nov 2000 11:21:30 -0500, rdeininger@mindspring.com (Robert >> Deininger) wrote: >> g >No it isn't rare. >aB >Take something like SAP, no users connect to the backend database= >server at all except for administration purposes. The users u< >connect to apps servers and the apps server connect (for a 8 >very long time e.g as long as the apps server stays up)4 >to the database server. This is a common model and 5 >one reason why people like Oracle are now trying to  7 >charge on a CPU basis rather than a user basis becauses/ >there are no users connected to the databases.t >e  K The SAP apps servers aren't written to cope with network outages etc ??????h  4 This is the real world short outages are inevitable.    
 David Webb VMS and Unix team leader CCSS Middlesex University  	 >Regards t >Andrew Harrison >Enterprise IT Architect   ------------------------------  % Date: Thu, 16 Nov 2000 14:45:14 +0000 0 From: andrew harrison <andrew.nospam@uk.sun.com>' Subject: Re: Sun "uptime" belly laugher ) Message-ID: <3A13F2FA.C222888@uk.sun.com>A   jlsue wrote: > 5 > On Tue, 14 Nov 2000 16:04:17 +0000, andrew harrison-# > <andrew.nospam@uk.sun.com> wrote:: > H > >> Andrew, you should have used a VMS cluster. Then you could turn offF > >> the node and kept the replication running on the other node. In aI > >> properly configured cluster with a decent database like RDB you tell C > >> RDB on one node to shut itself down after the last transactionp@ > >> conpletes and not to accept any new transactions. Once thatI > >> transaction completes you do wahtever work you need to do and reboot  > >> the node. > >>D > >That would be fine if the whole of the replication server processB > >ran on UNIX (or VMS) it however does not, part of it runs on NT( > >making it too complicated to migrate. > >lE > >This will change as Seibel have ported the whole of Seibel to UNIX(E > >but my customer does not have that version which was not available # > >when they rolled out the system.  > G > Then WHAT THE HELL was your point in bringing it up?!!!  The point ofmG > the conversation you were responding to (from me, anyway) was whethermD > the capabilities of VMSclusters were really what Compaq folks wereG > saying.  In response, I put forth proof that it can be done, as Kerry @ > has been trying to get through your thick skull, and gave some4 > information as to the environments (e.g., telnet). >   D Excuse me, my point which you disagree with but others don't is thatC there are long running processes that unlike users cannot easily bee. logged out and moved from one node to another.  @ I fully agree that it can be done, thats not in question but it @ can only be done with the right app and in the right environment. something that Kerry has reluctantly admitted.  B He recently added the caveat which is "if" and its an important if, you can log users/processes out of the node.  A Seibel is just an illustration of a widely used application that M> falls into the category of an application where Kerrys caveat D applies because while you can log the users out of a node you cannot> disconnect the NT replication server without causing downtime.  H > Somehow, in countering my discussion you bring up something that isn'tD > even relevant?!!!  What a crock.  This is the best that Sun has to > throw at us? >   > Oh it is relevant because it is an illustration of a commonly  used architecture.  ? Now you can argue that the architecture is a crock and that itssA difficult to build a reliable system because of it, but thats thedA whole reason why Kerrys and your argument is also a crock becauseN? a huge number of applications are not architected in a way that>, makes Kerrys scenario a viable one for them.  ? Of course you can architect apps so that they are well behaved h> but more often than not people havn't and if you can architect> an app to be resiliant in an OpenVMS environment you can also > do this with not much more work in other environments as well.   Regardso Andrew Harrison  Enterprise IT Architectk   ------------------------------  % Date: Thu, 16 Nov 2000 10:59:42 -0500 - From: "Peter Weaver" <peter.weaver@stelco.ca>c' Subject: Re: Sun "uptime" belly laugher / Message-ID: <t181ejeg41u2a0@corp.supernews.com>   = "andrew harrison" <andrew.nospam@uk.sun.com> wrote in message $ news:3A13E6E0.61C879A4@uk.sun.com... >... > Take something like SAP,  $ Don't know about SAP, never used it.  C >                          no users connect to the backend database = > server at all except for administration purposes. The users < > connect to apps servers and the apps server connect (for a9 > very long time e.g as long as the apps server stays up)-4 > to the database server. This is a common model and  E Yes, very common in the VMS world and why Clusters work great. Except=@ in the VMS world we put RDB on the Database Cluster and the appsE server makes its connections and breaks its connections as needed. WerD tell RDB on one node in the VMS cluster to close its databases afterD the last connection does its natural disconnect. Any new connectionsD to the database goes to the RDB instance on the node (or nodes) thatC still have the RDB database open. Once the final transaction on theuB database that we are closing disconnects we do whatever we want on@ that node. The users do not know, they do not need to know. Been there, done that.s  E Please Andrew, get yourself a couple of hobbyist machines, ask OracleeC for a temporary RDB license and build yourself a cluster. Come back F next week (I'm assuming you can find 3 or 4 hours between now and nextD week to install two VMS machines, setup the cluster, install RDB andA play with it since that should be all it takes - its that simple! E really, try it) and tell us what you think. Setting up RDB is nowherec9 near as complicated as setting up Oracle Parallel Server.s  E Please do not come back and tell us that Seibel or SAP or whoever has-B long running processes that means you can not disconnect from yourD Unix system. You (if I am following this thread correctly, but I mayE not be) are arguing that Kerry's method of doing work on VMS clusters+F means that he has to disconnect users, I am agreeing with Kerry that aC properly configured VMS cluster means that the users do not need tonC know that one (or more, depending on the size of the cluster) nodeseB have been taking down for maintenance. Seibel or SAP or whoever onE your Unix system does not come into it, we are saying that a VMS nodenD in a VMS cluster can be shutdown without any downtime for the users.D There may be individuals that stay connected to a machine for a longC time, but they are few and far between at sites I have worked at. IrF have spent the last 9 years at sites that operate 7x24 where 30 minuteD outages are measured in the 10's of thousands of dollars, what Kerry says can and does work.l   ------------------------------  % Date: Thu, 16 Nov 2000 16:55:10 +0000 0 From: andrew harrison <andrew.nospam@uk.sun.com>' Subject: Re: Sun "uptime" belly laugherb* Message-ID: <3A14116E.A8632AA0@uk.sun.com>   Peter Weaver wrote:i > ? > "andrew harrison" <andrew.nospam@uk.sun.com> wrote in message & > news:3A13E6E0.61C879A4@uk.sun.com... > >... > > Take something like SAP, > & > Don't know about SAP, never used it. > E > >                          no users connect to the backend databasei? > > server at all except for administration purposes. The usersn> > > connect to apps servers and the apps server connect (for a; > > very long time e.g as long as the apps server stays up)p6 > > to the database server. This is a common model and > G > Yes, very common in the VMS world and why Clusters work great. ExcepteB > in the VMS world we put RDB on the Database Cluster and the appsG > server makes its connections and breaks its connections as needed. WerF > tell RDB on one node in the VMS cluster to close its databases afterF > the last connection does its natural disconnect. Any new connectionsF > to the database goes to the RDB instance on the node (or nodes) thatE > still have the RDB database open. Once the final transaction on thenD > database that we are closing disconnects we do whatever we want onB > that node. The users do not know, they do not need to know. Been > there, done that.y > ; Of course you have but you included a the "last connection  ; does its natural disconnect" clause in your statement. What B happens if your are in the situation that Curtis found himself in.  = You have to backout a patch or add a patch but you have long w: running applications that havn't "naturally disconnected".  < The scenario you paint I can also acheive easily with Oracle< Parallel Server or Sybase OpenSwitch running on a SunCluster' or any other cluster that supports OPS.n    G > Please Andrew, get yourself a couple of hobbyist machines, ask OraclenE > for a temporary RDB license and build yourself a cluster. Come backeH > next week (I'm assuming you can find 3 or 4 hours between now and nextF > week to install two VMS machines, setup the cluster, install RDB andC > play with it since that should be all it takes - its that simple!gG > really, try it) and tell us what you think. Setting up RDB is nowhere-; > near as complicated as setting up Oracle Parallel Server.s >   G An OPS setup for HA isn't complicated to setup provided you don't want 8B it to scale which for your scenario is unescessary and since I can acheivenH your limitted definition of HA using OPS why would I bother with OpenVMSF and RDB. Isn't that the problem with your argument, if you can shut a F node down "once users have naturally disconnected" then you don't even need a cluster.>  G > Please do not come back and tell us that Seibel or SAP or whoever hastD > long running processes that means you can not disconnect from yourF > Unix system. You (if I am following this thread correctly, but I mayG > not be) are arguing that Kerry's method of doing work on VMS clusters H > means that he has to disconnect users, I am agreeing with Kerry that aE > properly configured VMS cluster means that the users do not need to E > know that one (or more, depending on the size of the cluster) nodes D > have been taking down for maintenance. Seibel or SAP or whoever onG > your Unix system does not come into it, we are saying that a VMS nodeSF > in a VMS cluster can be shutdown without any downtime for the users.F > There may be individuals that stay connected to a machine for a longE > time, but they are few and far between at sites I have worked at. IgH > have spent the last 9 years at sites that operate 7x24 where 30 minuteF > outages are measured in the 10's of thousands of dollars, what Kerry > says can and does work.e  E But you are only right if users are able to naturally disconnect fromlC the system, Curtis's Cull was nothing like that presumably because sB he could not afford to wait for a natural logout of all the users.   regards  Andrew Harrisont Enterprise IT Architectr   ------------------------------  % Date: Thu, 16 Nov 2000 13:27:53 -0500m- From: "Peter Weaver" <peter.weaver@stelco.ca> ' Subject: Re: Sun "uptime" belly laughero/ Message-ID: <t189t7oh8gav51@corp.supernews.com>u  = "andrew harrison" <andrew.nospam@uk.sun.com> wrote in messages$ news:3A14116E.A8632AA0@uk.sun.com... > Peter Weaver wrote:  > >.A > > "andrew harrison" <andrew.nospam@uk.sun.com> wrote in messagec( > > news:3A13E6E0.61C879A4@uk.sun.com... > > >...< > Of course you have but you included a the "last connection= > does its natural disconnect" clause in your statement. WhatsD > happens if your are in the situation that Curtis found himself in. >t> > You have to backout a patch or add a patch but you have long< > running applications that havn't "naturally disconnected".  F I do not know the details of the situation Curtis was in. I have neverB had a situation where I could not wait for the users to disconnect8 from one node in a cluster so I can not comment on this.  > > The scenario you paint I can also acheive easily with Oracle> > Parallel Server or Sybase OpenSwitch running on a SunCluster) > or any other cluster that supports OPS.a  F Fine, I never said that you could not. But based on your comment below< I doubt you understand the scenario. One of my sites will beE installing OPS on a VMS Cluster and then I will be able to comment onm this.l  B > > Please Andrew, get yourself a couple of hobbyist machines, ask OracleB > > for a temporary RDB license and build yourself a cluster. Come backE > > next week (I'm assuming you can find 3 or 4 hours between now andc nextD > > week to install two VMS machines, setup the cluster, install RDB andoE > > play with it since that should be all it takes - its that simple! A > > really, try it) and tell us what you think. Setting up RDB iso nowhereb= > > near as complicated as setting up Oracle Parallel Server.i > >  >nC > An OPS setup for HA isn't complicated to setup provided you don'tt wantD > it to scale which for your scenario is unescessary and since I can  @ Fine, setup OPS on your VMS hobbyist cluster and try what we areA saying. Having installed both RDB and Oracle (but not OPS) on VMS F systems I find that installing RDB (either standalone or on a cluster)" is sufficently easier than Oracle.  	 > acheivecB > your limitted definition of HA using OPS why would I bother with OpenVMSs  D Why is my definition "limitted?" I see nothing "limitted" at all. It4 is statements like this that give you the FUD label.  E > and RDB. Isn't that the problem with your argument, if you can shuty a C > node down "once users have naturally disconnected" then you don't  even > need a cluster.   F You just do not get it do you????? This is why I said above "But based? on your comment below I doubt you understand the scenario." TheeB cluster is there so the application is avalable 7x24, a particularB node may or may not be available but who cares? The application isF there for the users. Please setup your own hobbyist system and try it.  E > > Please do not come back and tell us that Seibel or SAP or whoever: hasSF > > long running processes that means you can not disconnect from yourD > > Unix system. You (if I am following this thread correctly, but I mayr@ > > not be) are arguing that Kerry's method of doing work on VMS clustersC > > means that he has to disconnect users, I am agreeing with Kerrye that aD > > properly configured VMS cluster means that the users do not need toA > > know that one (or more, depending on the size of the cluster)  nodestF > > have been taking down for maintenance. Seibel or SAP or whoever onD > > your Unix system does not come into it, we are saying that a VMS nodeA > > in a VMS cluster can be shutdown without any downtime for the* users.C > > There may be individuals that stay connected to a machine for as longE > > time, but they are few and far between at sites I have worked at.r I C > > have spent the last 9 years at sites that operate 7x24 where 30w minuteB > > outages are measured in the 10's of thousands of dollars, what Kerryr > > says can and does work.t >tB > But you are only right if users are able to naturally disconnect fromD > the system, Curtis's Cull was nothing like that presumably becauseD > he could not afford to wait for a natural logout of all the users.  D Again, I do not know exactly what his situation was. I can only tell> you that I have worked on systems configured the way Kerry hasB described and I know that what Kerry has described works. It works well.e   ------------------------------  % Date: Thu, 16 Nov 2000 12:48:26 +0000 ( From: Daniel O'Reilly <doreilly@csc.com>& Subject: Re: TCP/IP and load balancing' Message-ID: <3A13D79A.9363AB4E@csc.com>v   Mike,i  F I have just seen your note. I have done a bit of work with TCP/IP LoadJ Balancing on UCX 4.2 and TCP/IP V5. You are correct that under UCX 4.2 youH had to have to have a VMS DNS server to do load balancing. However underG TCP/IP V5 that's all changed you now have a VMS load broker system that:G communicates with a DNS server. The load broker system talks to the VMSeC systems in the cluster running the metric software (to get the load G balance), it then sends a DNS update to the DNS server to tell it which@I system should be used. The DNS server must be able to support DNS dynamic I updates.......our's does and is a Unix system. I have this working with at* local VMS cluster........hope this helps..   Dan.   Mike Price wrote:n  F > I am starting to look at using load balancing and TCP/IP rather than > LAT as we normally use.aH > I understand that it goes via DNS. - OK so the round robin bit is easy$ > but doesn't do any load balancing.I > The UCX 5.0 (a) docs talk about the load broker that talk to the metricn	 > server.oI > Unfortunatly we are still on 4.2 and I can't quite see what you need to H > do for that version. This is complicated by the fact that we are not aD > pure VMS site (OK OK I know!!) so the DNS system runs on a Solaris	 > system.gC > Can we still use the metric server on UCX 4.2 and get the SolarisiF > system to talk to it - if so what are the correct UNIX buzz words toI > say to get the right information (I have tried reading ths DNS and BIND0E > books but can't see anyting about metric server and load balancing)n >pE > The way I read it we need to set up a VMS system as a secondary DNSuG > server and get the other VMS system to use it as their DNS host. - isa? > this right?? - if so I think upgrading to 5.0 would be easiere >eF > Finally, once we upgrade to 5.0a, can the load broker talk to a UnixI > DNS system and update it the right way - it looks from the docs that itwF > can but as I am not sure of the exact mechanism on the Unix I am not > sure.n >tI > Can anyone help - sorry if this is a bit basic - I speak UCX but havn'ti" > had to do this DNS stuff before. >tA > pointers to manuals would be welcome if I have missed something  >> > As always  TIA >n > Mike > --D > All opinions are my own and do not necessarily reflect those of my
 > employer > ( > Sent via Deja.com http://www.deja.com/ > Before you buy.    ------------------------------  % Date: Thu, 16 Nov 2000 02:23:58 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>h' Subject: Re: Technology of US Electionse+ Message-ID: <3A137D7D.F818D82@videotron.ca>w   Jordan Henderson wrote:t? > And now, I apologize to comp.os.vms for participating in this ? > extended off-topic debate.  I promise not to say another word ? > here about it again.  Perhaps many of my long-winded wrangles ; > on this forum add little but text...  If I've wasted your  > time, I apologize again.    F Actually, this issue has very much to do with computing. Most computerF applications need time to be debugged and are generally repetitive andH established. You run this report every week, that bacth job every night,N paychecks every 2 weeks, and respond to customer inquiries 5000 times per day.  N But when the time comes to run something which you only run once every 4 yearsN and when you do need it, it must run fast and no time for real checks, then is$ when you risk running into problems.  L Punched card or not, bad ballot or not, if they had had a flexible computingN architecture, when the allegations of many spoiled ballots due to buhcanan andA gore being "in the same box" came out, if they had had a flexible L architecture, they should have run a job to analyse those spoiled ballots toN see what percentage of the spoiled ones had both buchanan and gore punched. IfM the percentage was roughly the same as Buchanan-Bush, or Bush-Nader, then the M authorities could have provided information that the design of the ballot dido8 not produce results against one candidate in particular.    H Also, if their ballots were numbered in some way, by rerunning the cardsI through the card reader, they could compare the results of the first readmM against that of the second read to see how many ballots were read differently  by the machine.a  M The election folks need to have the tools to not only ensure integrity of the Q data, but also tools to analyse the results to try to isolate suspected problems.a  L Someone said that it would cost $20 million bucks to outfit a county such asN Palm Beach with electronic voting. Even with VMS's exhorbitant prices, I doubt! it would come to that, would it ?   N If all of florida were to standardize on a single system, I am sure they couldQ negotiate favourable license prices, and volume discounts for VMS and DS10 boxes.e   ------------------------------  % Date: Thu, 16 Nov 2000 19:20:11 +0010.% From: paddy.o'brien@zzz.tg.nsw.gov.aue' Subject: Re: Technology of US Elections 5 Message-ID: <01JWMANZYCEA006O58@tgmail.tg.nsw.gov.au>e   J-F,  H >This reminds me of Star Trek where Kirk gives some master computer thatI >controls all of a planet a riddle the computer hasn't been programmed tooM >solve, and the computer starts to emit smoke and eventually dies with a niceaJ >show of fireworks. Seems to me that the election process has been given aM >problem it hasn't been programmed to solve and it doesn't know how to handle= >this. V  M Share the secret, was it Sun, NT or whatever the variants of windows are now t called?  :-)   Regards, Paddy   Paddy O'Brien, Transmission Development,M
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australias   Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people,c; but that little whizz-bang apostrophe gives me little spam.w   ------------------------------  % Date: Thu, 16 Nov 2000 08:38:59 +0000n8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>' Subject: RE: Technology of US electionslL Message-ID: <35666012DF4CD411BE940090279FA240111FBF@ppnt41.physics.ox.ac.uk>  D >The 17:00 deadline is statutory, it is law.  The Secretary of StateE >claims she has no discretion.  The history of such deadlines is thatwB >they are to prevent local authorities having much opportunity to H >manipulate the ballots and records to favor one candidate over another.C >In particular, there used to be a practice where local party hacksoD >would wait until they heard the results statewide and somehow theirA >local totals would be adjusted accordingly to aid their desired Y	 >result.    F If the system had been implemented to prevent one result affecting theJ outcome of another there would have been a single poll closing time acrossK the US and/or a single time for counting to start. I know it's awkward whenhK the country spans time zones but there really isn't any other way to do it.fJ The results from the East Coast must have affected the voting of the later voters on the West Coast.t  K The US elctoral system looked fine while majorities were large. Now that webL have this dead heat situation the perception from outside the US is that theJ whole system is corrupt to the core. It may be open but it's corrupt. WhenL we see one (Republican)group in Florida deciding in favour of the RepublicanI requests to stop recounting and then another (Democrat) group deciding infK favour of the Democrat request to have recounts there's only one conclusion-L an outside observer can reach. There doesn't appear to any body in the wholeC country which isn't tied to one party or the other. It would take aDK courageous individual indeed or someone of substantial independent means tos deviate from the party line.  H Other countries have their failings too but with the US claiming to be aL model for democracy some things will have to change.  Surely establishing anK electoral system where votes can be counted accurately has to be one of the  first steps?   John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)    ------------------------------  % Date: Thu, 16 Nov 2000 08:47:31 +0000e8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>' Subject: RE: Technology of US ElectionssL Message-ID: <35666012DF4CD411BE940090279FA240111FC0@ppnt41.physics.ox.ac.uk>  6 >Sorry but you're out of touch with reality.  They are9 >not counting ballots, they are counting holes in ballotse  H Even I can tell, from thousands of miles away, that they're not countingH holes. They're counting the ballots which have only one hole in them butF only after they have decided, of course, what constitutes a hole! If aL ballot has more than one hole in it, where only one hole should be, it's not counted.   John   -----Original Message-----. From: Ebinger . Eric [mailto:EEbinger@drc.com]* Sent: Wednesday, November 15, 2000 9:22 PM To: John Macallister' Subject: RE: Technology of US Electionst   ------------------------------  % Date: Thu, 16 Nov 2000 11:12:21 +0100t/ From: Nigel Arnot <sysmgr@maxwell.ph.kcl.ac.uk>r' Subject: re: Technology of US Electionst7 Message-ID: <009F3346.8D847A2D.26@maxwell.ph.kcl.ac.uk>    > >u= > >By the way, imagine the mess if a popular vote were close!d > O > That's because currently the definition of a tie in our elections is way too iL > narrow.  Most places it has to be an exact tie, but that's ridiculous whenE > dealing with millions of votes.  If it was required that the winneraJ > "convincingly beat" everybody else things would be simpler.  It would beH > considered an (N-way) tie between the highest vote candidates who wereM > within 0.5% of the candidate with the highest total vote, and the tie wouldeG > be broken by coin toss, card draw, or other unbiased method.  0.5% isp9 > arbitrary of course, it could just as well be 1% or 2%.i > M > The reason to consider this is that while a 101 to 100 win in basketball isSL > clearly a win, a 50,000,000 to 50,000,300 "victory" in an election just asM > clearly does not represent a clear signal that the voters really prefer onet > candidate over the other.  > K > Of course you could still get into a mess if the second candidate came intI > .5001% behind, but I'd also ban recounts (except for outright electoralnL > fraud) once a broader definition of a tie was in place.  It's a lot easierL > for me to accept that a candidate barely missed the tie cutoff at .5% thanH > it is to accept a win by one candidate over the other by .0001% of theL > vote, which is clearly within the statistical errors of the whole process.  J You can make this system fairer by weighting the randomness to reflect theN count. For example, you might say that a margin of less than 0.5% needs randomC determination. If A gets 50.2% and B gets 49.8%, give A a chance ofoH (50.2-49.8)/0.5 = 0.8, and B a chance of 0.2. This way there wouldn't beM any risk of a huge fuss about recounts, because they'd never be able to move a- the odds by an (a priori) significant extent.l  F Sadly, it's far too sensible for any politicians anywhere to agree to.  J A still more scientific approach would weight the odds using scientificalyL correct factors based on the normal distribution and the standard deviation O of vote-counting accuracy. But most of the electorate wouldn't understand that,t? so it's probably a less good idea even though more scientific.       	Yours,i
 		Nigel Arnote- 		NRA@MAXWELL.PH.KCL.AC.UK                   -  7 		"In the beginning there was nothing, which exploded."a   ------------------------------    Date: 16 Nov 2000 07:17:25 -0500* From: young_r@eisner.decus.org (Rob Young)' Subject: RE: Technology of US Electionsp+ Message-ID: <4cmOIgn5n3WW@eisner.decus.org>6   In article <35666012DF4CD411BE940090279FA240111FC0@ppnt41.physics.ox.ac.uk>, John Macallister <J.Macallister1@physics.ox.ac.uk> writes:E7 >>Sorry but you're out of touch with reality.  They areb: >>not counting ballots, they are counting holes in ballots > J > Even I can tell, from thousands of miles away, that they're not countingJ > holes. They're counting the ballots which have only one hole in them butH > only after they have decided, of course, what constitutes a hole! If aN > ballot has more than one hole in it, where only one hole should be, it's not
 > counted. >   = 	They are counting ballots that are "pregnant" too... i.e. a o: 	"bump" in them.  So at that point they are divining voter; 	intent.  And of course the intent is ALWAYS interpreted toa: 	mean: "I mean to punch this through but my little fingersC 	don't have quite the strength."  Instead of the just as plausible: $ 	"Should I vote for the guy?   Nah!"   				Rob    ------------------------------  # Date: Thu, 16 Nov 2000 12:52:16 GMTS= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)l' Subject: RE: Technology of US Electionsp0 Message-ID: <009F332A.9A6F1894@SendSpamHere.ORG>  X In article <4cmOIgn5n3WW@eisner.decus.org>, young_r@eisner.decus.org (Rob Young) writes: >In article <35666012DF4CD411BE940090279FA240111FC0@ppnt41.physics.ox.ac.uk>, John Macallister <J.Macallister1@physics.ox.ac.uk> writes:8 >>>Sorry but you're out of touch with reality.  They are; >>>not counting ballots, they are counting holes in ballotsB >>  K >> Even I can tell, from thousands of miles away, that they're not counting K >> holes. They're counting the ballots which have only one hole in them butaI >> only after they have decided, of course, what constitutes a hole! If a O >> ballot has more than one hole in it, where only one hole should be, it's noti >> counted.  >> n >r> >	They are counting ballots that are "pregnant" too... i.e. a ; >	"bump" in them.  So at that point they are divining votern< >	intent.  And of course the intent is ALWAYS interpreted to; >	mean: "I mean to punch this through but my little fingersnD >	don't have quite the strength."  Instead of the just as plausible:% >	"Should I vote for the guy?   Nah!"n >D >				Rob >   K Personally, I find this whole situation laughable.  The democrats have beenaK pushing the new technology agenda (internet/computer access for schools anduK the "lower classes" -- hey, their term not mine) and yet, when it suits themJ goal of pushing Al into office, technology is not to be trusted and is not "good enough".  Hippocrits.   J I don't like the idea of the hand count bias issues that have been raised.K I also don't buy the "we are way to freakin' stupid to punch out the proper K hole for the candidate we actually meant to vote for" excuse to migrate theS" Buchanan vote to 'son of Clinton'.  J I don't trust this party -- never have, never will -- nor the legal systemJ and the liberal slanted Judges that preside over it.  I'm not going to getK off on a rant of the "money buys the decision" court system here in the US.,  J I think the dems should give it up and not complain and remember the blue J dress with the stain.  Hasn't our country been embarrassed enough with the antics of this crowd?   I FYI, I'm not pro-Bush either and this thread is way way way off topic forn c.o.v. --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMu            hO city, n., 1. a place where trees are cut down and streets are named after them.t   ------------------------------    Date: 16 Nov 2000 08:24:40 -0500, From: koehler@eisner.decus.org (Bob Koehler)' Subject: Re: Technology of US Electionst+ Message-ID: <58bbQwKFquaq@eisner.decus.org>e  Z In article <8uucco$5p$3@info.service.rug.nl>, helbig@astro.rug.nl (Phillip Helbig) writes: > J > If Bush barely scrapes in, and the popular vote is still overwhelmingly K > in favour of Gore, and only 1 or 2 electors need to swing over, I'm sure u3 > they will.  They will become Heroes of Democracy.  > G > I'm sure Nader is waiting in the wings to subject the election rituall! > the same treatment he gave GM.   >   A One of the Congressional leaders is now circulating a note to allsE members of Congress that Congress can, when it approves the ElectoraleH College vote, reject the vote on a state by state basis if it feels that state's votes were suspect.o  H Somehow I don't see a Republican Congress rejecting 25 Florida electoral$ votes if they really do go for Bush.  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporationt= NASA GSFC Flight Software       | Federal Sector, Civil GroupdE                                 | please remove ".aspm" when replyinge   ------------------------------    Date: 16 Nov 2000 08:39:29 -0500, From: koehler@eisner.decus.org (Bob Koehler)' Subject: Re: Technology of US elections + Message-ID: <wvQJARGvz8Bz@eisner.decus.org>l  ] In article <3A12BADF.21905BD3@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:  > ? > Because of this it is difficult to see how Bushes position isp@ > other than anti-democratic because each count that his lawyers6 > get stopped means more voters being disenfranchised. >   E Bush trusts the people to know how to spend thier own money.  This isu4 not the same as trusting the people to vote his way.  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporationt= NASA GSFC Flight Software       | Federal Sector, Civil GroupmE                                 | please remove ".aspm" when replying)   ------------------------------    Date: 16 Nov 2000 08:42:08 -0500, From: koehler@eisner.decus.org (Bob Koehler)' Subject: Re: Technology of US electionsr+ Message-ID: <yyNLyoQfG53A@eisner.decus.org>t  \ In article <8uuhik$1bf$1@lisa.gemair.com>, jordan@lisa.gemair.com (Jordan Henderson) writes: > - > The 17:00 deadline is statutory, it is law.a  F Only the time is unambiguously law.  The law actually has it both ways" on whether or not it's a deadline.  H Now do we understand why the government buys so much Microsoft software?  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporation = NASA GSFC Flight Software       | Federal Sector, Civil Group E                                 | please remove ".aspm" when replying    ------------------------------  % Date: Thu, 16 Nov 2000 08:43:24 -0500S0 From: David Beatty <David.Beatty.NOSPAM@sas.com>' Subject: Re: Technology of US elections 2 Message-ID: <QeMTOjibOb5kNsiQGfkdK3EmwxOo@4ax.com>  4 On Thu, 16 Nov 2000 08:38:59 +0000, John Macallister( <J.Macallister1@physics.ox.ac.uk> wrote:  E >>The 17:00 deadline is statutory, it is law.  The Secretary of StateeF >>claims she has no discretion.  The history of such deadlines is thatC >>they are to prevent local authorities having much opportunity to 1I >>manipulate the ballots and records to favor one candidate over another.oD >>In particular, there used to be a practice where local party hacksE >>would wait until they heard the results statewide and somehow theirhB >>local totals would be adjusted accordingly to aid their desired 
 >>result.  >vG >If the system had been implemented to prevent one result affecting thevK >outcome of another there would have been a single poll closing time acrossbL >the US and/or a single time for counting to start. I know it's awkward whenL >the country spans time zones but there really isn't any other way to do it.K >The results from the East Coast must have affected the voting of the lateri >voters on the West Coast. >yL >The US elctoral system looked fine while majorities were large. Now that weM >have this dead heat situation the perception from outside the US is that thelK >whole system is corrupt to the core. It may be open but it's corrupt. WhengM >we see one (Republican)group in Florida deciding in favour of the Republican"J >requests to stop recounting and then another (Democrat) group deciding inL >favour of the Democrat request to have recounts there's only one conclusionM >an outside observer can reach. There doesn't appear to any body in the whole D >country which isn't tied to one party or the other. It would take aL >courageous individual indeed or someone of substantial independent means to >deviate from the party line.. > I >Other countries have their failings too but with the US claiming to be a M >model for democracy some things will have to change.  Surely establishing aniL >electoral system where votes can be counted accurately has to be one of the
 >first steps?  >h >Johni >fC >Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukaI >Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKsB >Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax) >n  C Wonder what the implications would be if all states were to go backa= to paper ballots, marked using indelible ink, hand counted inlD small blocks (say 50 or 100) by at least two different voters in theD precinct, then called into the election committee and posted for the public to see?  F Might not get the immediate results the "people" (actually, the media)@ are looking for, but I suspect the accuracy would increased, theB call for recounts would decrease, and the potential for vote fraud would decrease as well.e   Just a thought ...   ------------------------------    Date: 16 Nov 2000 08:58:20 -0500, From: koehler@eisner.decus.org (Bob Koehler)' Subject: Re: Technology of US Electionsi+ Message-ID: <Ob7Ww9AVXnxo@eisner.decus.org>o  ] In article <01JWMANZYCEA006O58@tgmail.tg.nsw.gov.au>, paddy.o'brien@zzz.tg.nsw.gov.au writes:r > J-F, > I >>This reminds me of Star Trek where Kirk gives some master computer that J >>controls all of a planet a riddle the computer hasn't been programmed toN >>solve, and the computer starts to emit smoke and eventually dies with a niceK >>show of fireworks. Seems to me that the election process has been given a+N >>problem it hasn't been programmed to solve and it doesn't know how to handle >>this.  > O > Share the secret, was it Sun, NT or whatever the variants of windows are now   > called?  :-)  * Can't be a Sun product, it was after 2038.  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporatione= NASA GSFC Flight Software       | Federal Sector, Civil GroupsE                                 | please remove ".aspm" when replyingr   ------------------------------  % Date: Thu, 16 Nov 2000 14:54:28 +0000s. From: Peter Jackson <peter.jackson@oracle.com>' Subject: Re: Technology of US electionsh* Message-ID: <3A13F524.AF56E229@oracle.com>  , This is a multi-part message in MIME format.& --------------13E62D21BDFDA115F12E1B86* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bitt   > Surely establishing anM > electoral system where votes can be counted accurately has to be one of thet > first steps?  J There will always be some error. So long as the margin is greater than theO estimated error there is no problem. If it isn't then a recount, under the sameaL rules as the original count, will give a more accurate and reduced estimatedM error. Repeat until the estimated error is less than the margin (or you get atN tied result). Who estimates the error? A fixed one such as Florida had for theN first recount is reasonable. Letting the candidates call for a recount if theyM want one as in UK elections works (providing the cost of a recount is not too P high). Allowing a different method of counting to see if rejected ballots can beN accepted when using different criteria would be OK if applied across the wholeM state (only rejected ballots need be recounted this way - putting them to one M side in the first count and then counting them automatically if the result isyN close would be fair, but the rules for the counting should be unambiguous, theO same everywhere, and preferably agreed in advance). Allowing selective recountssN by a different method, with a short time limit for requesting them favours theN candidate behind on the first count, who is slightly more likely to be the oneO who got least votes, and thus is undemocratic (democratic methods should favoure  the one who got the most votes).   Petero  & --------------13E62D21BDFDA115F12E1B86- Content-Type: text/x-vcard; charset=us-ascii;n  name="peter.jackson.vcf"K Content-Transfer-Encoding: 7bitn+ Content-Description: Card for Peter Jacksonn  Content-Disposition: attachment;  filename="peter.jackson.vcf"1   begin:vcard  n:Jackson;Peter  tel;fax:01344 415100 tel;work:0118 9249165c x-mozilla-html:FALSE org:Oracle Corporation UK Ltd.
 adr:;;;;;; version:2.19' email;internet:peter.jackson@oracle.comu fn:Peter Jackson	 end:vcardf  ( --------------13E62D21BDFDA115F12E1B86--   ------------------------------  % Date: Thu, 16 Nov 2000 11:08:46 -0500u2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Technology of US ElectionseL Message-ID: <rdeininger-1611001108470001@user-2ivea3d.dialup.mindspring.com>  I In article <8uujvt$2io$1@info.service.rug.nl>, helbig@astro.rug.nl wrote:e    lG > "Unfair at any count" might be a good title.  What's bad for General n  > Motors is bad for the country.  - If it's possible to get any more off-topic...o  w"Unsafe at Any Speed" (Nader's first big splash) was pure junk science.  There was no data to support his claims that the Corvair was unsafe.  After the customers were scared away and GM stopped making the car, many were still on the road.  It turned out (based on accident statistics) that the Corvair was about average in terms of safety, compared to other cars of the era.e  N Much of Nader's career seems to have been based on the same sort of flim-flam.  & (Yes, I've actually driven a Corvair.)   -- c Robert Deininger rdeininger@mindspring.com    ------------------------------  % Date: Thu, 16 Nov 2000 11:13:17 -0500t2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Technology of US ElectionsnL Message-ID: <rdeininger-1611001113170001@user-2ivea3d.dialup.mindspring.com>  [ In article <3A12F5A4.525F6201@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> wrote:     M > Why is a political party's selection of a leader something which the statesdP > mingle in or regulate ? Shouldn't this is completely internal to the political > party itself ?  dIt is pretty much up to the party.  The states get involved in trying to keep the elections honest, supplying the equipment and some election judges, etc.  Usually the party has to pay for everything.  Most states attempt to standardize the election dates, so both major parties vote on the same day at the same locations.  Then the parties split the cost.  E Minor parties usually don't have primaries, but there are exceptions.h   -- u Robert Deininger rdeininger@mindspring.como   ------------------------------  % Date: Thu, 16 Nov 2000 11:20:16 -0500i2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Technology of US ElectionsgL Message-ID: <rdeininger-1611001120160001@user-2ivea3d.dialup.mindspring.com>  [ In article <3A136187.F7FB487C@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> wrote:n   > Jordan Henderson wrote:,J > > It's the lawyers on Al Gore's side who are taking to the courts tryingH > > to get even dimples and "pregnant" chad validated as possible votes.G > > George Bush wants none of this subjective fishing around for votes.t > N > But if both side were to agree on a logical and realistic and reasonable wayL > to manually count ballots, things would be much better.  A dimple is not aP > clear indication of a voter's intentions. But chad is. If the recount includesO > votes where the intentions are not clear, the results are just as invalid (ifi# > not more) than the machine count.    Bingo!  That's the point.  The hand recounts are NOT just recounts.  They are deciding, ballot-by-ballot, how to "count" a ballot that was so ambiguous that the machine could not deal with it.    This is being done is selected (by Gore) counties, starting in selected (by Gore) precincts that voted overwhelmingly for Gore.  The starting precints in West Palm Beach county were something like 1000 to 1100 votes for Gore, and 50 to 60 votes for Bush.  That is the statistical "sample" that was used (by partisan Democrat election judges) to find a need to "recount" the entire county.   The 3-person election panel in one county is all Democratic.  Though one member had been telling the press he is a Republican, it turns out he is not.  Y Now do you see why Bush is objecting to this "recount" in counties selected by Democrats?w   -- l Robert Deininger rdeininger@mindspring.com    ------------------------------  % Date: Thu, 16 Nov 2000 11:29:02 -0500 2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Technology of US ElectionsAL Message-ID: <rdeininger-1611001129020001@user-2ivea3d.dialup.mindspring.com>  Z In article <3A137D7D.F818D82@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> wrote:    N > Someone said that it would cost $20 million bucks to outfit a county such asP > Palm Beach with electronic voting. Even with VMS's exhorbitant prices, I doubt# > it would come to that, would it ?n  Actually, electionic voting is probably not a good solution, at least if it requires electricity.  We can't have elections stopped because the power goes off.  There's ususally a blizzard _somewhere_ in the U.S. on election day.  So we need voting machines that don't need power.    Or do you think the poor benighted election judges, who can't count very well and aren't sure whether non-citizens are supposed to vote, can manage to start up the diesel generators when the power goes off?  Technology isn't a quick fix for our voting troubles.  The folks who will have to set it up and make it work aren't technically literate.  And we need a LOT of these folks on election day.  Our typical election judges are most often in 2 categories:  party hacks, or retired folks who enjoy a bit of excitement and gossip, not to mention some extra spending money for the day's work.  Which sort do you want setting up these fancy, high-tech schemes?   -- p Robert Deininger rdeininger@mindspring.como   ------------------------------  % Date: Thu, 16 Nov 2000 16:53:22 +0000 8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>' Subject: RE: Technology of US elections.L Message-ID: <35666012DF4CD411BE940090279FA240111FC7@ppnt41.physics.ox.ac.uk>   >> Surely establishing anhJ >> electoral system where votes can be counted accurately has to be one of theo >> first steps?   ! >There will always be some error.e  L Counting can be carried out with 100% accuracy. Measurement of anything willJ have some error. The vote as a measurement of the wishes of the electorateH has an associated error but the value of the counted "valid" vote can beL determined with 100% accuracy. Have some people really reached the stage nowG where 44 + 53 = 96.999999999999999 (Amazing accuracy to so many decimala places!) rather than 97 ?  n  L I guess the first question any new government in the US will have to face isL whether there are 50, 51, 52 or 53 states in the Union? 50-53 states appearsJ to be consistent with the uncertainty most people are willing to accept inF the counting of votes in the election. But will we ever know the finalK result of the election? Who's going to determine how the delegates from the J 52nd and 53rd states have voted? The Bush camp will declare that the votesH from the 52nd and 53rd states cannot be accepted because the election isI over and he has already won several times. The Gore camp will ignore them-J because they're haven't voted Democrat. Congress will eventually rule thatH the votes of those two states must be counted ...  as soon as anyone can
 find them ...s   John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)e   ------------------------------   Date: 16 Nov 2000 16:52:55 GMT* From: helbig@astro.rug.nl (Phillip Helbig)' Subject: Re: Technology of US Elections . Message-ID: <8v13d7$p7f$3@info.service.rug.nl>  F > Bingo!  That's the point.  The hand recounts are NOT just recounts. J > They are deciding, ballot-by-ballot, how to "count" a ballot that was so5 > ambiguous that the machine could not deal with it. i > I > This is being done is selected (by Gore) counties, starting in selectedlG > (by Gore) precincts that voted overwhelmingly for Gore.  The startingfE > precints in West Palm Beach county were something like 1000 to 1100-. > votes for Gore, and 50 to 60 votes for Bush  > J > The 3-person election panel in one county is all Democratic.  Though oneJ > member had been telling the press he is a Republican, it turns out he is > not.   > D > Now do you see why Bush is objecting to this "recount" in counties > selected by Democrats? (  > Of course, the governor and secretary of state in Florida are  republicans.   ------------------------------   Date: 16 Nov 2000 16:48:39 GMT* From: helbig@astro.rug.nl (Phillip Helbig)' Subject: Re: Technology of US electionsc. Message-ID: <8v1357$p7f$1@info.service.rug.nl>  ? In article <QeMTOjibOb5kNsiQGfkdK3EmwxOo@4ax.com>, David Beatty & <David.Beatty.NOSPAM@sas.com> writes:   E > Wonder what the implications would be if all states were to go backs? > to paper ballots, marked using indelible ink, hand counted in-F > small blocks (say 50 or 100) by at least two different voters in theF > precinct, then called into the election committee and posted for the > public to see? > H > Might not get the immediate results the "people" (actually, the media)B > are looking for, but I suspect the accuracy would increased, theD > call for recounts would decrease, and the potential for vote fraud > would decrease as well.l >  > Just a thought ...  C This is how it is done in Germany.  Almost-final results are in by sG midnight, certainly fast enough.  And even if there were more than one hI time zone in Germany, folks would think it absurd to release any results yI or even inofficial things from polls and the like before all booths have . closed.o   ------------------------------  # Date: Thu, 16 Nov 2000 17:15:45 GMT-  From: jlsue <jlsuexxxz@home.com>' Subject: Re: Technology of US electionsh8 Message-ID: <hu381t8e66uaccv2vbvhq6ck6abjvggomn@4ax.com>  3 On Tue, 14 Nov 2000 18:42:12 +0000, andrew harrison ! <andrew.nospam@uk.sun.com> wrote:   D >Democracies only survive if the people can agree on the legitimacy F >and fairness of the electorial process used to elect their govenors. F >Ignoring the questions about the electorial college the fact is that E >if the count is stopped in Florida then a question will always hang 8E >over the outcome of the election. Did the majority of the voters in :> >Florida actually vote for Bush or did the inadequacies of the6 >ballot and counting process hide a majority for Gore. > ; >To a certain extent this would not matter so much if Bush e9 >had got the majority of the popular vote, but recount ord% >not he is unlikely to overtake Gore.o  E Well, Andrew, you are pretty far off in your analysis for a couple ofc reasons:  E 1.  First, the issue of the popular vote results.  The vote is *very* C close nationwide.  200,000 votes out of the hundreds of millions of C voters is an *extremely* small margin.  The accuracy of any kind oftB polling where only a representative number of the constituents areD polled, requires that there be some margin of error.  Since we neverF get anywhere near 100% of the voting that we should have, we have thisF margin of error, and the less than 1% of difference here means that weF are well within the margin of error... conclusion:  the *real* will ofA the people could go either way, and you'd be equally correct withw@ either choice.  In that sense, it will be legitimate either way.  ? 2.  Additionally, because of the media's coverage, actual voteryE turnout was influenced significantly, meaning that continual recountsiC will never give accurate results of the "will of the people".  FromgF the news I've been watching, as soon as states were called, with pollsF still opened mind you, the turnout dropped off precipitously.  PeriodsD of voter activity that have historically been very heavy (e.g., justB after work, or just before the polls closed) were completely dead.E When voters in California learned that the state was called for Gore, B for example, many voters left.  This also adds to the idea that it( should be equally legitimate either way.  A 3.  Add to that the fact that states were being "called" for GorelA almost as soon as (or even before in come cases) the voting polls A closed.  However, when a state went for Bush, the states were not.E called for Bush for many, many minutes - in at least one case, it wasaF even hours.  This naturally biases voter turnout since nobody wants to) waste time voting for the apparent loser.7  A 4.  Also note, manual counts are notoriously bad and error prone. E Think about it:  millions and millions of votes to count by hand?  NouE way people's boredom won't get the best of them.  Not to mention that C here they want someone to "determine the will of the voter" on eachf ballot.  Give me a break.i     > : >The BBC interviewed Robert Mugabe's campaign manager last8 >week and he pointed out that if Zimbabwe had conducted < >their vote in the way that the US had then the UN observers9 >would have said that the election was not free and fair.. > ; >The BBC reporter had rather reluctantly to agree with his 0 >point.r  E Only because the BBC report is completely ignorant, as well as Robert E Mugabe.  Nobody outside of the US has any right to comment on whether<= the voting was conducted properly... particularly in war-tornnD countries that have a much worse, wide-spread fraud problem than the US has ever had.  E Basically, I laugh at anyone's ability, from outside the US, to judge'> how things are going here.  Our problems will not, and can notE paralyze the country.  We have an established way to handle this, andt@ it will all be worked out peacefully, without need for riots and
 bloodshed.  1 Not speaking for anyone, certainly not DEC/Compaq - (get rid of the xxxx in my address to e-mail)t   ------------------------------  # Date: Thu, 16 Nov 2000 17:17:52 GMTp  From: jlsue <jlsuexxxz@home.com>' Subject: Re: Technology of US elections-8 Message-ID: <89581tg0h6d5dt9ch0mlmqfuqgu3qm8mof@4ax.com>  , On Tue, 14 Nov 2000 17:10:06 -0400, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:i   >Jordan Henderson wrote:; >> The important points that you, Robert Mugabe and the BBS-7 >> reporter are missing here is that recounts are being2> >> done only in areas where Gore has strong electoral support,8 >> with a process that is without standards and is being( >> overseen by rabid Democrat partisans. >mO >I thought that the media made it quite visible that there were representativestK >from both parties present at the manual recount, and that the cameras werei0 >allowed in to verify that the process was fair.  D In some cases, there are three people, and TWO of them are democrat.   >eN >The initial recount was in fact very logical. The vote difference was below aK >certain threshold and it triggered an automatic recount. The problem, as IaS >understand it is that the automatic recount reduced the spread between the 2 guys.i  E This only covers the need for a machine recount.  And realize that iniD the three machine counts that they've done, NONE of them have agreedF with one another.  If the machine's can do it consistently, how do you recommend PEOPLE to do it.    1 Not speaking for anyone, certainly not DEC/CompaqA- (get rid of the xxxx in my address to e-mail)f   ------------------------------  # Date: Thu, 16 Nov 2000 17:42:28 GMT   From: jlsue <jlsuexxxz@home.com>' Subject: Re: Technology of US Electionst8 Message-ID: <an681ts8tp2sdvkgfsnv8r4r3nto7d0f07@4ax.com>  4 On Tue, 14 Nov 2000 17:10:07 +0000, John Macallister( <J.Macallister1@physics.ox.ac.uk> wrote:  / >> No one has come up with an error free systemh >,I >The manual count can be completely error free when counting valid balloteK >papers. Once agreement has been reached on what constitutes a valid ballott8 >paper they can be counted manually with 100% accuracy.  >e3 >Take an example. Eleven votes cast in an election.i# >1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11. # >Total votes counted. Accuracy 100%e >oE >While the accuracy may not be 100% first time round, with methodical H >counting and more than one pass it's possible to achieve 100% accuracy.L >Rememeber that manual counting usually only involves counting small bundles >at a time.l  B Um... how long do we need to go on with it before we're satisifiedF that it's 100% accurate?  If the machines didn't get the same count inF three runs, do you *really* believe that humans will be more accurate?B How long do you think each manual count will take - i.e., how muchD care would they give.  Do you really think that hundreds of millions8 of votes could be done 100% error free, by human beings?  1 Not speaking for anyone, certainly not DEC/Compaql- (get rid of the xxxx in my address to e-mail)0   ------------------------------  % Date: Thu, 16 Nov 2000 18:49:05 +0100v6 From: "Martin Knoblauch" <martin.knoblauch@compaq.com>' Subject: Re: Technology of US electionsh6 Message-ID: <8v171j$r4b$1@mailint03.im.hou.compaq.com>  - "jlsue" <jlsuexxxz@home.com> wrote in messageo2 news:hu381t8e66uaccv2vbvhq6ck6abjvggomn@4ax.com...5 > On Tue, 14 Nov 2000 18:42:12 +0000, andrew harrisoni# > <andrew.nospam@uk.sun.com> wrote:  >t >tC > 4.  Also note, manual counts are notoriously bad and error prone.eG > Think about it:  millions and millions of votes to count by hand?  NoeG > way people's boredom won't get the best of them.  Not to mention thattE > here they want someone to "determine the will of the voter" on each  > ballot.  Give me a break.' >a  K  Wrong. As stated a few times from people who know, manually counting works J in other large countries well enough. Germany has a population of about 80G million. Let that be 30 Million people of voting age. With about 50-60%lJ participation let that be 15 Million votes. Our voting bothes close at 6pmJ usually. For the federal elections, the "preliminary official results" areE announced usually a bit after midnight. those preliminary results arerK usually correct within a margin well below 1%. The final results take a bitt. longer, but are not much different most times.  L  OK, the US has more people. But not more than a factor of 2.5 or so. Manual  counting would be well possible.   >iG > Only because the BBC report is completely ignorant, as well as RobertlG > Mugabe.  Nobody outside of the US has any right to comment on whethere? > the voting was conducted properly... particularly in war-torntF > countries that have a much worse, wide-spread fraud problem than the > US has ever had. >a  H  You are correct. Of course, if the situation was the other way round, IE just do not want to imagine what kind of accusations would go towards1/ Mugabes country. But this is a different story.R  G > Basically, I laugh at anyone's ability, from outside the US, to judgee@ > how things are going here.  Our problems will not, and can notG > paralyze the country.  We have an established way to handle this, andoB > it will all be worked out peacefully, without need for riots and > bloodshed. >e  I  Sure, the world will not stand still because of a puny election. This iscJ just fact. Even if it is the election of the President of the US. And yes,L it is completely your business. We Europeans have just a lot of fun watching it :-)   Martin   ------------------------------  % Date: Thu, 16 Nov 2000 13:18:19 -0500r5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>n' Subject: Re: Technology of US Electionsl, Message-ID: <8v18c6$76ia$1@lead.zk3.dec.com>  L Absolutely party hacks.  Example: the pin-head Secretary of State in Florida is an Elector.    ' Crowley, George M. wrote in message ...eK >I thought the electors where nothing more than party 'hacks'.  Is that nots >true? >u
 >> ----------i3 >> From: Bob Koehler[SMTP:koehler@EISNER.DECUS.ORG] + >> Sent: Tuesday, November 14, 2000 3:09 PMc  >> To: INFO-VAX@LISTSERV.UGA.EDU* >> Subject: Re: Technology of US Elections >>A >> In article <8us450$36o1$1@lead.zk3.dec.com>, "Fred Kleinsorge"o( >> <kleinsorge@star.zko.dec.com> writes: >> >J >> > But in reality, they could also release the electors from their bonds >> and' >> > allow a faithless vote in Florida.  >> > >>H >> Does Florida hold the electors to their vote?  One Republican elector* >> has already said she may vote for Gore. >>I >> ----------------------------------------------------------------------"B >> Bob Koehler                     | Computer Sciences Corporation@ >> NASA GSFC Flight Software       | Federal Sector, Civil GroupH >>                                 | please remove ".aspm" when replying >> >e   ------------------------------   Date: 16 Nov 2000 10:49:52 GMT1 From: "Kari Keronen" <kari.keronen@radiolinja.fi>o Subject: Re: Using PIPE for UCX22 Message-ID: <01c04fbb$49239040$5609a8c0@rlhkikker>  ? UCX SHOW DEVICE command shows the device sockets that are named A BGnnnn. If you want to find TNA (telnet session) devices, try SHOe
 USERS/FULL instead.   -Kari-  D trdorr@my-deja.com wrote in article <8uv8fd$4q6$1@nnrp1.deja.com>...E > I am trying to search for a tna device using PIPE but have not beeni
 > successful.a > The command I am using ist2 > PIPE UCX SHOW DEV/FULL | SEARCH SYS$INPUT TNAnnn > 5 > When I issue the above, I get "no strings matched".h > What am I doing wrong?	 > Thanks,  > Tomn >  > ( > Sent via Deja.com http://www.deja.com/ > Before you buy.r >    ------------------------------  % Date: Thu, 16 Nov 2000 12:32:19 -0500 # From: John Vottero <John@MVPSI.com>  Subject: RE: Using PIPE for UCXfD Message-ID: <C15945A9D9EFCF11BA8B08002BBF1CCC0CD8BE@berry.mvpsi.com>  ; It looks like UCX isn't reporting a write error.  Try this:   ' $ DEFINE/USER SYS$OUTPUT UCX_OUTPUT.TMPl $ UCX SHOW DEV/FULLt4 $ PIPE TYPE UCX_OUTPUT.TMP | SEARCH SYS$INPUT TNAnnn  I It looks like UCX just silently ignores the write errors so the output is # only part of the result.  Try this:   ) $ PIPE UCX SHOW DEV/FULL | TYPE SYS$INPUT      > -----Original Message-----6 > From: trdorr@my-deja.com [mailto:trdorr@my-deja.com], > Sent: Wednesday, November 15, 2000 7:07 PM > To: Info-VAX@Mvb.Saic.Com. > Subject: Using PIPE for UCX  >  > E > I am trying to search for a tna device using PIPE but have not been2
 > successful.p > The command I am using isS2 > PIPE UCX SHOW DEV/FULL | SEARCH SYS$INPUT TNAnnn > 5 > When I issue the above, I get "no strings matched".r > What am I doing wrong?	 > Thanks,o > Tomr >  > ( > Sent via Deja.com http://www.deja.com/ > Before you buy.a >    ------------------------------   Date: 16 Nov 2000 18:49:10 GMT* From: bleau@umtof.umd.edu (Lawrence Bleau)) Subject: What is ASYNCWRTER system crash? ) Message-ID: <8v1a76$9v5$1@hecate.umd.edu>e  J A system of mine - a VAXstation 4000-90 - crashed today with a wierd crash message:  - ASYNCWRTER, Asynchronous write memory failure   M This is what ANAL/CRASH shows, anyway.  Doing a SHOW CRASH, the program beings" run was LINK; nothing wierd there.  L Does anyone have a clue what "asynchronous write memory failure" means?  I'mN hoping for a pointer to a specific piece of hardware to troubleshoot.  Thanks.   Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edud   ------------------------------  % Date: Thu, 16 Nov 2000 09:37:30 -0700m( From: l_ricker@lto.locktrack.com (Lorin)/ Subject: Where are Compaq's Technical Journals?d. Message-ID: <00111609373090@lto.locktrack.com>  O A recent issue of Terry's SKC, Vol. 7 Nr. 14, reported that the following link:h  3   http://209.44.181.111/CompaqTechnicalJournal.htmln  K would yield the new web-based `Compaq Technical Journal'... this may be oldmI news, but when I tried this link today, it's broken.  What's the correct,b updated link?  TIA.R  0 Lorin Ricker            Lorin.Ricker@T-NETIX.com   ------------------------------   Date: 16 Nov 2000 18:16:14 GMT2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)3 Subject: Re: Where are Compaq's Technical Journals? 6 Message-ID: <8v189e$ra9$1@mailint03.im.hou.compaq.com>  Y In article <00111609373090@lto.locktrack.com>, l_ricker@lto.locktrack.com (Lorin) writes:oP :A recent issue of Terry's SKC, Vol. 7 Nr. 14, reported that the following link: :n4 :  http://209.44.181.111/CompaqTechnicalJournal.html :(L :would yield the new web-based `Compaq Technical Journal'... this may be oldJ :news, but when I tried this link today, it's broken.  What's the correct, :updated link?  TIA.    G   The URL for the Compaq Technical Journal (CTJ) is also listed in the nG   OpenVMS FAQ.  It presently appears that the CTJ web site is off-line.d    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Thu, 16 Nov 2000 11:17:03 +0000r0 From: andrew harrison <andrew.nospam@uk.sun.com>2 Subject: Re: [Fwd: RE: Technology of US elections]* Message-ID: <3A13C22F.A68599B0@uk.sun.com>   Glenn and Mary Everhart wrote: >  > Re Andrew's post:r > 9 > no; retract your statement. Bush cannot "easily" demandm5 > a recount of the rest due to Florida law, and there  > are resource issues. >   8 That is simply not true, the Judge ruled that while the 5 Secretary of state had the right under state statute  4 to insist that the counts were in by 17:00 that the / satutes also gave her the discretion to accept e! results after the 17:00 deadline.e    ? > Besides, if any of the Gore folks gave a damn about computingo > the true totals  they wouldr@ > 1. be trying recounts all over, not only in areas their voters" >         are concentrated in, and  = And that is exactly what the Gore camp has offered, at total  < recount and that is exactly what the Bush camp has rejected.9 So they do appear to give a damn and the Bush camp don't.g  ? Now this is all academic interest to me not being a US citizen.o= But to an outsider given the rejection by the Bush camp of a f; total recount and the refusal by the Secretary of State to .; use her discretionary powers, it looks to me as if the Bush 8 camp know that they would be likely to lose Florida if a: total recount takes place and are using a technicality to  try to win.v    B > 2. be running at least double recounts in a significant fractionD >         of the areas. You cannot even try to estimate the fractionE >         of new ballots due to bias in the counters without at leasteD >         2 tries. Even then the estimate is rather model-dependent. > @ > There is no reason to complain about machine errors unless youB > believe the machine is biased toward one or the other candidate.> > (These machines are punched card readers...tough to build in > a bias there.)  > Hang on then explain the Buchanan votes, the ballot paper was < confusing and he polled much higher than average in an area = where the vast majority of the residents where lets say very p> unlikely to vote for him. He has admited this and he has also = be gracious enough to say that he does not want to have votes > added to his total where the voters did not intend to vote for him.  ; This may not be a machine error but it is a combination of t a machine and input error.= > As long as the losses are a few percent and hit all ballotsn= > regardless of direction, it just means your estimate of theo? > result is maybe 98% of the vote instead of 100%. Won't affecto@ > outcome. Counties not in the news so much have had even larger> > numbers of invalid ballots. However the "abnormal numbers of? > rejected ballots" do not exist. They are entirely normal bothr: > historically and within this election. Palm Beach county? > had 14,000 invalid ballots in 1996 with a much lower turnout.n> > Another one reported this time, that Gore did not count, had@ > 28,000 invalid ballots. The Gore people simply have publicised@ > the hell out of the invalid ballots and so on in counties they< > carried, but fine-toothing only those counties and not the< > rest has the effect of making them more important than the= > rest; it effectively disenfranchises the rest of the state.i >   ? This is all is all interesting statistical analysis but totallyaE irelevant. You cannot elect a president on the basis of a statisticale> prediction of what votes might actually be in the ballot boxes you have to count them.x  > Agree on the counting rules and count them. There is no other ; way, do it for the whole state, there is plenty of time youa: have until January. It isn't as if there is an absence of 9 Government, Clinton is still president and will be until d January.   Regardsa Andrew Harrison  Enterprise IT Architectl   ------------------------------    Date: 16 Nov 2000 07:56:13 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) 2 Subject: Re: [Fwd: RE: Technology of US elections]+ Message-ID: <LsokhyQZFxrX@eisner.decus.org>s  Z In article <3A13281D.E4675529@gce.com>, Glenn and Mary Everhart <Everhart@gce.com> writes:  @ > There is no reason to complain about machine errors unless youB > believe the machine is biased toward one or the other candidate.> > (These machines are punched card readers...tough to build in > a bias there.) c  @ If the machine fairly skips a certain percentage of the ballots,> and is located in an area of the state where one candidate has> more support than the other, it hurts that candidate more than
 the other.  5 That bias is not in the machine but in its placement.0   ------------------------------    Date: 16 Nov 2000 08:52:44 -0500, From: koehler@eisner.decus.org (Bob Koehler)2 Subject: Re: [Fwd: RE: Technology of US elections]+ Message-ID: <jvUExCcJYPLt@eisner.decus.org>r  ] In article <3A13C22F.A68599B0@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes: @ > Agree on the counting rules and count them. There is no other = > way, do it for the whole state, there is plenty of time youn< > have until January. It isn't as if there is an absence of ; > Government, Clinton is still president and will be until  
 > January. >   I No, we have until December 18.  That's when the Electoral College votes. dH If Florida isn't settled by then the Electoral College must vote without< them.  However, IMHO we're likely to know the results beforeF Thanksgiving. Only the Republicans are in a hurry ("Make them stop now while we're winning").  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporationi= NASA GSFC Flight Software       | Federal Sector, Civil GroupeE                                 | please remove ".aspm" when replyingh   ------------------------------  % Date: Thu, 16 Nov 2000 14:22:43 +0000t0 From: andrew harrison <andrew.nospam@uk.sun.com>2 Subject: Re: [Fwd: RE: Technology of US elections]* Message-ID: <3A13EDB3.F8221FC9@uk.sun.com>   Larry Kilgallen wrote: > \ > In article <3A13281D.E4675529@gce.com>, Glenn and Mary Everhart <Everhart@gce.com> writes: > B > > There is no reason to complain about machine errors unless youD > > believe the machine is biased toward one or the other candidate.@ > > (These machines are punched card readers...tough to build in > > a bias there.) > B > If the machine fairly skips a certain percentage of the ballots,@ > and is located in an area of the state where one candidate has@ > more support than the other, it hurts that candidate more than > the other. >   A Discussions abput the merits or otherwise of machine counting are A fine but I am suprised that the Bush camp is trying to uphold thei accuracy of a machine count. C  @ George W's own state prefers the manual counting method. Perhaps> the Democrats should suggest a manual recount for the whole of* Florida using the Texas counting rules !!    Regardsg Andrew Harrisont Enterprise IT Architect2   ------------------------------  % Date: Thu, 16 Nov 2000 16:10:41 +0000n0 From: andrew harrison <andrew.nospam@uk.sun.com>2 Subject: Re: [Fwd: RE: Technology of US elections]* Message-ID: <3A140700.B27C58B9@uk.sun.com>   Bob Koehler wrote: > _ > In article <3A13C22F.A68599B0@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:pA > > Agree on the counting rules and count them. There is no othere? > > way, do it for the whole state, there is plenty of time you = > > have until January. It isn't as if there is an absence ofe< > > Government, Clinton is still president and will be until > > January. > >o > J > No, we have until December 18.  That's when the Electoral College votes.J > If Florida isn't settled by then the Electoral College must vote without> > them.  However, IMHO we're likely to know the results beforeH > Thanksgiving. Only the Republicans are in a hurry ("Make them stop now > while we're winning"). > C I heard an interview with a couple of constitutional legal experts n> who said that the only non-negociable deadline is the January @ one, they both claimed that the December 18th Electorial College deadline can be moved.  @ Not that it really matters, how long would it take to hand count> all the Florida Votes (I am sure there would be no shortage of volunteers.-  B I can understand the Republican motivation, but they are in dangerA of ending up in a nightmare scenario where Bush gets to announce nB that he has won but the continuing legal process ends up forcing a? recount which he then loses. The political damage this would dov> to the Republicans would be enough to cause me to think again  if I was in his shoes.  ? The Democrats have an altogether easier decision, they have wono? the popular vote, the Republican Legal shenanigans make it lookl? as is the Republicans also think that Gore would win Florida ifsA there was a full manual recount. So they mount a legal challenge t; to the Secretary of States decision, if they lose the Bush  ; presidency will still have to labour with being a minority  7 administation who may have won in the courts but not ino the ballot boxes.s     Regards  Andrew Harrisonu Enterprise IT Architecto   ------------------------------   Date: 16 Nov 2000 16:50:10 GMT* From: helbig@astro.rug.nl (Phillip Helbig)2 Subject: Re: [Fwd: RE: Technology of US elections]. Message-ID: <8v1382$p7f$2@info.service.rug.nl>  B > George W's own state prefers the manual counting method. Perhaps@ > the Democrats should suggest a manual recount for the whole of, > Florida using the Texas counting rules !!   E Right---George W. actually signed legislation in Texas which favours a hand counts.  E I find myself agreeing with Andrew much more on politics than on VMS!e   ------------------------------  % Date: Thu, 16 Nov 2000 12:19:21 -0500u2 From: rdeininger@mindspring.com (Robert Deininger)2 Subject: Re: [Fwd: RE: Technology of US elections]L Message-ID: <rdeininger-1611001219210001@user-2ivea3d.dialup.mindspring.com>  \ In article <3A140700.B27C58B9@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> wrote:   > Bob Koehler wrote: > > a > > In article <3A13C22F.A68599B0@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:oC > > > Agree on the counting rules and count them. There is no other,A > > > way, do it for the whole state, there is plenty of time youd? > > > have until January. It isn't as if there is an absence ofh> > > > Government, Clinton is still president and will be until > > > January. > > >  > > L > > No, we have until December 18.  That's when the Electoral College votes.L > > If Florida isn't settled by then the Electoral College must vote without@ > > them.  However, IMHO we're likely to know the results beforeJ > > Thanksgiving. Only the Republicans are in a hurry ("Make them stop now > > while we're winning"). > > E > I heard an interview with a couple of constitutional legal experts :@ > who said that the only non-negociable deadline is the January B > one, they both claimed that the December 18th Electorial College > deadline can be moved. > B > Not that it really matters, how long would it take to hand count@ > all the Florida Votes (I am sure there would be no shortage of
 > volunteers.- > D > I can understand the Republican motivation, but they are in dangerC > of ending up in a nightmare scenario where Bush gets to announce  D > that he has won but the continuing legal process ends up forcing aA > recount which he then loses. The political damage this would dok@ > to the Republicans would be enough to cause me to think again  > if I was in his shoes. > A > The Democrats have an altogether easier decision, they have won> > the popular vote,e   Still being counted...  And in any case it makes no difference.  It doesn't count.  Whoever wins the popular vote will do it by such a small margin that nobody but rabid partisans will take notice.   / > the Republican Legal shenanigans make it lookCA > as is the Republicans also think that Gore would win Florida if # > there was a full manual recount. "  The legal shenanigans are mainly on the Democratic side.  They started the recounts in "safe" areas, with democrats doing the hand-counting, and quite a lot of fraud taking place.  (They aren't just "counting", they are assigning votes where none was clearly present before.)  Democrats are in court trying to legitimize this nonsense, and the other side is in court trying to stop it.a  " > So they mount a legal challenge = > to the Secretary of States decision, if they lose the Bush o= > presidency will still have to labour with being a minority s9 > administation who may have won in the courts but not inU > the ballot boxes.      A court will not determine the winner.  Our courts are very reluctant to get involved in elections.  They will confine themselves to whether or not the existing laws and procedures have been applied uniformly and fairly.     If Florida certifies it's election results for either candidate, that fellow has won Florida.  Every state always certifies it's elections; in some sense the votes don't count, only the certification.  Florida happens to specify that the secretary of state has the ultimate responsibility to do this; the current one happens to be a Republican.  All she really does is add up the certified results sent by the counties.  She has essentially no discretion.  The judge went well beyond the statute when he said she   You may have seen the Florida Attorney General in the news, denouncing the Secretary of State as a partisan, and issuing "rulings" that it is ok for the selective recounts to proceed as the Democrats requested.  In fact, the Attorney General has absolutely nothing to do with Florida elections, as a matter of law.  The Secretary of State _does_ have authority.  (The Governor also has some role, but he properly stepped aside, since he happens to be brother to one of the presidential candidates.)  A lot of our  The legislature set a hard deadline for the counties to submit their counts, and they did it because there has been big trouble in the past with some counties continuing to "revise" their results in close state-wide races, until the overall result was changed.  It is real, and people have gone to jail for it fairly recently in Florida.  The legislature tried to put in place a system that limits such abuses.  The county election boards have MUCH more discretion in deciding to recount.  Even so, some of the counties went beyond their rights.  ("Our guy lost" is the reason several counties started hand-counting at Gore's request, and that is NOT a legitimate reason in the statutes.)   There is a separate deadline in the statutes (Friday at midnight) for receiving ballots from oversees.  Those ballots will be counted and added to the other totals, and the result will determine the winners.W  1(While we appreciate political advice from the Brits, we worry that your information may be faultly across the pond.  And your instincts seem wrong quite often.  This was all settled around 1776 or so, but nevertheless I'm going to toss a tea-bag in the pond as I drive home from work.  Just in case.. :-)e   --   Robert Deininger rdeininger@mindspring.comh   ------------------------------  % Date: Thu, 16 Nov 2000 17:43:10 +0000i0 From: andrew harrison <andrew.nospam@uk.sun.com>2 Subject: Re: [Fwd: RE: Technology of US elections]* Message-ID: <3A141CAE.EFC0D551@uk.sun.com>   Phillip Helbig wrote:e > D > > George W's own state prefers the manual counting method. PerhapsB > > the Democrats should suggest a manual recount for the whole of- > > Florida using the Texas counting rules !!T > F > Right---George W. actually signed legislation in Texas which favours > hand counts. > G > I find myself agreeing with Andrew much more on politics than on VMS!c  7 Perhaps what we have here is a case of sibling rivalry.s  > George W prefers hand counts so Jeb (is that his name) I know = it sounds like a character from the Beverley Hillbillies goeso for machine counting. :):):):)     Regards  Andrew Harrison  Enterprise IT Architectd   ------------------------------  % Date: Thu, 16 Nov 2000 18:00:45 +0000-0 From: andrew harrison <andrew.nospam@uk.sun.com>2 Subject: Re: [Fwd: RE: Technology of US elections]) Message-ID: <3A1420CD.7D73A44@uk.sun.com>0   Robert Deininger wrote:r > ^ > In article <3A140700.B27C58B9@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> wrote: >  > > Bob Koehler wrote: > > > c > > > In article <3A13C22F.A68599B0@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:nE > > > > Agree on the counting rules and count them. There is no otherhC > > > > way, do it for the whole state, there is plenty of time youeA > > > > have until January. It isn't as if there is an absence ofe@ > > > > Government, Clinton is still president and will be until > > > > January. > > > >i > > >nN > > > No, we have until December 18.  That's when the Electoral College votes.N > > > If Florida isn't settled by then the Electoral College must vote withoutB > > > them.  However, IMHO we're likely to know the results beforeL > > > Thanksgiving. Only the Republicans are in a hurry ("Make them stop now > > > while we're winning"). > > >cF > > I heard an interview with a couple of constitutional legal expertsA > > who said that the only non-negociable deadline is the January D > > one, they both claimed that the December 18th Electorial College > > deadline can be moved. > > D > > Not that it really matters, how long would it take to hand countB > > all the Florida Votes (I am sure there would be no shortage of > > volunteers.l > > F > > I can understand the Republican motivation, but they are in dangerD > > of ending up in a nightmare scenario where Bush gets to announceF > > that he has won but the continuing legal process ends up forcing aC > > recount which he then loses. The political damage this would doaA > > to the Republicans would be enough to cause me to think again- > > if I was in his shoes. > >sC > > The Democrats have an altogether easier decision, they have won  > > the popular vote,  >  > Still being counted...  And in any case it makes no difference.  It doesn't count.  Whoever wins the popular vote will do it by such a small margin that nobody but rabid partisans will take notice.  > 1 > > the Republican Legal shenanigans make it lookeC > > as is the Republicans also think that Gore would win Florida ifm$ > > there was a full manual recount. > _ > The legal shenanigans are mainly on the Democratic side.  They started the recounts in "safe"eY > areas, with democrats doing the hand-counting, and quite a lot of fraud taking place.  ha > (They aren't just "counting", they are assigning votes where none was clearly present before.) o] >  Democrats are in court trying to legitimize this nonsense, and the other side is in court 1 > trying to stop it. >   D I cannot comment on the fraud allegation, though it seems a sweeping
 statement.F But the point about slection of recounts has been overtaken by events, the G Democrats have offered the Republicans a complete recount for the whole  ofH Florida without the inclusion of any of results from the existing manual	 recounts.     # > > So they mount a legal challenge > > > to the Secretary of States decision, if they lose the Bush> > > presidency will still have to labour with being a minority; > > administation who may have won in the courts but not ini > > the ballot boxes.h > [ > A court will not determine the winner.  Our courts are very reluctant to get involved in rc > elections.  They will confine themselves to whether or not the existing laws and procedures have .$ > been applied uniformly and fairly.  G Quite, so what happens if the say the supreme court rules that existinga
 rules have; not been applied uniformly and fairly and orders a recount.o  G The court are not deciding the outcome of the election, but the outcomet ofI a recount will then decide the election. This is what as an outsider the e Bush camp seems afraid of.   > a > If Florida certifies it's election results for either candidate, that fellow has won Florida.  -] > Every state always certifies it's elections; in some sense the votes don't count, only the aZ > certification.  Florida happens to specify that the secretary of state has the ultimate ^ > responsibility to do this; the current one happens to be a Republican.  All she really does ] > is add up the certified results sent by the counties.  She has essentially no discretion.   _ > The judge went well beyond the statute when he said she had to have a _reason_ to reject latea- > results.  She ended up having good reasons.)  H Did she, the impression we have over here is that the reasoning upermost in her mind wasa( her affiliation to the Republican party.   Regardst Andrew Harrisoni Enterprise IT Architectr   ------------------------------  # Date: Thu, 16 Nov 2000 18:04:39 GMTt+ From: Bru, Pierre <Pierre.Bru@spotimage.fr>  Subject: [Q] MadGoat FTP) Message-ID: <8v17jh$nbi$1@nnrp1.deja.com>e   hi,j  A does MadGoat FTP server implements error recovery and restart (itY4 musr implements the REST command if memory serves) ?  @ does it understand Un*x like syntax (for ex. RETR /myfile.dat) ?   TIA  Pierre.u    & Sent via Deja.com http://www.deja.com/ Before you buy.f   ------------------------------   End of INFO-VAX 2000.641 ************************