1 INFO-VAX	Mon, 30 Oct 2006	Volume 2006 : Issue 597       Contents: Advanced Server problem  Re: Advanced Server problem ! Re: Another Vax/Alpha/VMS triumph 1 Re: Banking security Was:PARSEC: SPAMMERS (not !) + Re: DS10L hanging problem, tracking it down + Re: DS10L hanging problem, tracking it down  Re: Fired Twice (a little tail)  Re: Fired Twice (a little tail)  Free cool VMS Email account  Re: Free cool VMS Email account  Re: Free cool VMS Email account  Re: Free cool VMS Email account - Free MicroVAX 3100-30 computers for hobbyists . getting write lock error while formatting tape Re: HP support* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server* Re: Java Applet pulled from VMS Web Server@ Java vs. Javascript (was Java Applet pulled from VMS Web Server)P Re: LSE Language Definitions for Fortran 90 or Fortran 95 Free-Format Source? So Re: Mylex controller update 4 Re: Observations on Samba for VMS (CIFS Kit from HP)4 Re: Observations on Samba for VMS (CIFS Kit from HP)4 Re: Observations on Samba for VMS (CIFS Kit from HP)( OpenVMS v7.3-2 Bug in SORT/MERGE utility, Re: OpenVMS v7.3-2 Bug in SORT/MERGE utility PAKs for Itanium Re: PAKs for Itanium Re: PAKs for Itanium1 Re: simple improvements to disk write performance 1 Re: simple improvements to disk write performance 1 Re: simple improvements to disk write performance 1 Re: simple improvements to disk write performance 1 Re: simple improvements to disk write performance 1 Re: simple improvements to disk write performance 1 Re: simple improvements to disk write performance  Re: Spit Brook location  Re: Spit Brook location  Re: Spit Brook location  Re: Spit Brook location  Re: Spit Brook location  Re: Spit Brook location  Re: Time change questions ! % Re: Updated VMS Information (big one)  Re: VMS installation crash Re: VMS installation crashP why some times after inserting the tape in tape drive, comes out automatically ?P Re: why some times after inserting the tape in tape drive, comes out automaticalP Re: why some times after inserting the tape in tape drive, comes out automaticalP Re: why some times after inserting the tape in tape drive, comes out automaticalP Re: why some times after inserting the tape in tape drive, comes out automatical  F ----------------------------------------------------------------------    Date: 30 Oct 2006 03:52:35 -0800% From: "Kari" <kari.keronen@digita.fi>   Subject: Advanced Server problemC Message-ID: <1162209155.052401.224160@h48g2000cwc.googlegroups.com>   ? I get strange error when adding a new share to Advanced Server:   * TKJR\\AHMA> add share tkjv3 c3$disk:[tkjv]/ %PWRK-E-ERRADDSHARE, error adding share "TKJV3" = -LM-E-NERR_UNKNOWNDEV, the device or directory does not exist    $ sho log *c3*     (LNM$SYSTEM_TABLE)      "C3$DISK" = "_$1$DGA106:[C3.]"   $ dire c3$disk:[000000]    Directory C3$DISK:[000000]  
 TKJV.DIR;1   Total of 1 file.  ? Disk $1$DGA106: (AHMA), device type HSV110, is online, mounted, 
 file-oriented G     device, shareable, device has multiple I/O paths, served to cluster  via MSCP%     Server, error logging is enabled.   <     Error count                    0    Operations completed   19193 1     Owner process                 ""    Owner UIC  [SYSTEM]0     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G:R,W ;     Reference count                3    Default buffer size      512 0     Current preferred CPU Id       1    Fastpath       1 ;     WWID   01000010:6005-08B4-0001-54CC-0001-0000-00BB-0000 9     Total blocks           209715200    Sectors per track      128 ;     Total cylinders            12800    Tracks per cylinder      128 <     Logical Volume Size    209715200    Expansion Size Limit	 330141696 $     Allocation class               1  >     Volume label            "AS-NET"    Relative volume number       0 9     Cluster size                 201    Transaction count        1 =     Free blocks            209512953    Maximum files allowed   519097 3     Extend quantity                5    Mount count        2 2     Mount status              System    Cache name "_$1$DGA100:XQPCACHE" F     Extent cache size             64    Maximum blocks in extent cache 20951295>     File ID cache size            64    Blocks in extent cache       0 D     Quota cache size               0    Maximum buffers in FCP cache    3580 0     Volume owner UIC           [1,1]    Vol Prot S:RWCD,O:RWCD,G:RWCD,W:RWCD   G   Volume Status:  ODS-5, subject to mount verification, file high-water  marking,!       write-back caching enabled. !   Volume is also mounted on SUSI.     - OpenVMS 7.3-2, Advanced Server V7.3A. Ideas ?    -Kari-   ------------------------------    Date: 30 Oct 2006 07:15:13 -0800' From: "benitos" <benitos@stcc.cc.tx.us> $ Subject: Re: Advanced Server problemC Message-ID: <1162221313.234928.170540@k70g2000cwa.googlegroups.com>    Hello Kari,   B Since you have the disk defined as a logical and not a dsa device,* Advance Server might not look at logicals., Try creating the share using the dga device.  $  ADD SHARE tkjv3 $1$DGA106:[C3.tkjv]   Let me know if this helps.       Kari wrote: A > I get strange error when adding a new share to Advanced Server:  > , > TKJR\\AHMA> add share tkjv3 c3$disk:[tkjv]1 > %PWRK-E-ERRADDSHARE, error adding share "TKJV3" ? > -LM-E-NERR_UNKNOWNDEV, the device or directory does not exist  >  > $ sho log *c3* >  >  > (LNM$SYSTEM_TABLE) > " >   "C3$DISK" = "_$1$DGA106:[C3.]" >  > $ dire c3$disk:[000000]  >  > Directory C3$DISK:[000000] >  > TKJV.DIR;1 >  > Total of 1 file. > A > Disk $1$DGA106: (AHMA), device type HSV110, is online, mounted,  > file-oriented I >     device, shareable, device has multiple I/O paths, served to cluster 
 > via MSCP' >     Server, error logging is enabled.  > > >     Error count                    0    Operations completed	 >   19193 3 >     Owner process                 ""    Owner UIC 
 > [SYSTEM]2 >     Owner process ID        00000000    Dev Prot > S:RWPL,O:RWPL,G:R,W = >     Reference count                3    Default buffer size 	 >     512 2 >     Current preferred CPU Id       1    Fastpath	 >       1 = >     WWID   01000010:6005-08B4-0001-54CC-0001-0000-00BB-0000 ; >     Total blocks           209715200    Sectors per track 	 >     128 = >     Total cylinders            12800    Tracks per cylinder 	 >     128 > >     Logical Volume Size    209715200    Expansion Size Limit > 330141696 & >     Allocation class               1 > @ >     Volume label            "AS-NET"    Relative volume number	 >       0 ; >     Cluster size                 201    Transaction count 	 >       1 ? >     Free blocks            209512953    Maximum files allowed 	 >  519097 5 >     Extend quantity                5    Mount count 	 >       2 4 >     Mount status              System    Cache name > "_$1$DGA100:XQPCACHE" H >     Extent cache size             64    Maximum blocks in extent cache
 > 20951295@ >     File ID cache size            64    Blocks in extent cache	 >       0 F >     Quota cache size               0    Maximum buffers in FCP cache	 >    3580 2 >     Volume owner UIC           [1,1]    Vol Prot > S:RWCD,O:RWCD,G:RWCD,W:RWCD  > I >   Volume Status:  ODS-5, subject to mount verification, file high-water 
 > marking,# >       write-back caching enabled. # >   Volume is also mounted on SUSI.  >  > / > OpenVMS 7.3-2, Advanced Server V7.3A. Ideas ?  >  > -Kari-   ------------------------------  % Date: Mon, 30 Oct 2006 20:41:44 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> * Subject: Re: Another Vax/Alpha/VMS triumph1 Message-ID: <ei4rms$etb$1@news-02.connect.com.au>    Hi JF,  H > Was there ever a consideration for developing an X-windows front end ?  L My understanding is that this was tried but the X-windows front-end made the< Alpha Servers run so hot that the sprinklers kept coming on.   Cheers Richard Maher  : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message2 news:53ec8$4543eff0$cef8887a$27735@TEKSAVVY.COM... > Sharon wrote: H > > Ha!  Actually we have!  We have one amazing Windows programmer who's@ > > been tasked with designing and implementing a client for us. > H > Was there ever a consideration for developing an X-windows front end ?   ------------------------------  % Date: Mon, 30 Oct 2006 23:17:03 +0800  From: prep@prep.synonet.com : Subject: Re: Banking security Was:PARSEC: SPAMMERS (not !)0 Message-ID: <871wopbzps.fsf@k9.prep.synonet.com>  1 Paul Sture <paul.sture.nospam@hispeed.ch> writes:   C > Prior to European-wide acceptance of debit cards, there were Euro D > Cheques. Now discontinued, but very handy in their day - you couldE > pop into another country for an evening meal and write a cheque out  > in the local currency.  % Not in the UK you couldn't as a rule.   ' But then it isn't really in EU is it :)    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 30 Oct 2006 08:02:22 -0800( From: "Rich Jordan" <jordan@ccs4vms.com>4 Subject: Re: DS10L hanging problem, tracking it downC Message-ID: <1162224142.931718.213170@e64g2000cwd.googlegroups.com>   # Island Computers, D B Turner wrote: * > Sounds like a power supply problem to me/ > We have seen this several times when it was.. % > We have them in stock for about $89  >  > DT >  > -- >   C Well, at first it looked like the new power supply helped but we're F right back to the same problem.  With the new power supply in place itG ran a successful self test for 24 hours (no errors logged), then booted E and ran for about 36 hours.  Then it hung.  I issued a reset from RMC G and it ran for a couple of  hours then hung.  Would not reset again, so < I powercycled.  Then it ran for about 1 hour and hung again.  G I then moved it from the APC BackUPS Pro its normally plugged into to a G standalone power conditioner (also APC, but not a UPS).  Ran 1 hour and  hung again.   G There are still no entries in the errorlog or operator log, and nothing  on the console when it happens.   A System temp recorded every 5 minutes shows 33C very steadily (its $ cooler in the house than last time).  > DS10L 6/466, 1GB RAM (Infineon DIMMs on risers), KZPBA-CA UWSEE controller, 4.5GB UWSE internal disk (7200RPM), OpenVMS V8.2, patched G up to 6/2006, external BA364 tower with two additional disks and CDROM.   ( Next step; change the KZPBA for a spare.  E If that doesn't work try moving the internal SCSI disk to an external > enclosure (actually replace it with another SBB in the BA364).  C If that doesn't work pull the CPU heat sink again (I pulled it last B time to reveal a damaged grafoil pad, which was replaced with highG quality heat sink grease) to make sure there's good contact between the D CPU and the heat sink (and verify its not getting too hot via grease color checking).  " If that doesn't work... what next?   Thanks for any advice or ideas.    Rich   ------------------------------  % Date: Mon, 30 Oct 2006 12:31:02 -0500 ' From: Dave Froble <davef@tsoft-inc.com> 4 Subject: Re: DS10L hanging problem, tracking it down9 Message-ID: <uZ-dnS-wu8eTqtvYnZ2dnUVZ_sGdnZ2d@libcom.com>    Rich Jordan wrote:% > Island Computers, D B Turner wrote: + >> Sounds like a power supply problem to me 0 >> We have seen this several times when it was..& >> We have them in stock for about $89 >> >> DT  >> >> --  >> > E > Well, at first it looked like the new power supply helped but we're H > right back to the same problem.  With the new power supply in place itI > ran a successful self test for 24 hours (no errors logged), then booted G > and ran for about 36 hours.  Then it hung.  I issued a reset from RMC I > and it ran for a couple of  hours then hung.  Would not reset again, so > > I powercycled.  Then it ran for about 1 hour and hung again. > I > I then moved it from the APC BackUPS Pro its normally plugged into to a I > standalone power conditioner (also APC, but not a UPS).  Ran 1 hour and 
 > hung again.  > I > There are still no entries in the errorlog or operator log, and nothing ! > on the console when it happens.  > C > System temp recorded every 5 minutes shows 33C very steadily (its & > cooler in the house than last time). > @ > DS10L 6/466, 1GB RAM (Infineon DIMMs on risers), KZPBA-CA UWSEG > controller, 4.5GB UWSE internal disk (7200RPM), OpenVMS V8.2, patched I > up to 6/2006, external BA364 tower with two additional disks and CDROM.  > * > Next step; change the KZPBA for a spare. > G > If that doesn't work try moving the internal SCSI disk to an external @ > enclosure (actually replace it with another SBB in the BA364). > E > If that doesn't work pull the CPU heat sink again (I pulled it last D > time to reveal a damaged grafoil pad, which was replaced with highI > quality heat sink grease) to make sure there's good contact between the F > CPU and the heat sink (and verify its not getting too hot via grease > color checking). > $ > If that doesn't work... what next? > ! > Thanks for any advice or ideas.  >  > Rich >   E It could be anything.  No, don't yell at me for stating the obvious.  G All I'm saying is that finding intermittent problems is a real pain in  I the ass.  At some time you have to ask yourself, "Hey, I can get another  @ system for $200 plus shipping, how much more time will I waste?"  D You could have a bad CPU, bad memory, or motherboard component that < isn't completely broke.  (Completely broke would be better.)   --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Mon, 30 Oct 2006 08:01:22 -0500 * From: "FredK" <fred.nospam@nospam.dec.com>( Subject: Re: Fired Twice (a little tail)* Message-ID: <4545f7a3@usenet01.boi.hp.com>  5 "Bill Todd" <billtodd@metrocast.net> wrote in message A news:w8OdnTL6l5rNn9jYnZ2dnUVZ_v6dnZ2d@metrocastcablevision.com...  > Neil Rieck wrote:  > G > And I do get the impression that they're 99.9+% worthless, making the H > search for useful content there a pretty poor use of one's time.  ThisH > blog, for example, appears to be written by a Johnny-One-Note with tooI > much time on his hands and apparently (from the paucity of comments) no I > audience to speak of, let alone one sufficiently interested to create a E > less boring environment:  that may be useful catharsis for the blog @ > owner, but from the viewpoint of the Internet it's just noise. >   G It would seem to make you a perfect candidate for having your own blog.    ------------------------------  % Date: Mon, 30 Oct 2006 13:44:52 -0500 ( From: Bill Todd <billtodd@metrocast.net>( Subject: Re: Fired Twice (a little tail)G Message-ID: <L-ednfaSEtK51dvYnZ2dnUVZ_qOdnZ2d@metrocastcablevision.com>    FredK wrote:7 > "Bill Todd" <billtodd@metrocast.net> wrote in message C > news:w8OdnTL6l5rNn9jYnZ2dnUVZ_v6dnZ2d@metrocastcablevision.com...  >> Neil Rieck wrote: >>H >> And I do get the impression that they're 99.9+% worthless, making theI >> search for useful content there a pretty poor use of one's time.  This I >> blog, for example, appears to be written by a Johnny-One-Note with too J >> much time on his hands and apparently (from the paucity of comments) noJ >> audience to speak of, let alone one sufficiently interested to create aF >> less boring environment:  that may be useful catharsis for the blogA >> owner, but from the viewpoint of the Internet it's just noise.  >> > I > It would seem to make you a perfect candidate for having your own blog.   H Nah - my observations tend to generate *lots* of interesting discussion G (though admittedly sometimes also including the occasional incompetent   response such as yours).   - bill   ------------------------------  % Date: Mon, 30 Oct 2006 08:45:52 -0500 > From: "Island Computers, D B Turner" <dturner-at-islandco.com>$ Subject: Free cool VMS Email account0 Message-ID: <12kc0ghl5mf8pea@news.supernews.com>  H If you want a  vms-oriented email address (for non usenet use only) then send us a request 9 We will issue a password for  YOURNAME(at)VMSUSER(dot)COM      --   Island Computers US Corp 2700 Gregory St  Savannah GA 31404  Tel: 912 447 6622 x201# Mail: dturner-atnospam-islandco-com % (You know what to do with the dashes)    ------------------------------    Date: 30 Oct 2006 05:58:42 -0800  From: "Ian Miller" <ijm@uk2.net>( Subject: Re: Free cool VMS Email accountC Message-ID: <1162216722.901526.182300@f16g2000cwb.googlegroups.com>   0 You can get a openvms-rocks.com email address at http://www.openvms-rocks.com/    ------------------------------  + Date: Mon, 30 Oct 2006 16:58:34 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) ( Subject: Re: Free cool VMS Email account( Message-ID: <ei5avp$ut9$1@pcls4.std.com>  @ "Island Computers, D B Turner" <dturner-at-islandco.com> writes:  
 >That's great   K >Our vmsuser(dot)com server is set up for users of vms (in a webmail format D >also) that want to share ideas (just like redhat's) and information2 >without being hounded by the normal usenet creeps  I Why not set up VAXNotes/DECNotes or whatever the final name was, and some 9 captive accounts to use it?  Or is that too last century?    ------------------------------    Date: 30 Oct 2006 09:42:26 -0800  From: "Ian Miller" <ijm@uk2.net>( Subject: Re: Free cool VMS Email accountB Message-ID: <1162230146.050901.168080@m7g2000cwm.googlegroups.com>   Michael Moroney wrote:K > Why not set up VAXNotes/DECNotes or whatever the final name was, and some ; > captive accounts to use it?  Or is that too last century?   7 http://www.encompasserve.org/conferences/VMS_new_1.html    ------------------------------    Date: 30 Oct 2006 09:45:43 -0800" From: "Tad" <twinters@netzero.net>6 Subject: Free MicroVAX 3100-30 computers for hobbyistsC Message-ID: <1162230343.082998.267030@h48g2000cwc.googlegroups.com>   < I have 2 MicroVAX 3100 model 30 systems I need to get rid of immediately.  See this posting: 5 http://portland.craigslist.org/mlt/sys/227684599.html    ------------------------------    Date: 30 Oct 2006 04:56:06 -0800! From: "alok" <alok.net@gmail.com> 7 Subject: getting write lock error while formatting tape C Message-ID: <1162212966.520330.282870@m73g2000cwd.googlegroups.com>    Hi  F I am using Initialize mka700: job1 command to format tape but it gives wrtie lock error. 
 Plz help me .    ------------------------------  % Date: Sun, 29 Oct 2006 20:07:52 -0500 ' From: Dave Froble <davef@tsoft-inc.com>  Subject: Re: HP support 9 Message-ID: <XJSdnWScyJQ_zdjYnZ2dnUVZ_vWdnZ2d@libcom.com>    DaveG wrote: > Dave Froble wrote:0 >> That rumor rag, The Inquirer, is at it again. >>8 >> http://www.theinquirer.net/default.aspx?article=35357 >>H >> IT LOOKS LIKE Steve Smith, VP of Services, resigned from HP yesterday@ >> with no explanation, and with no immediate successor planned. >>E >> People on the inside describe it as another of the 'rats leaving a E >> sinking ship', but particular phrase was not in the internal memo.  >>D >> One has to wonder.  Was Mr Smith discourged by the downgrading ofI >> services?  Was Mr Smith the instigator of said downgrading, and seeing B >> the impending result, jumping ship before the shit hit the fan? >>! >> Or possibly, personal reasons.  >> >> -- 7 >> David Froble                       Tel: 724-529-0450 A >> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com  >> DFE Ultralights, Inc. >> 170 Grimplin Road >> Vanderbilt, PA  15486 > G > All good questions.  Trouble is we'll never know the real story.  But < > there will probably be some here that will speculate.  ;-) > 	 > Dave...  >   H I've been contacted by a member of the IT press.  Problem is, I'm not a F current software support customer.  I'm not a good source to talk to. B I'm posting the exchanges I've had with this individual, with his I permission.  The story will be told, but the question is, how will it be   told.   F I'd suggest that current software support customers contact this guy, G and give him 'on the record' or 'off the record' statements, whichever  ( is more appropriate for each individual.   -------------------------    Dave,   D Thanks for getting back to me. I've started contacting customers to E gather material for this story. Here is a copy of the note that I've  G sent, at least the first version of it. Please feel free to forward it  E to whomever you think may have an interest in this topic. To build a  I solid and defensible story I'll need to get in touch with as many people   as possible     >>     H We are trying to assess the state of Hewlett-Packard's customer support 8 as it shifts some of its support operations to overseas.  F Complaints are turning up on technical message boards and the HP user H group Interex in Europe (Which didn't shut down unlike the U.S. group.) H is conducting an online survey of members. We have also been talking to % HP independent consultants in the EU.   F The only way for to assess the state of HP support is to reach out to : customers and solicit their feedback. Your help is needed.  F The questions:  Over the past year, has there been changes in dealing I with HP customer support -- What changes have you seen and what has been   the impact of these changes?. Have you raised any concerns with HP about it?  ; I'm gathering this material for a story in Computerworld.<<   , Thanks for your note and background on this.   Regards, Pat      Patrick Thibodeau 
 Senior Editor 
 Computerworld 
 Washington DC  o: 202 333 2448  c: 202 361 2011          Dave <davef@tsoft-inc.com>   10/27/2006 07:52 PM  	  To$ 	Patrick_Thibodeau@computerworld.com cc 	  Subject / 	Re: Computerworld inquiry regarding HP service    	         * Patrick_Thibodeau@computerworld.com wrote:  >  > Dave,  >I  > I've been reading the traffic on various groups about service for VMS, K  > in particular, but other systems as well and saw your post about service 
  > issues.  >H  > Can you help me out so I can understand and report on what's going on3  > with HP services and support over the last year?   >D  >  I'm trying to get in touch with users who can share experiences.  >  > Regards,   >  > Patrick Thibodeau  > Senior Editor  > Computerworld  > Washington DC  > o: 202 333 2448  > c: 202 361 2011  >  F As for the HP support issue, I may not be the best person to ask aboutB current issues since I do not now have a support contract with HP.E However, there are complaints posted to comp.os.vms by people who are D current support subscribers.  Some of these people appear to be VERYF unhappy.  Whether you can get an 'official' statement from all of themF may depend upon each individual's perception of what they feel they're 'allowed' to say.   C To really appreciate the issue, you must look at the history.  Some D people feel that the software organization developed by DEC (DigitalG Equipment Corporation) in the 1970s and 1980s was perhaps the best such F organization the world has ever seen.  Between development and support they were excellent.  D With first Bob Palmer at DEC, and then Compaq, and now HP, there hasH been a steady degradation of these organizations.  The latest chapter isC the sending of most software support calls to new centers in India, H Costa Rica, and another third world country who's name escapes me at the moment.   E Bottom line, people with 20-30 years of experience are being replaced G with people who have had several weeks of training.  All this to save a B few bucks to allow bigger executive bonuses and golden parachutes.  J I suggest you attempt to talk to customers with current support contracts.   --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Mon, 30 Oct 2006 18:39:42 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 3 Subject: Re: Java Applet pulled from VMS Web Server 1 Message-ID: <ei4ki4$4o4$1@news-02.connect.com.au>    Hi Arne,  B > I am not good at either Java Web Start or applets. I know peopleC > that I consider good in those that consider Java Web Start better  > than applets.   K Just had a (quick) read, and I too like the look of Java Web Start but feel G it seeks to address broader issues and is a sort of big-picture project D aimed at cross-version compatibility and the SUN/IBM defence againstJ Microsoft deprecating Java further. Down-loading JVMs and matching up JDKs= is all very interesting but I'm still just crackin' a nut :-)    Applets it is!  @ > Applets (or Java Web Start) will probably suit you better than
 > JavaScript.   J I suspect it's a combo of Applets *and* Javascript; but more in a reply to your other post.   Cheers Richard Maher  0 "Arne Vajhj" <arne@vajhoej.dk> wrote in message1 news:45453328$0$49201$14726298@news.sunsite.dk...  > Richard Maher wrote:K > >> PS: Have you investigated Java Web Start as an alternative to applet ?  > > $ > > Yet another tool/terminology :-( > > ' > > Never heard of it Arne, is it good?  > B > I am not good at either Java Web Start or applets. I know peopleC > that I consider good in those that consider Java Web Start better  > than applets.  > F > > This is the requirement-spec, and I'll use the screw-driver or the hammer;  > > you (or HP) tell me: - > > A > > 1) I need a TCP/IP socket interface/dll/api/method/widget-box G > > 2) Has to be available on 95% of browsers with no additional client  software > > reqdI > > 3) Has to be secure (within reason i.e. not uncommon browser settings  but no > > signatures) G > > 4) Has to permit connection to a server other than the html hosting  > > web-server > > $ > > Here are my asumptions so far: - > > E > > . Sockets aren't available in straight Javascript which led me to  applets J > > . Applets can connect back to their hosting server at "normal" browser > > security settings 7 > > . The initialize and destroy methods look shit hot! H > > . I really like the seperate thread idea for the applet! We can pull amazing  > > functionaly from this! > > 9 > > So I'm still at Applets at the moment; please advise.  > @ > Applets (or Java Web Start) will probably suit you better than
 > JavaScript.  >  > The problem is the 95%.  > / > Many newer Windows systems ship without Java.  > 1 > If you need your code to run on default Java on 6 > older Windows system you need to build your code for5 > Java 1.1, becaus ethat is what the MS JVM supports.  > 5 > One workaround is to embed the applet in a tag that : > asks users without Java support if they want to download > and install the newest Java. >  > Arne   ------------------------------  % Date: Mon, 30 Oct 2006 19:09:57 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 3 Subject: Re: Java Applet pulled from VMS Web Server 1 Message-ID: <ei4mar$7bq$1@news-02.connect.com.au>    Hi Arne,  D > I think this is what many consider limited. Probably not a problem9 > for you because you want to do thing serverside anyway.   J Apart from Request-initiation, Remote-task coordination, and Presentation, what else is there?     > AJAX uses JavaScript not Java. > C > JavaScript and Java does not have much in common except 4 letters  > in the name ...   I Are you saying that the Interfaces to the Socket Methods contained in the 0 Java Applet JAR file are not "surfaced" by AJAX?  H (I can't believe I just wrote that shit! It scares me even more to think) that the terminology could be correct :-)   * Look Arne, this is what I see happening: -  G 1) HTML page has an Applet tag so that, when the page is displayed, the $ Initialization method gets activated% 2) In the initialization, I choose to K 2a) Created a socket and connect back to the VMS server that is hosting the A applet (but different than the UNIX server hosting the html page) H 2b) Ask for VMS username/password and send them to the Tier3 Application Server on that node ' 2c) Recv the authorization success/fail F 3) A seperate AJAX thread in my browser is monitoring the state of theH Employee Name field and for every non-deterministic change I will Send aA "Lookup mf_personnel.employees message" to the server and refresh * "blink-free" the drop down under the name.K 4) When the browser is closed, the destroy method for the applet closes the  socket   Alles klar?   J Are you really saying that Javascript is incapable of accessing any of the9 input/output streams that have been opened by the Applet?   1 > AJAX is completely different from Java applets.  > I > AJAX uses traditional web frontend technologies: JavaScript, HTML, CSS.  > & > Java is a real programming language. > F > There are security rules for JavaScript also, but the rules are much > less well defined.  J Arne, you keep painting an either/or picture. Is it really that clear-cut?5 Can Javascript really not call/evoke/incantation Java  applets/methods/procedures?    Cheers Richard Maher  0 "Arne Vajhj" <arne@vajhoej.dk> wrote in message1 news:4545300c$0$49203$14726298@news.sunsite.dk...  > Richard Maher wrote:K > >> Applets sit at the client end (browser) and can't usually do very much " > >> because of security concerns. > > J > > It is my understanding that it is a common and secure practice to have yourI > > browser settings organized to let unsigned applets run in the sandbox  and F > > connect back to the server that they were downloaded from. Are you saying > > this is not the case?  > D > I think this is what many consider limited. Probably not a problem9 > for you because you want to do thing serverside anyway.  > K > > As far as Applets not being able to do "much" I've had some very useful K > > exchanges with people doing extremely interesting things with Java! For C > > example, using my Mickey Mouse employee surname lookup from the  mf_personnelK > > database people have described to me how they are using AJAX to monitor  the I > > input field and for every keystroke send the updated contents back to  the K > > server for a new list of matching employee surnames and display it as a H > > drop-down as the user is typing in. Similar to what Google does with search$ > > strings if you know what I mean. >  > Hmm. >   > AJAX uses JavaScript not Java. > C > JavaScript and Java does not have much in common except 4 letters  > in the name ...  > L > > Now, because no one seems to speak english in the Unix, Java, New world, I E > > have additional questions such as: - Is the "Asynch" part in AJAX 	 different J > > to the other sperate "threaded" examples people have shown me? Are anyB > > reduce security settings required for AJAX as opposed to other Javascript? > > > Is it available everywhere normal Javascript is available? > J > Asynch in in AJAX just means that the browser can refresh part of a page, > instead of the entire page ("blink free"). > 1 > AJAX is completely different from Java applets.  > I > AJAX uses traditional web frontend technologies: JavaScript, HTML, CSS.  > & > Java is a real programming language. > F > There are security rules for JavaScript also, but the rules are much > less well defined. >  > Arne   ------------------------------  % Date: Mon, 30 Oct 2006 06:56:28 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> 3 Subject: Re: Java Applet pulled from VMS Web Server < Message-ID: <4545e7f9$0$14837$9a6e19ea@news.newshosting.com>  1 "Arne Vajhj" <arne@vajhoej.dk> wrote in message  1 news:45454420$0$49200$14726298@news.sunsite.dk...  > Neil Rieck wrote: M >> At the end of the day, any flavour of Java solution (application, applet,   >> or L >> servlet) all need to run on a JVM (Java Virtual Machine) and you'll never< >> convince me that this is somehow better than native apps. >  [...snip...] > G >> 6) if you want to experience a real Java reality check, read Roland  M >> Hughes book titled "The Minimum You Need To Know About Java about Java on  : >> OpenVMS". You can read a small review of the book here:X >> http://www3.sympatico.ca/n.rieck/links/cool_openvms.html#TMYNTK-About-Java-on-OpenVMSI >> My employer purchased two copies of the book and we used the author's  I >> warnings to put the brakes on a couple of projects inside our company.  > - > Out of curiosity: what were those warnings.  >  > Arne  J Lots of comparisons between C/C++ and JAVA in this book proved to me that L JAVA should only be used where cross-platform portability is desired. Three  points that come to mind are:   J 1) object destructors in JAVA are called finalizers; why change the name? J because finalizers can't release an object or the memory allocated to it. - They can only mark it for garbage collection.   M 2) you can't force garbage collection. It happens automatically when the JVM  K decides the system has been quiet for a while. This means a system that is  G really humping along for a while (like most OpenVMS systems) may never  M "garbage collect" but you may still get an "out-of-memory" error when you do  B something like concat a string (which are all objects by the way).  K p.s. one of the main reasons people like OpenVMS is that these systems can  G hump for a long time (and even appear to be locked up) and then regain  L control of the situation without requiring a reboot. JVM garbage collection - restriction would make us loose that feature.   M 3) The section on comparing threads to processes makes me think that threads   should be used sparingly.   J For more information you should read the book. Unlike other authors, this M one uses "The Introduction" and "Chapter 1" to convince you to not use Java.  I It does tell you how to do RTL + SYS calls in OpenVMS. It does not cover  7 APPLETS and SERVLETS which may come in a later edition.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Mon, 30 Oct 2006 20:17:20 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 3 Subject: Re: Java Applet pulled from VMS Web Server 1 Message-ID: <ei4q93$crs$1@news-02.connect.com.au>    Hi Neil,  K > At the end of the day, any flavour of Java solution (application, applet,  orK > servlet) all need to run on a JVM (Java Virtual Machine) and you'll never ; > convince me that this is somehow better than native apps.   J I agree 100% I recommend native apps to everyone and don't want them to beG restricted by the browser sandbox and its idiosyncrasies. But I am also K pragmatic, lacking in conviction on the issue, and a purveyor of VMS server 4 software that doesn't care less what talks to it :-)  H I say "Use the Internet more and the browser less!". But  the browser isJ everywhere. You don't have to load/maintain/release *any* client software!G Sounds good to me - And, more to the point, good to the people with the  purse-strings!  K > At the end of the day, any flavour of Java solution (application, applet,  or
 > servlet) : 3 > 1) A Java servlet runs on a JVM at the web server   L I don't want to sound like Foghorn Leghorn, but my beak's a flappin' and youJ don't seem to be a listenin'! I'm not saying you're "as sharp as a bowlin'L ball" but let me spell it out for everyone: - 1) I couldn't give a toss whatG JVM or JDK is on VMS or what contortions one may have to perform to get F anything to work and 2) I will go to my grave before deploying a "Java7 servlet" on a VMS system as I just don't see the point!   L What's more (and if you'd've read my ITRC posts you'd already know) I'd likeI to use FTP (as opposed to HTTP) as the protocol to upload the .JAR applet I file from VMS, 'cos that way I wouldn't have to run with the baggage of a I Web-Server on VMS at all! I also floated the possibility of compiling and H archiving your applets on a Unix or Windows system and then just copyingJ your .JAR files to your VMS server to be upload to the PC browser clients.; Net result: - *NO* JVM and *NO* JDK on your VMS box  - Just + Application-Server grunt and reliability!!!   K Might not be for everyone, but if it proves to be technically feasible then  it sure works for me.   K But let's not scare everyone off - you can still use INETd or home-grown if  you prefer.   G > My employer purchased two copies of the book and we used the author's H > warnings to put the brakes on a couple of projects inside our company.  @ Like Arne, I'd like to know what those warnings were too please.   Cheers Richard Maher  L PS. Below is an example from Brian Reiter over at the ITRC thread. I hope itG helps somebody 'cos I simply can't do this. I have brought you to water I (Yeah, real quality H2O Himilayan shit Steve :-) but I simply cannot make F you drink :-( All you (someone) have to do is join up the bloody dots!  L If you find a stumbling block then I'll chase it down for you; otherwise get  off your arses and do something!  H Unfortunately, if anyone from HP works it out, then they'll just send upI some European proposal to put ten layers of concealing crud on it to keep . their mates employed for the next 10 years :-(  H Now, I have taught you how to fish. The next Wanker that say he's hungry get's it :-)  ? public class CobWebDisplay extends Applet implements Runnable {   /   private volatile Thread displayThread = null;    private SubsState subs[] ;   private SubsState subs[] ;     private String host ;   +     StatusStore store = new StatusStore() ;      tcpthread reader ;     Image cobweb_logo ;    Image serco_logo ;  
   int portid;      public void init() {      4      String portstring = getParameter("portnumber");      if ( portstring == null) 	         { F           System.out.println("Missing parameter -portstring- in APPLET tag");	         } 	      else 	         { 2           portid = Integer.parseInt (portstring) ;	         }   -      System.out.println ("Port "+portstring); 0      serco_logo = getPicImage("serco_logo.gif");-      cobweb_logo = getPicImage("cobweb.gif");   #      URL urlServer = getCodeBase();   !      host = urlServer.getHost() ;   %      store.updatestore ( "COBS 0 ") ;   )      store.updatestore ( "LCC 2 -1 -1") ; )      store.updatestore ( "LCC 3 -1 -1") ; )      store.updatestore ( "LCC 4 -1 -1") ; )      store.updatestore ( "LCC 5 -1 -1") ; '      store.updatestore ( "HAL -1 -1") ;   )      store.updatestore ( "IPL 1 -1 -1") ; )      store.updatestore ( "IPL 2 -1 -1") ;            repaint();     }       #   public void update (Graphics g) {         paint( g) ;   }   !   public void paint(Graphics g) {         store.paint( g) ;      )      g.drawImage(serco_logo, 0 ,0, this); +      g.drawImage(cobweb_logo, 0 ,50, this);      }     2  private Image getPicImage(String imageFileName) {       Image imgWork = null; +   MediaTracker mt = new MediaTracker(this);      try     {8     imgWork = getImage(getDocumentBase(),imageFileName);    }   catch(Exception e1)     {     System.out.println(e1);     }   mt.addImage(imgWork, 0);   try    { 1    showStatus("Loading image " + imageFileName );     mt.checkID(0, true);     mt.waitForID(0);    }    catch(Exception e2)    {     System.out.println(e2);   }    return imgWork;   }       public void start() { "       if (displayThread == null) {    :          reader = new tcpthread ( store , host , portid) ;          reader.start() ;   5          displayThread = new Thread(this, "Display");           displayThread.start();            }      }        public void run() {          int i = 0;    /       Thread myThread = Thread.currentThread(); +         while (displayThread == myThread) {                   repaint();                    try { #                 Thread.sleep(1000); /             } catch (InterruptedException e){ } 	         }      }      }              class COBSConnection {           Socket Sckt;         String Host = null;          DataOutputStream Out;          BufferedReader In;         String host ;          int port ;  )         COBSConnection(String h, int p) {                host = h ;             port = p ;	         }            void connect () {                  try {         5                         Sckt = new Socket(host,port); K                         Out = new DataOutputStream(Sckt.getOutputStream()); 3                         In = new BufferedReader(new * InputStreamReader(Sckt.getInputStream()));1                         if (Sckt.isConnected()) {                                }                            else {7         System.out.println("Not Connected to " + Host);                              }                  } '                 catch (Exception e) { } 	         }            void close() {                 try { %                         Sckt.close();                  } '                 catch (Exception e) { } 	         }            void send(String S) {                  S = S+"\n";                  try { *                         Out.writeBytes(S);                 } '                 catch (Exception e) { } 	         }            String recv() {                  String S;                  try { 8                         while ((S=In.readLine())==null);!                         return S;                  } %                 catch (Exception e) { $                         return null;                 } 	         }  }       class tcpthread extends Thread {           StatusStore  store ;#         COBSConnection cobs = null;          boolean connected ;            String host ;          int port ;  8         tcpthread ( StatusStore s , String h , int p ) {                 store=s ;                  host = h ;                 port = p ;	         }                public void init () {     3         System.out.println("Host name is " + host);   8                 cobs = new COBSConnection( host,port ) ;                   cobs.connect();       	         }            public void run () {                       init() ;  "                 connected = true ;    0                 store.updatestore ( "COBS 1 ") ;                   while (true) {  %                     while (connected)                      { !                         String S;   )                         S = cobs.recv() ;   $                         if (S!=null)                         { 2                             store.updatestore(S) ;                           }                          else                         { +                         connected = false ;                          }                      }   $                     cobs.connect() ;  &                     connected = true ;                 }    //              cobs.close() ;  	         }   + //      public void stop() {cobs.close(); }    }     4 "Neil Rieck" <n.rieck@sympatico.ca> wrote in message5 news:45453db6$0$7476$9a6e19ea@news.newshosting.com...  > @ > "Richard Maher" <maher_rj@hotspamnotmail.com> wrote in message- > news:ei3ah1$5kj$1@news-02.connect.com.au...  > > Hi Neil, > >  > [...snip...] > > J > > It is my understanding that it is a common and secure practice to have > > yourI > > browser settings organized to let unsigned applets run in the sandbox  and F > > connect back to the server that they were downloaded from. Are you sayingG > > this is not the case? (Having said that my IE on Windows2000 server  seems  > > toJ > > show Allow "All" network connections or "None". Does the registry have to > > beI > > modified to get finer granularity or am I looking in the wrong place? 
 > > AnyoneF > > with other browsers care to share their Java-setting experiences?) > > K > > As far as Applets not being able to do "much" I've had some very useful K > > exchanges with people doing extremely interesting things with Java! For C > > example, using my Mickey Mouse employee surname lookup from the  > > mf_personnelK > > database people have described to me how they are using AJAX to monitor  > > the I > > input field and for every keystroke send the updated contents back to  the K > > server for a new list of matching employee surnames and display it as a H > > drop-down as the user is typing in. Similar to what Google does with
 > > search$ > > strings if you know what I mean. > > L > > Now, because no one seems to speak english in the Unix, Java, New world, I E > > have additional questions such as: - Is the "Asynch" part in AJAX 
 > > different J > > to the other sperate "threaded" examples people have shown me? Are anyB > > reduce security settings required for AJAX as opposed to other Javascript? > > > Is it available everywhere normal Javascript is available? > >  > J > I pretty much agree with all your points. A few years back I got "sucked in" K > by all the Java hype and ended up taking a few semesters of Java training  atG > a local college. Like off shoring, many companies want programmers to  check H > out Java solutions because they programs are mostly cross platform and cheap  > (or even free).  > K > At the end of the day, any flavour of Java solution (application, applet,  orK > servlet) all need to run on a JVM (Java Virtual Machine) and you'll never ; > convince me that this is somehow better than native apps.  > J > So hopefully these answers will clear up some of the mystery (but I'm no$ > expert; just a johnny-come-lately) > E > 0) A java application runs on a local JVM (I guess this is obvious)  > 3 > 1) A Java servlet runs on a JVM at the web server  > K > 2) A Java applet runs on a JVM at the client end (after being sent by the 	 > server)  > J > 3) There are some cool Java applets out there which can communicate with
 > servlets > L > 4) Like any C/C++ program, a properly written Java applet can use TCPIP toJ > open a connection to a remote database. Because Java has been used to do all L > kinds of evil things, I find the default environment for most JVMs is veryJ > restricted. (but users are allowed to remove most of these restrictions) > L > 5) Like LAMP, AJAX is a very cool collection of technologies (AsynchronousI > Javascript And XML) used together to pull off some desired task. I have K > never heard of AJAX using JAVA, just Javascript (the two technologies are L > completely different). The Asynch phrase relates to the ability to have anF > asynchronous communications channel working behind the scene in yourK > browser, watching what you do, and possibly offering choices to you while J > you type, and before you click submit. For a good AJAX demo just surf onK > over to http://maps.google.com Notice how the page never reloads when you I > submit data? I have found a lot of good AJAX articles are linked to the I > bottom of this page: http://en.wikipedia.org/wiki/AJAX as well as here:  > o http://www-128.ibm.com/developerworks/views/web/libraryview.jsp?search_by=Mastering+Ajax&ca=dgr-wikiaMasterAJAX  > L > 6) if you want to experience a real Java reality check, read Roland HughesD > book titled "The Minimum You Need To Know About Java about Java on	 OpenVMS". / > You can read a small review of the book here:  > U http://www3.sympatico.ca/n.rieck/links/cool_openvms.html#TMYNTK-About-Java-on-OpenVMS G > My employer purchased two copies of the book and we used the author's H > warnings to put the brakes on a couple of projects inside our company. > * > I hope you find this information useful. >  > Neil Rieck > Kitchener/Waterloo/Cambridge,  > Ontario, Canada.: > http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Mon, 30 Oct 2006 20:33:13 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 3 Subject: Re: Java Applet pulled from VMS Web Server 1 Message-ID: <ei4r6t$e65$1@news-02.connect.com.au>    Hi Neil,  L Yep, to me, everything the book has to say, screams out "Don't try this shit" on VMS!". But then, why would you?  6 Just to keep a couple more CV-Builders employed at HP?  J Let the web server pimp it's web page. Apart from that, grow up and do the
 job properly!    Cheers Richard Maher  4 "Neil Rieck" <n.rieck@sympatico.ca> wrote in message6 news:4545e7f9$0$14837$9a6e19ea@news.newshosting.com... > 2 > "Arne Vajhj" <arne@vajhoej.dk> wrote in message3 > news:45454420$0$49200$14726298@news.sunsite.dk...  > > Neil Rieck wrote: F > >> At the end of the day, any flavour of Java solution (application, applet,  > >> or H > >> servlet) all need to run on a JVM (Java Virtual Machine) and you'll never > > >> convince me that this is somehow better than native apps. > >  > [...snip...] > > H > >> 6) if you want to experience a real Java reality check, read RolandK > >> Hughes book titled "The Minimum You Need To Know About Java about Java  on< > >> OpenVMS". You can read a small review of the book here: > >>U http://www3.sympatico.ca/n.rieck/links/cool_openvms.html#TMYNTK-About-Java-on-OpenVMS J > >> My employer purchased two copies of the book and we used the author'sK > >> warnings to put the brakes on a couple of projects inside our company.  > > / > > Out of curiosity: what were those warnings.  > >  > > Arne > K > Lots of comparisons between C/C++ and JAVA in this book proved to me that G > JAVA should only be used where cross-platform portability is desired.  Three  > points that come to mind are:  > K > 1) object destructors in JAVA are called finalizers; why change the name? K > because finalizers can't release an object or the memory allocated to it. / > They can only mark it for garbage collection.  > J > 2) you can't force garbage collection. It happens automatically when the JVM L > decides the system has been quiet for a while. This means a system that isH > really humping along for a while (like most OpenVMS systems) may neverK > "garbage collect" but you may still get an "out-of-memory" error when you    doD > something like concat a string (which are all objects by the way). > L > p.s. one of the main reasons people like OpenVMS is that these systems canH > hump for a long time (and even appear to be locked up) and then regainB > control of the situation without requiring a reboot. JVM garbage
 collection/ > restriction would make us loose that feature.  > F > 3) The section on comparing threads to processes makes me think that threads  > should be used sparingly.  > K > For more information you should read the book. Unlike other authors, this H > one uses "The Introduction" and "Chapter 1" to convince you to not use Java. J > It does tell you how to do RTL + SYS calls in OpenVMS. It does not cover9 > APPLETS and SERVLETS which may come in a later edition.  >  > Neil Rieck > Kitchener/Waterloo/Cambridge,  > Ontario, Canada.: > http://www3.sympatico.ca/n.rieck/links/cool_openvms.html >  >  >  >    ------------------------------  % Date: Mon, 30 Oct 2006 07:40:46 -0500 6 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= <arne@vajhoej.dk>3 Subject: Re: Java Applet pulled from VMS Web Server ' Message-ID: <4545F2CE.50902@vajhoej.dk>    Neil Rieck wrote: 3 > "Arne Vajhj" <arne@vajhoej.dk> wrote in message  3 > news:45454420$0$49200$14726298@news.sunsite.dk...  >> Neil Rieck wrote:H >>> 6) if you want to experience a real Java reality check, read Roland N >>> Hughes book titled "The Minimum You Need To Know About Java about Java on ; >>> OpenVMS". You can read a small review of the book here: Y >>> http://www3.sympatico.ca/n.rieck/links/cool_openvms.html#TMYNTK-About-Java-on-OpenVMS J >>> My employer purchased two copies of the book and we used the author's J >>> warnings to put the brakes on a couple of projects inside our company.. >> Out of curiosity: what were those warnings.  L > 1) object destructors in JAVA are called finalizers; why change the name? L > because finalizers can't release an object or the memory allocated to it. / > They can only mark it for garbage collection.   E Because they are not the same. C# started by calling them destructors F in V1.x but changed the name to finalizers to avoid the confusion. And! besides nobody uses them in Java.   O > 2) you can't force garbage collection. It happens automatically when the JVM  M > decides the system has been quiet for a while. This means a system that is  I > really humping along for a while (like most OpenVMS systems) may never  O > "garbage collect" but you may still get an "out-of-memory" error when you do  D > something like concat a string (which are all objects by the way).   ????  @ You will only get out of memory if there are no memory available and no memory can be GC'ed.   9 It do try to GC before it runs out. It is like paging and = swapping. You dont't really know when it is happening, but if   memory is needed it will happen.  M > p.s. one of the main reasons people like OpenVMS is that these systems can  I > hump for a long time (and even appear to be locked up) and then regain  N > control of the situation without requiring a reboot. JVM garbage collection / > restriction would make us loose that feature.   ) I think that is a misunderstanding of GC.   : The only drawback of GC is that it has very poor real time characteristics.  O > 3) The section on comparing threads to processes makes me think that threads   > should be used sparingly.   = It is wise to keep it below 100 or so (depending on the box).    Arne   ------------------------------  % Date: Mon, 30 Oct 2006 07:45:04 -0500 6 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= <arne@vajhoej.dk>3 Subject: Re: Java Applet pulled from VMS Web Server 7 Message-ID: <4545f3cb$0$49209$14726298@news.sunsite.dk>    Richard Maher wrote:! >> AJAX uses JavaScript not Java.  >>D >> JavaScript and Java does not have much in common except 4 letters >> in the name ... > K > Are you saying that the Interfaces to the Socket Methods contained in the 2 > Java Applet JAR file are not "surfaced" by AJAX?   Yes.  < AJAX uses an XmlHttpRequest object to send HTTP requests not
 real sockets.   I > 1) HTML page has an Applet tag so that, when the page is displayed, the & > Initialization method gets activated' > 2) In the initialization, I choose to M > 2a) Created a socket and connect back to the VMS server that is hosting the C > applet (but different than the UNIX server hosting the html page) J > 2b) Ask for VMS username/password and send them to the Tier3 Application > Server on that node ) > 2c) Recv the authorization success/fail H > 3) A seperate AJAX thread in my browser is monitoring the state of theJ > Employee Name field and for every non-deterministic change I will Send aC > "Lookup mf_personnel.employees message" to the server and refresh , > "blink-free" the drop down under the name.M > 4) When the browser is closed, the destroy method for the applet closes the  > socket > 
 > Alles klar?    True except for #3.   L > Are you really saying that Javascript is incapable of accessing any of the; > input/output streams that have been opened by the Applet?   : Not quite. You can access the Java applet from JavaScript.  0 But I would probably do it the other way around.  ) Java applet -> JavaScript -> HTML form !!   2 >> AJAX is completely different from Java applets. >>J >> AJAX uses traditional web frontend technologies: JavaScript, HTML, CSS. >>' >> Java is a real programming language.  >>G >> There are security rules for JavaScript also, but the rules are much  >> less well defined.  > L > Arne, you keep painting an either/or picture. Is it really that clear-cut?7 > Can Javascript really not call/evoke/incantation Java  > applets/methods/procedures?   ( Java and JavaScript can call each other.  : Through a very old NetScape interface. STill working fine.   Arne   ------------------------------  + Date: Mon, 30 Oct 2006 18:12:26 +0000 (UTC) - From: klewis@MAZDA.MITRE.ORG (Keith A. Lewis) 3 Subject: Re: Java Applet pulled from VMS Web Server . Message-ID: <ei5faa$1qu$1@newslocal.mitre.org>   "Richard Maher" <maher_rj@hotspamnotmail.com> writes in article <ehufap$s97$1@news-02.connect.com.au> dated Sat, 28 Oct 2006 10:33:39 +0800:
 >Hi Keith, > 9 >> Last I heard, applets were restricted to communicating * >> with the server they  were loaded from. > I >Apparently not, and I don't even think you need a http server process on G >your secondary Applet-Server box. See the following for *much* more: - M >http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1071300   I The reason for this applet security is to prevent somebody from making an J applet which does other things like send spam.  Some people have mentionedK signed jarfiles; I have never done that so I'm not in a position to advise.   ; >> OK, now you have to decide how to implement your server.  > M >Nope. I decided how to implement the server(s) 15 years ago and did it. It's L >how to get at 'em from the ubiquitous web-browser is what I have to decide,& >and I think I'm gettin' pretty close.  L Sounds like all you have left to do is set up a bare-bones VMS web server to@ supply your applet's codebase.  And write the applet, of course.  D Somebody else mentioned servlets.  Yeah you should be aware of theirD existence but IMHO they don't go well with applets.  Servlets are anJ excellent tool for running HTML "sessions" over HTTP 1.0, a protocol which2 closes its session after each page load (I think).  L Your existing server probably maintains a 1:1 session-to-connection mapping,D so you don't have any issue there which would be helped by servlets.  I >PS. I'll send you and evaluation copy of Tier3/hotTIP V3.1 if that's ok?   L No, thanks.  I'm not really doing anything with web servers or applets rightJ now.  The helloworld thing was something I dusted off from 1997 because I  thought it might help you.     0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Mon, 30 Oct 2006 09:17:21 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> I Subject: Java vs. Javascript (was Java Applet pulled from VMS Web Server) < Message-ID: <454608fd$0$14814$9a6e19ea@news.newshosting.com>  0 "Arne Vajhj" <arne@vajhoej.dk> wrote in message1 news:4545f3cb$0$49209$14726298@news.sunsite.dk...  [...snip...] > * > Java and JavaScript can call each other. > < > Through a very old NetScape interface. STill working fine. >  > Arne  H The following comments are probably a little off-topic but I thought I'dG pass them on anyway since some people lurking on this thread may not be  aware of some of the facts:   G Java is an object oriented compiled language originally produced by SUN I which purports to be a better C++ (no pointers, better strings, etc.) but F IMHO seems to be a language which will allow non-programmers to becomeF application developers. The complied code will only run on a JVM (JavaK Virtual Machine) and is almost always capable of running on any platform. I K guess it goes without saying that if a Java program has been extended to do L RMS calls on the host platform, then that program will not run on a platform! that does not have RMS installed.   J JavaScript is an object oriented scripting language originally produced by Netscape and called LiveScript. D http://wp.netscape.com/comprod/columns/techvision/innovators_be.htmlI At first blush, programming in a scripting language may not make any more J sense than programming in DCL. But we all have places where we would only L use DCL and there are many places where JavaScript is superior to Java. One G of those places is when using DOM (built-into all popular browsers) to   access objects on a web page.   H Today you will hear many large companies refer to Java by another name, / ECMAscript (which may not be entirely accurate) ' http://en.wikipedia.org/wiki/Ecmascript       
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  # Date: Mon, 30 Oct 2006 14:54:48 GMT & From: John Reagan <john.reagan@hp.com>Y Subject: Re: LSE Language Definitions for Fortran 90 or Fortran 95 Free-Format Source? So 0 Message-ID: <Ymo1h.1727$_9.296@news.cpqcorp.net>   Craig Dedo wrote:   G >     Recommendation to HP OpenVMS Engineering:  You may wish to update 2 > LSE to recognize free-format source for Fortran.  D Well, OpenVMS engineering doesn't have the DECset support.  That is I another group.  We do have the Fortran compilers however so any template  8 updates would be a joint effort.  I have forwarded your 7 question/suggestion to the appropriate product manager.    --   John Reagan 5 HP Pascal/{A|I}MACRO/COBOL for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------    Date: 30 Oct 2006 09:58:50 -0800 From: etmsreec@yahoo.co.uk$ Subject: Re: Mylex controller updateC Message-ID: <1162231130.368524.319950@h48g2000cwc.googlegroups.com>   C I've thought of doing that a few times, but recent experiences with + SWXCRs mean I don't put much faith in them.   A Still, I guess I should do a backup of my hobbyist 1000A sometime  soon...!   Steve    Paul Sture wrote: 7 > In article <45447282$0$20664$bf4948fe@news.tele2.nl>, ' >  "H Vlems" <hvlems@freenet.de> wrote:  > J > > Backup has always been a very close friend.... I decided to format theL > > drives as JBOD's and use them as separate devices. I might yet decide toO > > create a bindset but never liked it much. With RA80's it seemed a nice idea 2 > > but restoring bindset backups is a royal pain. > J > Or shadow two of them for important data, and use the other as a scratch	 > device?  > I > That way, if the scratch disk fails, you can simply point whatever uses G > it elsewhere, or "borrow" one of the shadow set members until you can  > fit a replacement. >  > --   > Paul Sture   ------------------------------  % Date: Mon, 30 Oct 2006 10:07:34 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> = Subject: Re: Observations on Samba for VMS (CIFS Kit from HP) J Message-ID: <paul.sture.nospam-0EBE14.10073430102006@mac.sture.homeip.net>  8 In article <9thak2trfmcvd2be88namr0gjerumpood3@4ax.com>,1  Clay M. Denton <denton@orison.dsserv.com> wrote:   O > I recently installed the CIFS kit on an AlphaServer to see if it could start  Q > to replace my Advanced Server (Pathworks) setup.  I ran into a couple of things = > that I wanted to see if anyone else had seen - or resolved.  > C > Server - AlphaServer DS20, OpenVMS 8.2, TCP/IP 5.5 - SAN w/HSG80s  > PC - Windows XP Professional@ > Network - 100Mb Full Duplex - hard coded on switch and systems > O > 1)  It's slow. I could only get 1.25MB/sec transfer rates between the PC and  
 > the server. M > From the same PC to the same Alpha with ftp I could get 7MB/sec.  From the  ^ > same PC to an AlphaServer cluster running advanced server I could get 11MB/sec (pretty much  > the max with overhead).   I I haven't tried to measure in MB/sec, but have noticed that with a large  G number of files in a directory, Finder (the OS X equivalent of Windows  F Explorer), can send Samba up to over 12,000 logical name translations  per second.   K > 2) File size limitations.  One of the walls I keep hitting with Advanced  M > Server is the 4GB file size limit.  Samba on Linux has supported very large N > files for a while now.  I was very surprised to find that the CIFS kit maxes > out at 2GB files!   F Could this be a compilation option? When you say "Samba on Linux", is 5 that a general thing, or a specific flavour of Linux?    --  
 Paul Sture   ------------------------------    Date: 30 Oct 2006 02:34:59 -0800 From: etmsreec@yahoo.co.uk= Subject: Re: Observations on Samba for VMS (CIFS Kit from HP) A Message-ID: <1162204499.044478.8160@h48g2000cwc.googlegroups.com>    Hi Clay,  C The kit that's out now is only an evaluation release, not the final E production version.  As such, there's probably going to be a shedload F of changes between what you've got now and what you'll get in the end.  ; Proposals (I'm told) for the final product are extended VMS = compatibility/support a la Advanced Server, some more AS-like F management programs and a conversion utility to take the configuration  from AS and import it into CIFS.  1 As usual, these are not firm commitments from HP. G Apparently, the production release will be based on Samba 3.x since the " 4.x release is just a bit too new.  G These are all my understandings though and don't reflect any commitment ( from HP since I don't work for them! :o)   Steve    Clay M. Denton wrote: \ > I recently installed the CIFS kit on an AlphaServer to see if it could start to replace my[ > Advanced Server (Pathworks) setup.  I ran into a couple of things that I wanted to see if % > anyone else had seen - or resolved.  > C > Server - AlphaServer DS20, OpenVMS 8.2, TCP/IP 5.5 - SAN w/HSG80s  > PC - Windows XP Professional@ > Network - 100Mb Full Duplex - hard coded on switch and systems > Z > 1)  It's slow. I could only get 1.25MB/sec transfer rates between the PC and the server.Z > From the same PC to the same Alpha with ftp I could get 7MB/sec.  From the same PC to an\ > AlphaServer cluster running advanced server I could get 11MB/sec (pretty much the max with > overhead). > \ > 2) File size limitations.  One of the walls I keep hitting with Advanced Server is the 4GBY > file size limit.  Samba on Linux has supported very large files for a while now.  I was B > very surprised to find that the CIFS kit maxes out at 2GB files!   ------------------------------  % Date: Mon, 30 Oct 2006 07:03:27 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> = Subject: Re: Observations on Samba for VMS (CIFS Kit from HP) < Message-ID: <4545e99c$0$14842$9a6e19ea@news.newshosting.com>  = "Clay M. Denton" <denton@orison.dsserv.com> wrote in message  2 news:9thak2trfmcvd2be88namr0gjerumpood3@4ax.com...H >I recently installed the CIFS kit on an AlphaServer to see if it could  >start to replace myK > Advanced Server (Pathworks) setup.  I ran into a couple of things that I   > wanted to see if% > anyone else had seen - or resolved.  > C > Server - AlphaServer DS20, OpenVMS 8.2, TCP/IP 5.5 - SAN w/HSG80s  > PC - Windows XP Professional@ > Network - 100Mb Full Duplex - hard coded on switch and systems > K > 1)  It's slow. I could only get 1.25MB/sec transfer rates between the PC   > and the server. M > From the same PC to the same Alpha with ftp I could get 7MB/sec.  From the   > same PC to an K > AlphaServer cluster running advanced server I could get 11MB/sec (pretty   > much the max with  > overhead). > K > 2) File size limitations.  One of the walls I keep hitting with Advanced   > Server is the 4GB H > file size limit.  Samba on Linux has supported very large files for a  > while now.  I was B > very surprised to find that the CIFS kit maxes out at 2GB files!  M 2 GB limit? That sounds suspiciously similar to the 31-bit / 32-bit / 64-bit  L discussions going on in this NG 6 months ago relating to limitations in ZIP.  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.9 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html     ------------------------------    Date: 30 Oct 2006 09:37:43 -0800' From: "sapienzaf" <sapienza@noesys.com> 1 Subject: OpenVMS v7.3-2 Bug in SORT/MERGE utility C Message-ID: <1162229863.457281.221390@e64g2000cwd.googlegroups.com>   F Back in v7.2-1 days there was a bug in SORT/MERGE involving the lengthF of logical names used on the command lines.  At the time my client was@ on support and there was a patch issued to resolved the problem.  F This past weekend the cluster was upgraded to v7.3-2 and lo and behold: the same bug is back.  But this time, no support contract.  G Here's the way to duplicate the problem.  In order for this to work the E expectation is that the data file being merged has at least one prior D version in the same directory.  The goal is to take the current fileD version, merge it into the prior version, and generate a file with a new version number.   C Don't ask why this needs to be done.  The fact is it's important at 
 this site.  
 Here goes:  0 $! just to be clear, Hypersort is NOT being used $ DEASSIGN SORTSHR $ ) $! somedatafile.dat has multiple versions  $ DEFINE MRG somedatafile.dat  $ 
 $! this works @ $ MERGE /NOCHECK /STABLE /SPEC=SYS$INPUT:   MRG:;-1,MRG:;   MRG:
   /KEY=(1) $ # $! this fails with access violation  $ DEFINE MRG1 somedatafile.datC $ MERGE /NOCHECK /STABLE /SPEC=SYS$INPUT:   MRG1:;-1,MRG1:;   MRG1:     /KEY=(1)  $   G So, same command, same target file, but in the failing case the logical C name has four characters whereas the succeeding case the logical is  only three characters.  / Is there a publicly available patch for v7.3-2?   D Can I take the patched SORTSHR.EXE from v7.2-1 and run it on v7.3-2?   ------------------------------  % Date: Mon, 30 Oct 2006 11:20:17 -0700 4 From: Norman Lastovica <norman.lastovica@oracle.com>5 Subject: Re: OpenVMS v7.3-2 Bug in SORT/MERGE utility * Message-ID: <45464261.7E3814D4@oracle.com>  > seems to work just fine on V7.3-2 with all patches up to date:  $ $ DEFINE /NOLOG MRG somedatafile.dat% $ DEFINE /NOLOG MRG1 somedatafile.dat A $ MERGE /NOCHECK /STABLE /SPEC=SYS$INPUT:   MRG:;-1,MRG:;   MRG:   /KEY=(1)D $ MERGE /NOCHECK /STABLE /SPEC=SYS$INPUT:   MRG1:;-1,MRG1:;   MRG1:  /KEY=(1) $      sapienzaf wrote: > H > Back in v7.2-1 days there was a bug in SORT/MERGE involving the lengthH > of logical names used on the command lines.  At the time my client wasB > on support and there was a patch issued to resolved the problem. > H > This past weekend the cluster was upgraded to v7.3-2 and lo and behold< > the same bug is back.  But this time, no support contract. > I > Here's the way to duplicate the problem.  In order for this to work the G > expectation is that the data file being merged has at least one prior F > version in the same directory.  The goal is to take the current fileF > version, merge it into the prior version, and generate a file with a > new version number.  > E > Don't ask why this needs to be done.  The fact is it's important at  > this site. >  > Here goes: > 2 > $! just to be clear, Hypersort is NOT being used > $ DEASSIGN SORTSHR > $ + > $! somedatafile.dat has multiple versions  > $ DEFINE MRG somedatafile.dat  > $  > $! this works B > $ MERGE /NOCHECK /STABLE /SPEC=SYS$INPUT:   MRG:;-1,MRG:;   MRG: >   /KEY=(1) > $ % > $! this fails with access violation   > $ DEFINE MRG1 somedatafile.datE > $ MERGE /NOCHECK /STABLE /SPEC=SYS$INPUT:   MRG1:;-1,MRG1:;   MRG1: 
 >    /KEY=(1)  > $  > I > So, same command, same target file, but in the failing case the logical E > name has four characters whereas the succeeding case the logical is  > only three characters. > 1 > Is there a publicly available patch for v7.3-2?  > F > Can I take the patched SORTSHR.EXE from v7.2-1 and run it on v7.3-2?   --  	 - - - - - 0  opinions expressed here are mine and mine alone.  and certainly are not intended in any way to 0  express or represent any opinions or commitment  of oracle corporation.   *  norman lastovica / oracle rdb engineering   ------------------------------  # Date: Mon, 30 Oct 2006 17:27:03 GMT " From:   VAXman-  @SendSpamHere.ORG Subject: PAKs for Itanium 0 Message-ID: <00A5DFA7.2B10B63E@SendSpamHere.ORG>  4 How do I now create PAKs which will load on Itanium?   --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------    Date: 30 Oct 2006 09:39:56 -0800  From: "Ian Miller" <ijm@uk2.net> Subject: Re: PAKs for Itanium C Message-ID: <1162229996.175505.139290@f16g2000cwb.googlegroups.com>   - you need a updated PAKGEN pak - talk to DSPP.    ------------------------------  % Date: Mon, 30 Oct 2006 12:46:52 -0500 2 From: "Stanley F. Quayle" <squayle@insight.rr.com> Subject: Re: PAKs for Itanium ; Message-ID: <4545F43C.21228.B6A9FF4@squayle.insight.rr.com>   8 On 30 Oct 2006 at 17:27, VAXman-@SendSpamHere.ORG wrote:6 > How do I now create PAKs which will load on Itanium?  D You can generate PAK's with /UNITS=100/ACTIVITY=CONSTANT=100.  That  seems to work.  F If you want to generate licenses on the Itanium system itself, you'll 2 need an Itanium PAK.  Mine have /OPT=(IA64,PPL)...  
 --Stan Quayle  Quayle Consulting Inc.  
 ----------8 Stanley F. Quayle, P.E. N8SQ  Toll free: 1-888-I-LUV-VAX3 8572 North Spring Ct., Pickerington, OH  43147  USA < stan-at-stanq-dot-com   http://www.stanq.com/charon-vax.html) "OpenVMS, when downtime is not an option"    ------------------------------  % Date: Mon, 30 Oct 2006 11:06:55 -0000 5 From: "Tom Garcia" <tgarcia-REMOVE-THIS@hivemind.org> : Subject: Re: simple improvements to disk write performance7 Message-ID: <4545dcd1$0$627$5a6aecb4@news.aaisp.net.uk>    Hein RMS van den Heuvel wrote:  G >> Although I  haven't requested it explicitly, I think this is default G >> when the file isn't opened with sharing ("shr=put" gives rec and stm  >> methods both dragged down >> to ~3MB/sec.).  > D > With SHR=PUT you are likely to be doing 1 IO per record put (and a > bunch of locks).F > IF you can fit mutlipel records in buffer (default 16kb these days),D > then you can improve on that by enabling Deferred Write (fop=dfw).C > here RMS postpones writing the buffer untill the application says   H Nice.. fop=dfw brings me up to 9MB/sec. Without this, as your statement D implies, record size is a clear determinant in how quickly the test K completes - for 512-byte records, speed goes down to about 100k/sec. Other   things I think I've learnt:   I (1) Record size of earlier file versions is honoured by fopen if none is   explicitly specified.   G (2) SET VOL/NOWRITETHROUGH (with its global enabler ACP_WRITEBACK) has  K nothing to do with allowing fop=dfw, but is about deferred writing of file  
 _headers_.  L (3) Setting RMS_SEQFILE_WBH=0 doesn't appear to slow any tests down, though K I'd have expected it to degrade performance for shr=nil when rop=wbh isn't  ( explicitly specified, per documentation:D     0 (default)    Do not enable writebehind feature. Preserve priorA                    behavior of using writebehind only if the user A                    requests it by setting RAB$V_WBH in RAB$L_ROP. I Maybe improvement for ctx=rec over ctx=stm _wasn't_ due to wbh after all?   , >> f = fopen("needless.bin", "w", "ctx=rec",7 >>     "rfm=stmlf", "deq=65535", "mbf=127", "mbc=127"); F >> /* nonsense has size CHUNK=32767 with all 'a's, terminating in '\n' >> */ B >> for (x=1; x<=200*MEG/CHUNK; x++) fwrite(nonsense, CHUNK, 1, f);
 >> fclose(f);  > D > mbc=127 : More then 3 or 4 buffers will only cause more memory useB > (pagefaulting) and offer no further advantage for linear writes. > E > mbf=127: That causes 'odd sized' IOs. As I replied earlier, you may G > have more luck with 124, or 112 (128-16). (hmmm, now i forget... does G > XFC stop cache at a given IO size like the 35 blocks for VIOC? No VMS  > handy right now to try).  H Noted (but swapped), thanks -- just for the record: mbc=96, mbf=8 gives M similar performance; mbf=4 adds about 15% to runtime. mbf=2 doubles runtime.   Cluster size is 16.    Thanks for your tips!    --  " Tom Garcia | tgarcia@hivemind.org    ------------------------------  % Date: Mon, 30 Oct 2006 10:04:12 -0500 ( From: Bill Todd <billtodd@metrocast.net>: Subject: Re: simple improvements to disk write performanceG Message-ID: <3aGdnbcP4OvwidvYnZ2dnUVZ_q2dnZ2d@metrocastcablevision.com>    Tom Garcia wrote:    ...   K > Maybe improvement for ctx=rec over ctx=stm _wasn't_ due to wbh after all?   I There's too much going on here to be really certain of anything, but the  H figure of interest is that a 10K rpm drive transferring 63.5KB every 1+ F rev (how much the "+" is depends on how large a percentage of a track D 63.5 KB occupies) dumps data into a contiguous file at no more than D around 9 MB/sec, no matter what other optimizations you may perform.  H Improvements beyond that could be due to write-behind, or opportunistic G transfers of multiple buffers when deferred write is specified (if RMS  G knows that it has to write *one* buffer to make room for more data but  H can see that it has several more dirty buffers that it could get rid of G in the same I/O operation, one might hope that it would take advantage  E of the opportunity to do so - and this could explain the performance  G decreases that you saw as you decreased the MBF value), or, of course,  % use of disk-level write-back caching.   E Just turning on the disk's write-back cache may not give you optimal  F performance, however, because RMS and the XQP should be bypassing the E disk's write cache (using the 'force unit access' modifier - I don't  I know what they'd do with ATA disks that don't provide that option, but I  I hope that they'd just always disable the write-back cache no matter what  B the user did) for some critical operations that must occur in the F correct order such that things won't be left in an unacceptable state H should, e.g., a power interruption occur.  Whereas when RMS and the XQP B have been told by the application that they're allowed to operate D asynchronously, they can keep the application from seeing most such  synchronization overheads.   - bill   ------------------------------  % Date: Mon, 30 Oct 2006 16:24:45 +0100 ( From: "Rudolf Wingert" <win@fom.fgan.de>: Subject: Re: simple improvements to disk write performance7 Message-ID: <000a01c6fc37$8791fbd0$994614ac@domina.fom>    Hello,   Bill Todd wrote:   >>> B because RMS and the XQP should be bypassing the disk's write cache <<<   G Sorry, but my experience is other. The MOUNT command disables the disks E onboard cache. If you enable this after the MOUNT command, it will be I unchanged till dismount or specific disable (user forced). As I did write K before, we use the RZ_DISK utility to enable the write cache. 9MB/s figures K very good for  a SCSI disk drive without write cache enable. I did see only K 3.xMB/s. Once I did disable the read cache and did see the same performance  for read and write (3.xMB/s).    Best regards R. Wingert    ------------------------------  % Date: Mon, 30 Oct 2006 10:59:50 -0500 ( From: Bill Todd <billtodd@metrocast.net>: Subject: Re: simple improvements to disk write performanceG Message-ID: <r_6dnTISebrrvNvYnZ2dnUVZ_q2dnZ2d@metrocastcablevision.com>    Rudolf Wingert wrote:  > Hello, >  > Bill Todd wrote: > D > because RMS and the XQP should be bypassing the disk's write cache > <<<  > I > Sorry, but my experience is other. The MOUNT command disables the disks G > onboard cache. If you enable this after the MOUNT command, it will be < > unchanged till dismount or specific disable (user forced).  H That doesn't contradict what I said:  I didn't say that RMS and the XQP D should *disable* the disk's write-back cache, just that they should @ *bypass* it for certain critical update sequences (using SCSI's * per-operation 'force unit access' option).  B ATA disks don't support that bit (I may have read that SATA disks G support something equivalent, but I'm not at all sure).  My suggestion  E that VMS should just always disable the write-back cache on such ATA  G disks may have been overly Draconian:  it should be sufficient just to  H wrap any update that would use 'force unit access' if it could inside a ! pair of ATA cache-flush commands.    - bill   ------------------------------  % Date: Mon, 30 Oct 2006 17:40:46 -0000 5 From: "Tom Garcia" <tgarcia-REMOVE-THIS@hivemind.org> : Subject: Re: simple improvements to disk write performance7 Message-ID: <45463920$0$626$5a6aecb4@news.aaisp.net.uk>    Bill Todd wrote:  E > That doesn't contradict what I said:  I didn't say that RMS and the B > XQP should *disable* the disk's write-back cache, just that theyH > should *bypass* it for certain critical update sequences (using SCSI's, > per-operation 'force unit access' option).  I I was guessing that enabling the write cache was Unsupported(tm), so all  I bets were off once I tweaked mode pages on a mounted drive. I attributed  K poor IDE performance to write cache being off, with no TCQ to nicely order  @ writes (and little interest in otherwise optimising the driver).  G I guess it would be nice to have a supported way of enabling the write  L cache, and choose whether to bypass for metadata updates. If performance is J more important than the risk of needing to restore from backup, FUA might L not be welcomed - e.g. shadow set to guard against drive failure, but write G cache enabled on some members so write performance remains reasonable?  A Though hardware failures are often concomitant with power loss...   C > ATA disks don't support that bit (I may have read that SATA disks H > support something equivalent, but I'm not at all sure).  My suggestion  4 Seagate's mumble on SATA's NCQ suggests FUA exists: O http://www.seagate.com/docs/pdf/whitepaper/D2c_tech_paper_intc-stx_sata_ncq.pdf   F Meanwhile, there seem to be quite a few discussions on the problem of L ensuring a classical IDE drive's cache is flushed before powerdown - random 	 examples:   F http://developer.apple.com/technotes/tn/tn1040.html implies one might I disable cache, send the write command (that actually flushes the cache),   then re-enable cache.   I Another gotcha... as late as 2004 there appears on the Linux kernel list  L http://www.ussg.iu.edu/hypermail/linux/kernel/0405.1/1018.html a discussion F on how some <120GB Maxtor drives(the SATA version of one such in this I workstation) falsely claim to support FLUSH CACHE EXT, when in fact they  $ don't accept 48-bit commands at all.  K Even assuming we do have a guaranteed flush, in the non-shutdown case this  K is still an unnecessary write of *everything* still cached. And so I learn    more about the evils of IDE :-).   --  ! Tom Garcia | tgarcia@hivemind.org    ------------------------------  % Date: Mon, 30 Oct 2006 13:28:42 -0500 ( From: Bill Todd <billtodd@metrocast.net>: Subject: Re: simple improvements to disk write performanceG Message-ID: <MdudndqbxOrG2dvYnZ2dnUVZ_r2dnZ2d@metrocastcablevision.com>    Tom Garcia wrote:  > Bill Todd wrote: > F >> That doesn't contradict what I said:  I didn't say that RMS and theC >> XQP should *disable* the disk's write-back cache, just that they I >> should *bypass* it for certain critical update sequences (using SCSI's - >> per-operation 'force unit access' option).  > K > I was guessing that enabling the write cache was Unsupported(tm), so all  = > bets were off once I tweaked mode pages on a mounted drive.   I My point was that it doesn't *have* to be unsupported if RMS and the XQP  E do the right thing when they find it enabled (and while there's some  H reasonable argument that it should *never* have to be enabled on a SCSI I drive, since higher-level software should be doing any write-caching and  G the TCQ mechanisms should provide optimal write performance otherwise,  E on ATA drives there's no way to obtain decent large-contiguous-write  I performance without enabling it, save for those few that support command  	 queuing).   E 'Force unit access' is what WinNT/2K/XP use for critical writes when  I they encounter enabled write-back disk caches (I don't know whether they  C bracket them with cache-flush commands for ATA disks that lack FUA  I capability, but they should), and it *may* be what RMS and the XQP do (I   just don't know if it is).     I attributedM > poor IDE performance to write cache being off, with no TCQ to nicely order  B > writes (and little interest in otherwise optimising the driver).  F It's not so much the ability to *reorder* writes as the fact that ATA G supports only smallish single writes (nominally up to 128 KB, but I've  H heard that some drives have quirks limiting them to as little as 32 KB) H and without any way to submits another write while the first portion of C a large transfer is being digested you have to miss an entire disk  = rotation between each one.  Small-write-reordering is a nice  G optimization in things like intense OLTP environments, but of far less  : general use than letting large writes execute efficiently.   > I > I guess it would be nice to have a supported way of enabling the write  N > cache, and choose whether to bypass for metadata updates. If performance is L > more important than the risk of needing to restore from backup, FUA might N > not be welcomed - e.g. shadow set to guard against drive failure, but write I > cache enabled on some members so write performance remains reasonable?  C > Though hardware failures are often concomitant with power loss...   I If FUA isn't welcomed, tough:  it's what's needed to ensure the internal  F integrity of the file system (you know, the thing that the OS depends F upon to function correctly as well as the application that, in and of  itself, may not give a damn).   = RMS and the XQP offer more than adequate facilities to allow  H applications *almost* completely asynchronous disk write behavior, with F forced synchronization only where it's important.  People who want to F run faster and looser than that can (and should) use raw disk access, 4 where whatever they do won't screw up anything else.   > D >> ATA disks don't support that bit (I may have read that SATA disksI >> support something equivalent, but I'm not at all sure).  My suggestion  > 6 > Seagate's mumble on SATA's NCQ suggests FUA exists: Q > http://www.seagate.com/docs/pdf/whitepaper/D2c_tech_paper_intc-stx_sata_ncq.pdf   G Yes - I just found an SATA 2.5 spec that indicates that FUA is part of  C the native command queuing command set (though not of the unqueued   command set, it appears).    > H > Meanwhile, there seem to be quite a few discussions on the problem of N > ensuring a classical IDE drive's cache is flushed before powerdown - random  > examples:  > H > http://developer.apple.com/technotes/tn/tn1040.html implies one might K > disable cache, send the write command (that actually flushes the cache),   > then re-enable cache.   H Interesting, but by now a decade out of date.  I *hope* that ATA drives I have improved somewhat in their consistency of write-back cache handling   over that period.    > K > Another gotcha... as late as 2004 there appears on the Linux kernel list  N > http://www.ussg.iu.edu/hypermail/linux/kernel/0405.1/1018.html a discussion H > on how some <120GB Maxtor drives(the SATA version of one such in this K > workstation) falsely claim to support FLUSH CACHE EXT, when in fact they  & > don't accept 48-bit commands at all.  G I've seen such discussions (the Linux driver code, in fact, is where I  E stumbled upon the 'as little as 32 KB' write-request size limitation  A that I mentioned above).  Again, I'm hoping that such bugs are a  B vanishing breed, but tend toward vendors like Seagate and Hitachi G because I have more confidence in their implementations than I have in  D vendors like Maxtor (RIP - though Seagate is still selling products I under that name, I hope that if they update them they'll fix them in the  	 process).    > M > Even assuming we do have a guaranteed flush, in the non-shutdown case this  M > is still an unnecessary write of *everything* still cached. And so I learn  " > more about the evils of IDE :-).  I Actually, in real-world use this probably isn't all that much worse (and  H certainly better in most cases than just not using the write-back cache I at all).  In any event, old-style ATA is a rapidly-dying breed, and SATA  D looks a lot better in multiple respects (while maintaining the same G attractive price points, even for those middle-of-the-road 'near-line'  C enterprise drives from Seagate and WD) - which is why I think that  B SCSI/FC may be on the way out, though will certainly be around in F significant numbers through at least the end of this decade (I really G like to see a hard-core comparison of Raptor performance with high-end  @ drives:  if it lives up to its specs, the latter will be really G hard-pressed to compete, since their performance is just not that much  , better compared with their far higher cost).   - bill   ------------------------------  % Date: Mon, 30 Oct 2006 13:38:04 -0500 ( From: Bill Todd <billtodd@metrocast.net>: Subject: Re: simple improvements to disk write performanceG Message-ID: <ibqdnT4iPYYQ29vYnZ2dnUVZ_sSdnZ2d@metrocastcablevision.com>    Tom Garcia wrote:    ...   K > Another gotcha... as late as 2004 there appears on the Linux kernel list  N > http://www.ussg.iu.edu/hypermail/linux/kernel/0405.1/1018.html a discussion H > on how some <120GB Maxtor drives(the SATA version of one such in this K > workstation) falsely claim to support FLUSH CACHE EXT, when in fact they  & > don't accept 48-bit commands at all.  E In fairness to Maxtor, this appears to be a documentation bug rather  G than a firmware bug - FLUSH CACHE EXT being a 48-bit command, when the  E disk doesn't support them (and thus returns an error after receiving  G it).  It does, however, support the normal 24-bit flush-cache command,  H and that's what Linux should have sent it (if you look at the preceding ? post in the thread, it appears that Linux may have performed a    multi-Boolean test incorrectly).   - bill   ------------------------------  # Date: Mon, 30 Oct 2006 15:02:04 GMT & From: John Reagan <john.reagan@hp.com>  Subject: Re: Spit Brook location0 Message-ID: <Mto1h.1728$_9.673@news.cpqcorp.net>   Alan Frisbie wrote:   A > When I was there in May 2005, the barcode still had two one-bit B > errors (a one bit in one character was swapped with the zero bit* > from another).   Have they fixed it yet? > > > In particular, on the bottom row, the first character shouldB > be an upper case "W" (hex 57).   Instead, it is hex 17 (Escape).< > Likewise, the seventh character is supposed to be an upper3 > case "D" (hex 44).   Instead, it is hex 46 ("F").   @ I believe they are still broken but I don't that entrance often.     --   John Reagan 5 HP Pascal/{A|I}MACRO/COBOL for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------    Date: 30 Oct 2006 07:10:58 -0800 From: etmsreec@yahoo.co.uk  Subject: Re: Spit Brook locationC Message-ID: <1162221058.244318.262910@i42g2000cwa.googlegroups.com>   6 So, in my ignorance, what should the barcode wall say?   Steve      Steve Lionel wrote:  > DaveG wrote: > I > > How long has ZK been the home of VMS engineering?  And thanks for the # > > links to the views of the area.  > H > August 1980 is when the first of VMS engineering moved in to the firstC > (lowest) floor of ZK1. For two years before that it had been in a H > converted shopping center in Tewksbury, MA.  I know that ZK2 opened in? > 1982, and languages moved there.  I think ZK3 opened in 1986.  > C > As for the helipad, it's the square in the middle of the image at  > http://maps.google.com/maps?f=q&hl=en&q=110+spit+brook+rd,+03062&ie=UTF8&om=1&z=18&ll=42.712712,-71.457752&spn=0.002523,0.005171&t=k > I > It does not seem to be specially marked, so perhaps it does not qualify E > anymore.  I have not seen any helicopters land there in many years. H > There is still the little waiting shed with instructions about waiting0 > for the pilot to signal that you can approach. > G > Regarding the barcode, I was not aware that there was an error in it. F > If there is, I doubt it has been fixed. I'll have to check that out. >  > Steve    ------------------------------    Date: 30 Oct 2006 07:17:09 -0800 From: etmsreec@yahoo.co.uk  Subject: Re: Spit Brook locationC Message-ID: <1162221429.024612.184320@f16g2000cwb.googlegroups.com>   6 So, in my ignorance, what should the barcode wall say?   Steve      Alphaman wrote:  > Alan Frisbie wrote: @ > > In particular, on the bottom row, the first character shouldD > > be an upper case "W" (hex 57).   Instead, it is hex 17 (Escape).> > > Likewise, the seventh character is supposed to be an upper5 > > case "D" (hex 44).   Instead, it is hex 46 ("F").  > G > Errr, Uhm, 0x17 is a 23 (decimal), which is NOT the escape character.   > Escape is 27 decimal, or 0x1B. > C > Does anyone have a photo of the wall they'd care to share with us , > geographically challenged trivia lovers??? > J > Okay, all this talk about the barcode wall and ZK3 has gotten me feelingD > nostalgic.  Anyone got a copy of ZK???  Oops, just found it on the > hobbyist site... > / > http://www.openvmshobbyist.com/downloads.html  > E > Cool!  Something to do this afternoon while my son plays the PS2...  >  > Aaron  > --F > "The deepest sin against the human mind is to believe things withoutJ > evidence. Science is simply common sense at its best -- that is, rigidlyG > accurate in observation, and merciless to fallacy in logic."  (Thomas  > Henry Huxley)    ------------------------------  + Date: Mon, 30 Oct 2006 17:06:11 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney)   Subject: Re: Spit Brook location( Message-ID: <ei5be3$ut9$2@pcls4.std.com>   etmsreec@yahoo.co.uk writes:  7 >So, in my ignorance, what should the barcode wall say?    It originally read:   
 DIGITALSOFTWA 
 REENGINEERING   5 but some VP found out (I think it was Compaq by then) / and he had it changed to some mealymouth slogan > which is, if I remember right: "Customers Win When We Deliver" or something equally inane.    ------------------------------  % Date: Mon, 30 Oct 2006 09:31:54 -0800 4 From: Alan Frisbie <Usenet02_REMOVE@Flying-Disk.com>  Subject: Re: Spit Brook location% Message-ID: <1162229514.405817@smirk>    etmsreec@yahoo.co.uk wrote: 8 > So, in my ignorance, what should the barcode wall say?   It is *supposed* to say:  
 Customers Win 
 WhenWeDeliver   9 (Note the clever lack of spaces in the bottom row to make  the two rows the same length.)  ? What it *actually* (as of June, 2005) says (in 7-bit ASCII) is:   
 Customers Win 
 .henWeFeliver   5 (The "." in the second row represents the unprintable  hex 17 character.)  4 I think this means that whatever is "Felivered" will have bugs in it.   :-)  5 I am told that an earlier incarnation of the bar code 7 said something like "Software Engineering", but I never 
 saw that one.   ; Could someone at ZKO please inspect the bar code and report 5 the colors of the first seven bars in the bottom row?    Alan   ------------------------------  % Date: Mon, 30 Oct 2006 09:34:41 -0800 4 From: Alan Frisbie <Usenet02_REMOVE@Flying-Disk.com>  Subject: Re: Spit Brook location% Message-ID: <1162229680.498705@smirk>    Michael Moroney wrote:  F > I don't know where the live.com "Bird's Eye" views came from.  I canG > tell you the different directions were shot at different times, since L > the view from one direction shows the ZK3 parking lots empty, from another > direction they're full.   9 No, it is because the cloaking device was only working in & certain directions at that time.   :-)   Alan   ------------------------------    Date: 30 Oct 2006 08:27:14 -0800 From: bob@instantwhip.com $ Subject: Re: Time change questions !C Message-ID: <1162225634.904128.293960@e64g2000cwd.googlegroups.com>    JF Mezei wrote: D > Oh no ! It is that time of the year again when there are plenty of  > questions about time changes ! > ) > I don't run that DECENET Phase 5 thing.   ! with DCL and Phase IV it is cake!   9 1.  Do a month and week check to see if it is time to set +     that nodes time forward or backward ...   ; 2. if it is, set the time of the current node + or - 1 hour   8 3.  feed the time to the other phase IV nodes via a task  7 simple!  And with TCPware you can do it over a phase IV > over IP connection and adjust your time for the different time) zones before sending it thru the task ...    ------------------------------  % Date: Sun, 29 Oct 2006 23:20:33 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> . Subject: Re: Updated VMS Information (big one), Message-ID: <45457D75.DE91AF49@teksavvy.com>  
 Sue wrote:D > First I would like to apologize for the size of this email message    F First, I would like to thank you for the size of your email message...  E Its contents (and your work in assembling it) are VERY appreciated by  the VMS community.   ------------------------------    Date: 30 Oct 2006 07:08:46 -0800  From: "Mike" <mlpoole@gmail.com># Subject: Re: VMS installation crash C Message-ID: <1162220925.995507.145130@m73g2000cwd.googlegroups.com>    Thanks!      Richard B. Gilbert wrote:  > JF Mezei wrote:  > H > > For what it is worth, my DS10L from the wonderful people at IslandCoK > > came with SRM version 7.2-1 dated June 9 2006 if I remember well. And I E > > have not have a problem with crashing VMS during install or while & > > running with both 8.2 and now 8.3. > > H > > But I did not install VMS from CD, I installed it via an MSCP served > > disk on another node.  > > L > > I'd have to crash the node again to get the PAL code version though, and= > > in doing so would lose this message before it is sent :-(  >  > JF,  > 0 > $ WRITE SYS$OUTPUT F$GETSYI("CONSOLE_VERSION")0 > $ WRITE SYS$OUTPUT F$GETSYI("PALCODE_VERSION") >   > It's all in knowing how. . . .   ------------------------------    Date: 30 Oct 2006 07:42:10 -0800  From: "Mike" <mlpoole@gmail.com># Subject: Re: VMS installation crash B Message-ID: <1162222930.045725.88610@i42g2000cwa.googlegroups.com>  G I had a defective halt switch that wasnt depressing all the way.  Whew!   That was an easy fix.  @ Shhh.. nobody tell David.  I think I voided my warranty heh heh.  C I cleared the NVRAM and I am setting the environment and boot stuff  now.   THANKS again!@         Mike wrote: 	 > Thanks!  >  >  > Richard B. Gilbert wrote:  > > JF Mezei wrote:  > > J > > > For what it is worth, my DS10L from the wonderful people at IslandCoM > > > came with SRM version 7.2-1 dated June 9 2006 if I remember well. And I G > > > have not have a problem with crashing VMS during install or while ( > > > running with both 8.2 and now 8.3. > > > J > > > But I did not install VMS from CD, I installed it via an MSCP served > > > disk on another node.  > > > N > > > I'd have to crash the node again to get the PAL code version though, and? > > > in doing so would lose this message before it is sent :-(  > >  > > JF,  > > 2 > > $ WRITE SYS$OUTPUT F$GETSYI("CONSOLE_VERSION")2 > > $ WRITE SYS$OUTPUT F$GETSYI("PALCODE_VERSION") > > " > > It's all in knowing how. . . .   ------------------------------    Date: 30 Oct 2006 04:59:03 -0800! From: "alok" <alok.net@gmail.com> Y Subject: why some times after inserting the tape in tape drive, comes out automatically ? C Message-ID: <1162213143.521992.292550@f16g2000cwb.googlegroups.com>   @ why some times after inserting the tape in tape drive, comes out automatically ?    ------------------------------    Date: 30 Oct 2006 14:18:22 +0100/ From: huber@NIRWANA-mppmu.mpg.de (Joseph Huber) Y Subject: Re: why some times after inserting the tape in tape drive, comes out automatical + Message-ID: <d2CnbI2o6jKb@vms.mppmu.mpg.de>   g In article <1162213143.521992.292550@f16g2000cwb.googlegroups.com>, "alok" <alok.net@gmail.com> writes: B > why some times after inserting the tape in tape drive, comes out > automatically ?   ' Could You be a bit more verbose please:    which system ?   which tape drive ?*   what was the command to mount the tape ?  L   If it is VMS: there should be a message from the mount command explaining "   the reason ("label mismatch" ?).  I   If it is not due to a mount command, then the drive itself is in error  E   (front-panel lights ?), or it does not like the tape/cartrige type. #   Are You loading a cleaning tape ?     --  @    Joseph Huber , Muenchen,Germany:  http://www.huber-joseph.de/   ------------------------------    Date: 30 Oct 2006 05:50:53 -0800$ From: "AEF" <spamsink2001@yahoo.com>Y Subject: Re: why some times after inserting the tape in tape drive, comes out automatical B Message-ID: <1162216253.355656.47060@e64g2000cwd.googlegroups.com>   Joseph Huber wrote: i > In article <1162213143.521992.292550@f16g2000cwb.googlegroups.com>, "alok" <alok.net@gmail.com> writes: D > > why some times after inserting the tape in tape drive, comes out > > automatically ?  > ) > Could You be a bit more verbose please:  >   which system ? >   which tape drive ?, >   what was the command to mount the tape ? > M >   If it is VMS: there should be a message from the mount command explaining $ >   the reason ("label mismatch" ?). > J >   If it is not due to a mount command, then the drive itself is in errorG >   (front-panel lights ?), or it does not like the tape/cartrige type. % >   Are You loading a cleaning tape ?  >  > --B >    Joseph Huber , Muenchen,Germany:  http://www.huber-joseph.de/  F I had exactly this problem with a TLZ09 4mm DAT drive. Sometimes I had; to push it in maybe 10 times for it to take. I replaced it.    AEF    ------------------------------    Date: 30 Oct 2006 09:47:16 -0800) From: "WWWebb" <William.W.Webb@gmail.com> Y Subject: Re: why some times after inserting the tape in tape drive, comes out automatical C Message-ID: <1162230436.647743.284200@i42g2000cwa.googlegroups.com>   
 AEF wrote: > Joseph Huber wrote: k > > In article <1162213143.521992.292550@f16g2000cwb.googlegroups.com>, "alok" <alok.net@gmail.com> writes: F > > > why some times after inserting the tape in tape drive, comes out > > > automatically ?  > > + > > Could You be a bit more verbose please:  > >   which system ? > >   which tape drive ?. > >   what was the command to mount the tape ? > > O > >   If it is VMS: there should be a message from the mount command explaining & > >   the reason ("label mismatch" ?). > > L > >   If it is not due to a mount command, then the drive itself is in errorI > >   (front-panel lights ?), or it does not like the tape/cartrige type. ' > >   Are You loading a cleaning tape ?  > >  > > --D > >    Joseph Huber , Muenchen,Germany:  http://www.huber-joseph.de/ > H > I had exactly this problem with a TLZ09 4mm DAT drive. Sometimes I had= > to push it in maybe 10 times for it to take. I replaced it.  >  > AEF    %MSG-I-HUH, ambiguous response    5 Did you replace the tape drive or the tape cartridge?    WWWebb   ------------------------------    Date: 30 Oct 2006 10:35:07 -0800$ From: "AEF" <spamsink2001@yahoo.com>Y Subject: Re: why some times after inserting the tape in tape drive, comes out automatical B Message-ID: <1162233307.664372.22560@m73g2000cwd.googlegroups.com>  
 WWWebb wrote:  > AEF wrote: > > Joseph Huber wrote: m > > > In article <1162213143.521992.292550@f16g2000cwb.googlegroups.com>, "alok" <alok.net@gmail.com> writes: H > > > > why some times after inserting the tape in tape drive, comes out > > > > automatically ?  > > > - > > > Could You be a bit more verbose please:  > > >   which system ? > > >   which tape drive ?0 > > >   what was the command to mount the tape ? > > > Q > > >   If it is VMS: there should be a message from the mount command explaining ( > > >   the reason ("label mismatch" ?). > > > N > > >   If it is not due to a mount command, then the drive itself is in errorK > > >   (front-panel lights ?), or it does not like the tape/cartrige type. ) > > >   Are You loading a cleaning tape ?  > > >  > > > --F > > >    Joseph Huber , Muenchen,Germany:  http://www.huber-joseph.de/ > > J > > I had exactly this problem with a TLZ09 4mm DAT drive. Sometimes I had? > > to push it in maybe 10 times for it to take. I replaced it.  > >  > > AEF  >   > %MSG-I-HUH, ambiguous response >  > 7 > Did you replace the tape drive or the tape cartridge?  >  > WWWebb   %ME-I-YURRITE, doh!  -MEM-I-ANSWER, the drive. F -MEM-I-SUPP, Actually, I don't think it ever went to 10 insertions. It just seemed that way!    AEF    ------------------------------   End of INFO-VAX 2006.597 ************************