1 INFO-VAX	Sun, 23 Apr 2000	Volume 2000 : Issue 226       Contents:! Re: Dropping DECnet..don't do it!  Re: Mozilla M15's out..." Re: VAX - Alpha upgrade assistance" Re: VAX - Alpha upgrade assistance" Re: VAX - Alpha upgrade assistance" Re: VAX - Alpha upgrade assistance Re: Verify of Backups   F ----------------------------------------------------------------------  # Date: Sun, 23 Apr 2000 00:51:16 GMT ( From: Terry Kennedy <terry@gate.tmk.com>* Subject: Re: Dropping DECnet..don't do it!' Message-ID: <FtG2DG.89D@spcuna.spc.edu>   2 JF Mezei <jfmezei.spamnot@vl.videotron.ca> writes:L > Sorry , I did not mean to say it was your customer's views. But most of myN > (former VMS) customers headed Palmer's call and stopped developping/building2 > on VMS platform and are building on other stuff.  H   Most of my customers changed from running VMS only to running a mix ofH VMS and "something else", where "something else" was usually a Unix box.H But they did this to use the best system for the particular application,. not out of any "VMS is dead/dying" perception.  H   But my customers probably differ from yours in that mine don't have toH justify their decision to others in the company - for example, one of my, 20-year customers is the CEO of his company.  J   The thing that caused *me*, personally, to stop upgrading my VMS systemsK was the inability to get an Alpha early on. We did a great deal of in-house H software development and wanted to get on the Alpha bandwagon. But whileI it seemed like every commercial developer had an Alpha - usually a Jensen H or a Flamingo, and later [I think] a Sandpiper, these were not orderable@ by "mere" customers, no matter how much you were willing to pay.  G   Once we wound up being a couple versions back, it was a lot easier to H look at the release notes and go "nope, nothing here for us". And as theG layered and 3rd-party products started requiring newer VMS releases, we  stopped upgrading them as well.   I   The only thing that changed this for me was the VMS V7.2-FT2 "everybody H can play" field test, and the fortuitous appearance of a load of surplusG AS200 4/233's for $250 or so. Putting those together with a generic CD- J ROM, more RAM, and some SCSI drives made me go "hey, this is pretty cool".  F   Since then, I've personally bought two DS10's, and my customers haveD bought DS10's and DS20's. And I attended my first DECUS Symposium inE ages. These boxes were bought with various add-ons as well (licences, H support uplifts, robotic media libraries, etc.). And all of this is justF because I happened to find a free field test and a cheap system at the
 same time.  G   I wonder how many [former] customers could be brought back by similar H low-cost things (like loaner DS10's to evaluate the latest VMS and hard- ware)?  E   To bring this back to the DECnet topic - if I'd made the mistake of F installing DECnet Phase V/DECnet-Plus on that first box, none of this H would probably have happened. I did install it on a test box for anotherG customer (with the problems described previously in this topic). When a E 3rd customer asked why I wanted to install Phase IV despite the "dire E warnings" in the release notes/installation guide, I said "you don't  G have it [DECnet] on a support contract anyway, so the support surcharge E doesn't affect you, and with Phase IV both I and other folks can help 7 you, while there's a lot less experience with Phase V".   - 	Terry Kennedy             http://www.tmk.com 5         terry@tmk.com             Jersey City, NJ USA    ------------------------------   Date: 22 Apr 2000 20:58:21 CST; From: wayne@tachyon.xxx.571648.killspam.0287 (Wayne Sewell) ! Subject: Re: Mozilla M15's out... . Message-ID: <PpqZTmfaz6GW@tachxxsoftxxconsult>  W In article <38FF5AE7.C66AD084@theblakes.com>, Colin Blake <colin@theblakes.com> writes:  > Peter LANGSTOEGER wrote: > ; >> GTK+ is some of kind of a [graphics] tool kit. Correct ?  > : > Yes. Mozilla doesn't use MOTIF as its GUI, it uses GTK+. > - >> It is a requirement for MOZILLA. Correct ?  >  > Yes. > L >> On VMS, GTK+ is not available standalone or VMS users are more uninformedK >> than on the other platforms and so the GTK+ gets installed automatically 4 >> in the background for and with MOZILLA. Correct ? > _ > GTK+ is not available for VMS. Not as a separate product, anyway. As part of the Mozilla port ^ > GTK+ has also been ported and work is underway to get the changes checked back into the GTK+] > source pool. Once that has happened I would expect GTK+ to be listed as a pre-requisite for ` > Mozilla/Netscape6 and for there to be a separate kit for it. Unless then it ships bundled into > the Mozilla kit. > 8 >> Why GTK+ doesn't get a separate (PCSI) product then ? >  > See above. > J >> >The other kits are in zip format. If you zip the OpenVMS PCSI file itsC >> >only 17 Mbytes instead of 30Mb for the self-extracting version.  >>H >> This raises again the question, why DCXEXE ? Use ZIP or [UN]ZIPSFX... > ] > Because some people don't like being given kits when some other tool that isn't part of the A > standard OpenVMS distribution is required to "unpack" that kit.  >   J That is not a factor with a self-extracting zip file, which can be done asJ easily with zip as with dcx.   The user does not need to have zip or unzipN installed on his system.   What, the run command is a tool not in the standardK openvms distribution?  If you run a program and get a file, what difference N does it make to the user what algorithm is used?   Other than the much smaller  size of the zip file, of course.     --  O =============================================================================== K Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachyon.xxx 8 http://www.tachyon.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-) O =============================================================================== N Butler:"Gentlemen!"  Curly(as he and other Stooges look around):"Who came in?"   ------------------------------  % Date: Sat, 22 Apr 2000 21:06:14 +0200 " From: "Hans Vlems" <hvlems@iae.nl>+ Subject: Re: VAX - Alpha upgrade assistance ( Message-ID: <8dsss7$nti$1@news.IAEhv.nl>  J Ahem, does that mean that I'm not the only one thinking that porting COBOL was & not that bad? Or PL/1 for that matter?  
 Hans Vlems+ Terry Marosites heeft geschreven in bericht D <1137A4A23A51D311B2D600105A1D5213019AEDAF@seantexch.unitedad.com>...: >Ahh the portability of C , not like that old Cobol stuff. >Just thinking aloud >Terry >  > -----Original Message-----8 >From: JF Mezei [mailto:jfmezei.spamnot@vl.videotron.ca]& >Sent: Friday, April 21, 2000 12:31 PM >To: Info-VAX@Mvb.Saic.Com, >Subject: Re: VAX - Alpha upgrade assistance >  >Larry Kilgallen wrote: I >> My recollection of stories at the time was that for those saddled with H >> legacy C code, converting from VAX C to DEC C was much more than half! >> the battle of moving to Alpha.  > H >Are there many instances where once you're converted your VAX-C code to >please H >the pedandic DEC-C compiler, that there would still be assumptions left that >poiters would be 32 bits ?  > I >I would thend to think that converting VAX-C to DEC-C would be closer to  85% % >of the task. Would that be correct ?  > L >However, having said this, if the use applications rely on middleware whichL >has not been ported to ALPHA (or whose funcionality was ported as a totallyA >separate product), then the task at hand might be more involved.    ------------------------------  % Date: Sat, 22 Apr 2000 16:12:12 -0400 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>+ Subject: Re: VAX - Alpha upgrade assistance / Message-ID: <39020799.69E9D339@vl.videotron.ca>    Hans Vlems wrote:  > L > Ahem, does that mean that I'm not the only one thinking that porting COBOL, > was not that bad? Or PL/1 for that matter?  N Cobol is tame and mature language. The problem with C is that it went from theK "glorified assembler that lets you do what you want" of K&R   to the "force E student to comply" of ANSI-C. The migration from the very lenient and L permissive K&R to the pedantic ANSI-C requires that you really go though all
 your code.  X Onc you get the hang of it, it becomes more or less automatic, but still time consuming.  8 COBOL hasn't undergone such a "puberty" for a long time.   ------------------------------  # Date: Sat, 22 Apr 2000 21:05:23 GMT 2 From: kilgallen@eisner.decus.org (Larry Kilgallen)+ Subject: Re: VAX - Alpha upgrade assistance ' Message-ID: <2000Apr22.170523.1@eisner>   5 In general, C and Macro-32 were the most troublesome. 7 Higher level languages ported easily, which is what one 
 would expect.    Larry Kilgallen   M In article <8dsss7$nti$1@news.IAEhv.nl>, "Hans Vlems" <hvlems@iae.nl> writes: L > Ahem, does that mean that I'm not the only one thinking that porting COBOL > was ( > not that bad? Or PL/1 for that matter? >  > Hans Vlems- > Terry Marosites heeft geschreven in bericht F > <1137A4A23A51D311B2D600105A1D5213019AEDAF@seantexch.unitedad.com>...; >>Ahh the portability of C , not like that old Cobol stuff.  >>Just thinking aloud  >>Terry  >> >> -----Original Message----- 9 >>From: JF Mezei [mailto:jfmezei.spamnot@vl.videotron.ca] ' >>Sent: Friday, April 21, 2000 12:31 PM  >>To: Info-VAX@Mvb.Saic.Com - >>Subject: Re: VAX - Alpha upgrade assistance  >> >>Larry Kilgallen wrote:J >>> My recollection of stories at the time was that for those saddled withI >>> legacy C code, converting from VAX C to DEC C was much more than half " >>> the battle of moving to Alpha. >>I >>Are there many instances where once you're converted your VAX-C code to  >>pleaseI >>the pedandic DEC-C compiler, that there would still be assumptions left  > that >>poiters would be 32 bits ? >>J >>I would thend to think that converting VAX-C to DEC-C would be closer to > 85% & >>of the task. Would that be correct ? >>M >>However, having said this, if the use applications rely on middleware which M >>has not been ported to ALPHA (or whose funcionality was ported as a totally B >>separate product), then the task at hand might be more involved. >  >    ------------------------------   Date: 22 Apr 2000 21:33:38 CST; From: wayne@tachyon.xxx.165349.killspam.0290 (Wayne Sewell) + Subject: Re: VAX - Alpha upgrade assistance . Message-ID: <+G5Hv$nHsDBg@tachxxsoftxxconsult>  y In article <rR0M4.27578$WF.1136293@bgtnsc04-news.ops.worldnet.att.net>, "John Nixon" <jorlnixon@worldnet.att.net> writes: M > My company is finally about to make the decision to tackle the VAX to Alpha M > VMS conversion.   I am looking for recommendations for  software conversion L > tools or consulting services, either within Compaq or external.  We are in > the Atlanta GA. area.  > K > The application is a fairly complex one that has evolved over the past 15 L > years and is mostly written in DECC.  There are some remnants of VAX MACRO > code.  > I > We have a programming staff, although an undermanned one.   We recently E > completed a huge project to convert from VAXC to DECC.  It was very E > difficult and took a long time.  Everyone is leery of another major L > coversion project at this time, so all we may really need is a little handD > holding and assurance that this project won't be as difficult as C
 > conversion.   O It probably won't be.  The compiler conversion tends to be 95% of the vax/alpha J conversion for programs written in C.  You've already done the hard part. H There will probably be more effort with the macro part, but you might be  surprised how easy this is, too.  J I maintain a lot of code written in dec c and targeted to both platforms. N There is very little platform-specific code.  Pretty much the same source code compiles on both vax and alpha.   F You might want to watch out for floating point data in files.  The twoK architectures use different formats.  Also, alignment of data is more of an K issue on alpha.  The alpha compiler will do it automatically, as it will on N vax, unless you tell it not to because of external data structures.  Unaligned+ data is more of a performance hit on alpha.      --  O =============================================================================== K Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachyon.xxx 8 http://www.tachyon.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-) O =============================================================================== N Butler:"Gentlemen!"  Curly(as he and other Stooges look around):"Who came in?"   ------------------------------   Date: 22 Apr 2000 19:14:02 CST; From: wayne@tachyon.xxx.573438.killspam.0283 (Wayne Sewell)  Subject: Re: Verify of Backups. Message-ID: <o5OpeRraRQns@tachxxsoftxxconsult>  V In article <2000Apr19.114247.1@eisner>, kaplow_r@eisner.decus.org (Bob Kaplow) writes:o > In article <8dkebq$fqn$1@mailint03.im.hou.compaq.com>, hammond@not@peek.ppb.dec.com (Charlie Hammond) writes: - >> In article <8djvr0$6kn$1@news.ihug.co.nz>, / >> "Ryan C. Price" <pricerc@ihug.co.nz> writes:  >> .. D >>>you can't trust your backup unless you've tested it. The only way, >>>to test it is to scan the whole tape. ... >>  E >> I disagree.  The only way to test your backup is to restore it and I >> check that the restored disk/files is/are correct.  Short of that, any M >> number of things -- Backup utility, media reliability, storage procedures, ! >> etc. -- can and will go wrong.  > H > Hear! Hear! At a previous employer, that was my #1 priority task everyI > Monday. 6 months before they hired me, they trashed their database, and M > found out that their backups were garbage. If you don't do at least regular N > restores from your backup tapes, you never know if it is possible to recover > 100% from those tapes. > M > Even further back, as a Digit, I worked with the Chicago Restart center. We L > had customers come out, restore their tapes, and run real processing. THeyJ > knew 100% for sure that not only were their backups good, but that their" > whole recovery process was good. > L >> As for /VERIFY -- if push came to shove I would probably cover myself andJ >> advise  to use it.  However, this option exists because the tape drivesO >> typically used for backup a decade or two ago were not paricularly reliable. I >> They would sometimes report a write operation completed, but not write K >> anything readable; /VERIFY caught this.  Today's backup devices are much > >> more robust.  The need for and value of /VERIFY is reduced. > L > We use Save Set Manager to dup all of our backups. The originals go to offF > site storage, and the dups to our second data center, where they areN > restored. That pretty much tells us that what we've got is good. With modern* > DLTs, /VERIFY does seem to be redundant. > : > BTW, I don't think it's been pointed out that if you useJ > BACKUP/IMAGE/IGNORE=INTERLOCK/VERIFY it is very likely that /VERIFY willN > FAIL. Any file that is either open at backup time, or is changed between theK > backup and verify passes will trigger errors. That renders /VERIFY pretty 
 > useless.    L I think "useless" is not the correct word here.  My experience has been thatJ 95% of the "errors" are merely open log files.  True, backup cannot verifyN those, but I don't really consider log files to be critical data anyway.  MostH of the rest of the "errors" are directory files to which files have beenN added/deleted and the like.  The fact that false errors are triggered does notJ diminish the value of the fact that most of the files have indeed verifiedN successfully.  I ignore the final return code of the backup and make a cursoryH scan of my backup logs to insure that no *unexpected* verify errors haveM occurred.  If the only errors are on logs, directories, and other files known / to be open, I consider it a useful verify pass.    Wayne      --  O =============================================================================== K Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachyon.xxx 8 http://www.tachyon.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-) O =============================================================================== N Butler:"Gentlemen!"  Curly(as he and other Stooges look around):"Who came in?"   ------------------------------   End of INFO-VAX 2000.226 ************************