STL Time Machine Report #25 - Tuesday 28 July 1998 (1998-07-28) It's been just over two weeks since the last TMR. It's summertime and not much is happening. It seems that Group 3 testing has had the usual setup problems but they're working through their tests. Some least critical applications are still being modified, but they can be tested with date override simulators. More on the "Summer of Upgrades" We're starting to upgrade our Unix servers. Since we have over 500 of them it will take nearly a year. They're doing it in conjunction with a disk upgrade to allow larger data files. One of our three mainframes is scheduled for an upgrade to faster engines on August 8. Technical services is already testing OS/390 on the new box and they're going to force all the programmers to convert to TSO/ISPF and SDSF. We currently use H&W SYSD, which is an ISPF lookalike that runs in CICS. We're also supposed to swap out our solid-state DASD for fast cached RAID DASD late in August. It's going to be a busy month, and we have more testing to do before the Thanksgiving holidays. COBOL I have a 1998 objective to get a minimum of 40 hours of training. I am now signed up for a class called "Beyond COBOL II" next November. Eighteen years experience and they keep changing the Language on me. High-Level Assembler. My longtime project leader is moving to a client/server project that's three years behind schedule. Last Friday at lunch We teased him about turning traitor. They asked him to take the slot (translation: We need somebody who can actually build something that works). He's cleaning out his desk and handed me the file folder on testing HLASM for Y2K. Thanks a lot. I really didn't want to see him go. At least he'll be in the same building. Actually, I had some fun a couple of weeks ago. During our Y2K inventory in 1996 I found an assembler program we run every day with no source code. I've been bugging one Tech sub-group for a disassembler for months, but they can't get me one. So I sat down with my 1984 S/370 yellow card, an old Feingold S/360 assembler textbook, an AMBLIST, and an IDCAMS dump and disassembled that sucker by hand. It starts out: 90ECD00C = STM 14,12,12(13) 18CF = LR 12,15 1821 = LR 2,1 4510C074 = BAL 1,116(12) DC 18F'0' DC CL32'** QR608010 &SYSDATE &SYSTIME **' 50D10004 ST 13,4(1,0) 501D0008 ST 1,(8,13) 18D1 LR 13,1 1812 LR 1,2 05C0 BALR 12,0 USING *,12 (I presume) 58A10000 L 10,0(1,0) 58B10004 L 11,4(1,0) 58910008 L 9,8(1,0) It was tedious, but interesting. That BALR 12,0 displaced all the addresses, which took some decoding. I built test JCL to assemble the decoded file and compare it to the production load module. I put in decent comments, labels, and datanames, and set up some JCL to run a parallel test against the production program, to verify the recovered source gives the same results. By Wednesday I had a decent source file for a colleague to review, and incorporated some good insights that made the recovered source easier to understand. This program doesn't have any date issues (apart from &SYSDATE), but it did have a very interesting use of a TR translate table, and it used a BXLE (uses 4 registers, but only specifies 3)! Source recovery is not too difficult, if it's a small program. I wouldn't try this out on a large COBOL program without a real disassembler. I had so damn much fun solving this problem I felt like taking a stab at writing my own disassembler. It would be fun to do, but so much easier to just buy one off the shelf. Miscellaneous Notes A buddy at work got a brochure with his Ameritech Cellular phone bill. Call them at 1-800-221-0994, or visit their Year 2000 web site at: http://www.ameritech.com/Y2K He also got an insert with his bank statement begging him to call his bank. "We'd be happy to discuss our Year 2000 Goals!" Awareness is rising, even if futile . There's been quite a bit of discussion in c.s.y2k about the SIA Wall Street testing. I think it's good news, but they're not done testing. They're taking baby steps, building their time machine. It's the beginning of the job, not the end. Here's the link: http://www.sia.com/year_2000/index.html Until the Y2K movie comes out, we can hold up on the comparisons to "Titanic" and start using "Saving Private Ryan" as a metaphor for Y2K work. The battle scenes are almost as violent as some of the meetings I've been in (or usenet discussions I've seen). Interesting URL's I've seen a couple of requests for the FAQ, but I haven't seen Pam Hystad repost it in the last few days, so here it is: http://www.computerpro.com/~phystad/csy2kfaq.html The entire collection of Cory Hamasaki's DC Y2K Weather Reports: http://www.sonnet.co.uk/muse/dcwrp.html As of 1998-07-28, My countdown now reads: 96 days until 1998-11-01 (Beta Test begins) 157 days until 1999-01-01 (External Testing begins) 522 days until 2000-01-01 (Rollover) Procrastination is futile! Previous Year 2000 Time Machine Reports are available at: http://home.att.net/~arnold.trembley/tmr.htm STANDARD DISCLAIMER: I am NOT an official corporate spokesperson. My opinions should not be held against my benevolent employer. -- Arnold Trembley http://home.att.net/~arnold.trembley/ "Y2K? Because Centuries Happen!"