STL Time Machine Report #33 - Friday 6 November 1998 (1998-11-06) Several milestones have passed since the last TMR. Cory Hamasaki's DC Y2K Weather Report rolled over from 99 to 100, obliterating my DOS 8.3 file naming convention. The end of October came, and instead of the DJIA falling below 7000 (as I predicted) it rocketed back up to 8592, and closed Thursday near 8900. If this had been a bar bet with -bks-, he would have won, big time. Most of the hassles in our final stage of internal compliance testing were related to setting up automated job schedules for the Time Machine. As I mentioned some time ago, some groups tested with JCL overrides, rather than building full job schedules. November and December are planned for beta testing of our Time Machine, then we open the doors in January 1999. HLASM, the TIME Macro, GMT, and Y2K There was quite a controversy a few months back about converting old S/370 assembler programs to the new High-Level Assembler (HLASM). They compromised by converting about seven of them, and testing them in the Time Machine. One of them broke in production a couple of weeks ago, due to a time bomb that had been sitting in the source code for fifteen years: ******************************************************************** ** ACQUIRE CURRENT DATE AND TIME ******************************************************************** TIME STCK,WORK8,LINKAGE=SYSTEM **Y2K** XC TIMEDATE(16),TIMEDATE ZERO DATE FIELD **Y2K** STCKCONV STCKVAL=WORK8,CONVVAL=TIMEDATE, **Y2K** DATETYPE=YYYYDDD The preceding code uses the STCK option of the TIME macro, which gets the date/time from the system in GMT. For fifteen years we've been booting up our MVS with GMT time set equal to Local Time. In October they installed a sysplex timer, and started setting true GMT time on bootup. The preceding code example now gives tomorrow's date after 19:00 Central USA Time, instead of today's date. Here's the corrected example: ********************************************************************* ** ACQUIRE CURRENT DATE AND TIME ** ********************************************************************* TIME DEC,TIMEDATE, GET DATE AND TIME **Y2K** X ZONE=LT,LINKAGE=SYSTEM, X DATETYPE=YYYYDDD * CONVERT TO TODCLOCK CONVTOD CONVVAL=TIMEDATE,TODVAL=WORK8,TIMETYPE=DEC, **Y2K**X DATETYPE=YYYYDDD It only took a couple of days to install this fix, but the error had caused us some squirrelly problems for several days before the bug was diagnosed. The code was fixed for Y2K to get a four digit year instead of a two digit year, but it didn't always return the correct date! The Summer of Upgrades - 1998 Well, summer has been over for a while and we've had some chilly days, but the pace of upgrades hasn't fallen off much. In the last TMR I mentioned we were planning to install LE/MVS as the default runtime libraries under MVS/ESA 5.2.2, for batch only. It went in, but there were a few problems. Some existing COBOL II programs were missing output DD's in their JCL. Under COBOL II, those files were silently created and deleted. Our LE/MVS was installed with an option to abend the program if the DD was missing. This restored the behavior that we used to get with OS/VS COBOL. But overall LE/MVS is working fine. This means that we can finally begin converting to COBOL for MVS and VM. That's a gradual change that will be much easier than conversion from OS/VS COBOL to COBOL II. Election Results You can skip this part if you want. I ran for Missouri State Senate as a Libertarian. I couldn't say that I campaigned, since I did the minimal amount required to be a legal candidate. I spent $24 to participate in a debate moderated by the League of Women Voters. I did not embarrass myself too badly, although I doubt I converted anyone to the Libertarian cause. Got to meet my Democrat and Republican opponents, and lost in a dignified manner: (R) Betty Sims (Incumbent) 25,269 54.25% (D) Howard Shalowitz 19,991 42.92% (L) Arnold J. Trembley 1,320 2.83% I don't recommend running for office. It takes a lot of time, even if you don't campaign. You will get phenomenal amounts of mail, constant deadlines to meet responding to special interest group surveys, newspaper requests, and campaign finance reporting. I attended a very interesting seminar sponsored by the Missouri Ethics Commission on campaign finance laws (and what I had to do to comply with them). Paying taxes is a lot easier. If you have to run for office, be sure to get some pictures of yourself for the newspapers, wallet size is acceptable, and some short position papers, "why you should vote for me", "my stand on the issues", that kind of thing. A lot of people around the office assumed I was going to quit my job and become a full-time politician. NOT! The woman who runs the function point group had no idea I was running, but she saw my name on the ballot and said, "Hey, I KNOW this guy!" She later told me she had never had the opportunity to vote for anyone she had personally met. So she voted for me. I'm not really a "people person", don't have the Clinton knack for remembering the name of everybody I've ever met, but personal contact has a big impact in political campaigns. If you're running to win in a small district, you want to ring every doorbell and make sure every household gets a flyer with your name on it. You don't want to pass up any opportunity for TV coverage (I skipped all of them). I don't know how anyone with a real job can find the time to run an effective political campaign. Another PHM story My buddy at the retailing company attended an all-employee meeting on Y2K. The CEO was very concerned, he told the entire staff that their IT dep't was going on a death march. There was a chart listing all their applications that have to be fixed. They're starting now and their strategy is still to replace everything with client/server and purchased software like PeopleSoft. They're starting to realize they're not going to make it. A Geek bails We lost a geek last week. Dan S. decided to move back to California. He has relatives there and I'm sure he got a big raise to take that new job. We wish him well, and we will miss his skills. Right now, the work is winding down a bit, and after I get done with the project I'm on right now, our workload looks a little light running into 1999. We'll have some time to tune the system, apply production fixes, and get ready for rollover. As of 1998-11-06, My countdown now reads: -5 days until 1998-11-01 (Beta Test has begun!) 56 days until 1999-01-01 (External Testing begins) 421 days until 2000-01-01 (Rollover) 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!"