STL Time Machine Report #31 - Monday 05 October 1998 (1998-10-05) We're still running our test cases, which means that our mission- critical internal applications will have been run through rollover 16 times. A week ago I received an internal e-mail saying that two PC's had been dedicated for testing non-standard desktop applications, inviting users to call and sign up for test time. As I understand it, the standard apps (Lotus Notes, Office 97, NT 4.0, et cetera) have already been tested. They also announced the shutdown of the last CICS 3.3 region. It was only used for an e-mail system, and everyone now uses Lotus. This is quite an achievement. We've convert 85+ CICS regions to Release 4.1 since last November, both production and test regions. We still had a few of them at CICS 2.1.2 as recently as last May. The Voice Mail system was shut down over the weekend for eight hours for installation of a Y2k upgrade. They now have a external number that responds to voice commands and directs calls. We've had this internally for a month or two now. I still think it's black magic, speaking a person's name into a computer and having it figure out who I want to talk to. Meanwhile, c.s.y2k is going nuts with threads on oral sex, politics, wild dogs, riots, gold, religion, and freemasonry. Within the last week I saw a post on comp.lang.asm370 asking if there was any forum for discussing OS/390 and Y2k, other than comp.software.year-2000. Why Aren't More Computers Crashing? This recent thread was worth noting. I believe problems are occurring now, but aren't visible to the general public. Most organizations will not draw attention to problems if they can solve them before they become obvious to their customers. And sometimes the problems are merely an inconvenience. We still haven't seen a Y2k bug take down an entire corporation. How soon will we see this milestone achieved? My guess is mid-1999 at the earliest. COBOL Stuff There have been several discussions recently on comp.lang.cobol and bit.listserv.ibm-main about what level of COBOL is Y2k compliant. I tested OS/VS COBOL and the compiler doesn't break, but it reports compile dates in 1900 instead of 2000. The native COBOL date calls provide correct dates with two-digit years. IBM dropped support for OS/VS COBOL in June, 1994, but says the executables are okay if run with Language Environment for MVS. IBM has stated that the VS COBOL II compiler and runtime library will be supported until the end of March, 2001, but the compiler is not "year 2000 ready". There's no native COBOL II function for getting a date with four-digit year from the system. You can, however, get an APAR for IGZEDT4, which is a callable routine for obtaining the current date in YYYYMMDD format. Here's a quick example of how to use IGZEDT4: 01 MY-YYYYMMDD-DATE PIC 9(8) USAGE IS DISPLAY. CALL 'IGZEDT4' USING MY-YYYYMMDD-DATE. For the IBM VS COBOL II runtime library the APAR number is PN76666 and the PTF number is UN83685. And for LE/MVS the APAR number is PN84080 and the PTF number is UN92085. APAR = Authorized Program Analysis Report PTF = Program Temporary Fix IGZEDT4 will be supported under COBOL/MVS or COBOL/390, but those compilers support the 1989 standards which include intrinsic date functions. There's also an APAR for COBOL/MVS to make the compiler support ACCEPT MY-DATE FROM DATE YYYYMMDD (or even YYYYDDD), but I don't have the APAR number at hand. Perhaps some kind soul will post it. A funny PHM story (maybe)... My buddy at another shop tells me his PHM has become sullen, spends all his time in his office surfing the web with his laptop. I guess the client/server replacements are not going to make it in time. "What does he do on the web all day long", I ask. "I don't know", says my buddy,"he won't let anyone see, but I'm pretty sure he's trading stocks during business hours." Needless to say, my buddy doesn't work at a brokerage house. Speaking of the stock market, I was talking to a contract programmer after helping him diagnose a COBOL problem, and we got to talking about the market. This contractor has been playing the market heavily for the last couple of years, making $20,000 bets on contracting companies like AIC and Keane. He said he did great last year, but this year is not so good. He's been buying on the dips, and can't understand why they don't go back up. I don't think he is shorting his own agency. There's no reason the DJI couldn't go down to 5000, or even 3000. If =bks= is still taking predictions I expect we won't see 9000 again this millennium, and we will probably see the DJI go below 7000 this October. Here's my chance to be proved wrong. As of 1998-10-05, My countdown now reads: 27 days until 1998-11-01 (Beta Test begins) 88 days until 1999-01-01 (External Testing begins) 453 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!"