STL Time Machine Report #23 - Sunday 5 July 1998 (1998-07-05) Group 2 testing is back on track now that Unix log files are coming in with post-2000 dates. We found a new defect in one of our batch systems. This is the second time I've seen this kind of problem. It should go into the category of "stupid windowing tricks". Here's how it happens. A record comes in with a two-digit year in the date. So it must be windowed into a four-digit year in order to be processed correctly. But the windowing logic tests the two- digit year of the current date instead of the record date. Voila (as the Germans would say), we now have records with dates in the year 2099! The last time I saw this was in an on-line system, and records a few seconds old were incorrectly windowed. This time it happened in a batch system, and an entire input file was incorrectly converted. I'm surprised this one got through unit testing. Don't let this happen to you! The Summer of Upgrades (Again) Last week was a good week. We upgraded several CICS regions from release 3.3 to the Y2K-compliant release 4.1. This puts us ahead of our own technical services department (not that we're competitive or anything). To be fair to them, they've had some problems upgrading the CICS regions that include the numerous third-party packages, including our database. It's much more difficult to integrate all the purchased software. Guerilla Remediation I've passed on several anecdotes about my buddy with the PHM at another company. They recently had a success when they tested an entire batch system in their VM/ESA Time Machine, without any authorization from management. Some minor bugs were found and corrected. The original author of the system still works for them, and was glad to have the opportunity to do some coding. This system was fixed for Y2K without unit testing. I think it's better to approach the problem in an organized way, with a standard approach, but guerilla remediation is better than no remediation. Interesting URL's Staffan Ottosson posted this URL in bit.listserv.ibm-main on July 2: http://www.software.ibm.com/year2000/y2ktestsystem.html It's on IBM's Y2K pages and includes a lot of tips for setting up a Time Machine on a System 390 computer. They cover MVS, VSE, VM, and PR/SM. They have tips on isolating your disks and managing tapes. It's worth a look. I've added it to my TMR web page. Several times I have mentioned the IBM Time Machine for CICS. This is a supportpak, a free download, licensed software with no fee (and no guarantee). I can tell you it works, and is very handy for unit-testing your CICS applications: http://www.software.ibm.com/ts/cics/txppacs/ch15.html You may have to search lower to get to the download button, but this should still be the right address. Since I'm throwing around URL's, here's a few more (I'm trying to boost my stats). I hope they are still up-to-date. My thanks to Chris Anderson, who posted most of them already: The COBOL Center: http://www.infogoal.com/cbd/cbdhome.htm Click on "COBOL Example Projects and Code", then "date.zip" to download the full set of date routines. Microsoft patches for Windows 3.1: http://www.microsoft.com/ithome/topics/year2k/product/wind31.htm Microsoft patches for Win95: http://www.microsoft.com/ithome/topics/year2k/product/win95.htm Windows NT 4.0 Workstation fixes: http://backoffice.microsoft.com/downtrial/moreinfo/y2kfixes.asp NT 4.0 Server, standard & enterprise editions: http://www.microsoft.com/ithome/topics/year2k/product/WinNt40SvrUS.htm AT&T's Year 2000 page: http://www.att.com/year2000/ Here's all of Cory Hamasaki's DC Y2K Weather Reports: http://www.sonnet.co.uk/muse/dcwrp.html Re-read 'em, trade 'em with your friends, collect the whole set. Here's the REXX program I use to calculate my countdown to New Year's Evil: /* rexx */ teotwawki = 730119 /* 2000-01-01 */ dd1999 = teotwawki - 365 /* 1999-01-01 */ dd1998 = dd1999 - 61 /* 1998-11-01 */ today = date('b') date1 = dateconv(today, 'b', 'n') date2 = dateconv(dd1998, 'b', 'n') date3 = dateconv(dd1999, 'b', 'n') date4 = dateconv(teotwawki, 'b', 'n') count2 = dd1998 - today count3 = dd1999 - today count4 = teotwawki - today say ' ' say date1 'is' count2 'days from' date2 say date1 'is' count3 'days from' date3 say date1 'is' count4 'days from' date4 say ' ' exit 0 Here's the results: 5 Jul 1998 is 119 days from 1 Nov 1998 5 Jul 1998 is 180 days from 1 Jan 1999 5 Jul 1998 is 545 days from 1 Jan 2000 Unlike Cory's REXX (published long ago in a DC Y2k Weather Report) my routine assumes that 1999-12-31 is one day away from 2000-01-01. Cory Hamasaki's routine assumes you don't have ANY full days remaining on 1999-12-31. Please let me know if I should convert my countdown to Cory's count. As of 1998-07-05, My countdown now reads: 119 days until 1998-11-01 (External Customer Test Beta begins) 180 days until 1999-01-01 (External Customer Testing begins) 545 days until 2000-01-01 (Rollover) Previous STL 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!"