STL Time Machine Report #35 - Saturday 12 December 1998 (1998-12-12) I survived my on-call rotation week carrying three beepers. Even though I was paged every day, I never had to come in, and none of of them were for Time Machine problems. We haven't had any date problems in our beta testing. I recently attended another all-employee meeting, and Y2k was on the powerpoint screens. A very high level manager expressed high confidence in our readiness, and congratulated the geeks who built our Time Machine. But he also announced that Saturday, 2000-01-01 would be a work day. He went on to say that we might get some advance warning of problems, since rollover hits in New Zealand, Australia, Japan, and Europe several hours before St. Louis. Normally, Monday 2000-01-03 would be a holiday for us, but staff is planned to be on-site and on-call to cover service level agreements. There was no mention of fix-on-failure. "VISA is Toast" Andrew Rowland's thread on "Visa is Toast" was interesting, but I don't completely buy it. I don't believe that date failures will change a $100 credit card transaction into a $1 Million purchase, but hey, I could be wrong - I'm not a Big Brain. I passed some of the more interesting posts on to our Y2k project director. This is a guy who gets around. He meets with other companies on Y2k conferences. Visa has about 50% of the credit card market, about twice as big as their nearest competitor. Visa has the money to do the job right. If the codeheads in San Mateo can't get their act together, Paul Milne will be going BUHWAHAHAHA when the lights go out. As reported in c.s.y2k, someone tested a POS terminal in 1997 with the phone line unplugged, and it rejected an '00 card as expired without attempting to dial out. Most of those pesky POS terminals have been fixed for Expiration Date edit problems, but how does the POS terminal know what the current date is? It must have some kind of internal clock/calendar. Some proportion of those POS terminals will not correctly roll the current date on 2000-01-01. So how do you set the date on those POS terminals? Meanwhile, c.s.y2k is going nuts with threads on Jack-Booted Thugs, drug laws, impeachment, stock market manipulations, purple spiky haired cannibal mutants, food shortages, nationalization of the trucking industry, and my personal favorite - can we delay the rollover and call 2000 a different year? The "Jo Anne Effect" Don Scott doesn't quite believe in JAE, but I know it's happening. The real question is, will it cause any failures that can be seen by the public? I don't think so. But if we keep score, if we hear of a big-name company that goes bankrupt due to Y2k that will be a clue. "All of 1999 for Testing" -bks- made some interesting points in the "all of 1999 for testing" thread. Y2k is a collection of maintenance projects. It does not take a year to test a single system for Y2k compliance. In my limited experience an application with a dedicated programming staff of five to eight people takes 3-6 months to analyze, repair, and test the app. Time Machine testing took two additional months per application, and a separate team to support the Time Machine. The reason you need "all of 1999" for testing, is to test your interfaces with your customers/suppliers, and to re-test to make sure that regular maintenance doesn't introduce new date problems. The Y2k code is already in production. Internal compliance is hard to achieve, but it's easy compared to testing for external compliance. Normally, we have two software releases a year, but we won't do a second one in 1999 in order to keep the system stable for rollover. This has the curious side-effect that we are actually letting contract programmers go because there's less programming to do! And December is a slow month. There's the Christmas parties, the potluck lunches, the plates of cookies and donuts. People are using up the vacation time they didn't get to take during the summer of upgrades. I'm hearing about contractors going on the bench because Companies don't start new projects between Thanksgiving and Christmas. Two years ago, I was coming to work in an ice-storm to perform scans on our mainframe code. The temperature dipped to minus 10. This December we had 70 degree weather until this week. A tremendous amount of work has been done in two years. There's still a lot of testing to do. Joke of the day, seen on a manager's cube wall: A Bus stops at a Bus Station. A Train stops at a Train Station. On my desk, I have a Work Station. Another PHM story (it's a Dilbert World!) Here's the latest chapter in the ongoing saga of my buddy at the company with the CIO/PHM. The auditors came in and interviewed the applications staff. I don't know if the sound of forehead slaps could be heard in the bullpen, but the universal answer was, "we don't do project plans here". So, no Y2k plans, no inventory list, no risk assessment, no specs. The auditors are now writing up their report in a closed-door office. I wonder what it will say? The CFO wants quarterly status reports. As of 1998-12-12, My countdown now reads: -41 days until 1998-11-01 (Beta Test underway!) 20 days until 1999-01-01 (Testing begins) 385 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!"