To: John Mansfield who wrote (1384 ) 4/3/1998 2:27:00 PM From: John Mansfield Respond to of 9818
'Y2K code fixes will cause problems' ********************* Well recently IRS computers erroneously declared about 1,000 taxpayers who were current in their tax installment agreements were suddenly declared in default due to a programming error. The source of the problem, technicians found was the result of an attempt to fix a Year 2000 issue in one of the IRS computers. **************************y2knet.com Well, this is proof of what several of us quality-oriented people here have been preaching over the last week or two: Y2J code fixes will cause problems - some of them horrendous. The software development life-cycle MUST include several phases of testing, finding new bugs, recoding, retesting, etc., etc. or there WILL be bugs at delivery date, and there WILL be outages and failures. Normally, software is delivered/shipped with some pretty horrendous bugs and the end users are USED (as in abused) for system testing purposes. Their inconveniences are not a big deal, harr harr, as they perform a necessary function to software developers who unscrupulously foist their trash code on paying customers. Y2K is not the normal scenario, though, and shipping buggy code will be horrendous in its result. Bad habits, driven by greed and unprofessional behavior is the real culprit in Y2K. The IRS is obviously making fixes and implementing them on-the-fly to some degree, and this is going to be the norm for all agencies or companies who are pressed for time (virtually ALL). As the delivery date draws closer, more fixes will be made by bleary-eyed and over-caffeinated programmers at midnight. The fixes will be implemented with little of no testing. You can guess the result of putting that code on line, for real, with real folks' money and lives on the line. If this happens with the utility grid, how many government or private side programmers and QA people can geek out in their cubes by the week fixing all these "new" bugs discovered on 2000? Few, if we're lucky. No electricity, no geeking. No geeking, no fixing. No fixing, no society as we know it today. Imagine a terabyte or more sized database having corrupted data (corrupted by implementing last minute, untested, bad Y2K fixes on 2000) discovered only after it's too late? Can you imagine trying to fix a terrabyte sized database that is botched? Can you imagine if it's a tax database or bank database or mutual fund database and your money is involved? Jim Lord recommends getting hard copies of all tax status and account status before this all goes down. I'm thinking he's onto something. We suffer from a seriously faulty development life-cycle in software development today. I believe that it will manifest in a pandemic sort of way, worldwide and simultaneously, and is going to be felt in real-time by billions of frightened people. This is not a crisis of confidence (in banks and government), as some naive people want to believe. It is a crisis of a software development life-cycle that has been perpetually abused, worldwide, and is not respected or planned for or adhered to by people who think they know better. - pl _______ From: paul leblanc Newsgroups: comp.software.year-2000 Subject: IRS Y2K fix backfires Date: Fri, 03 Apr 1998 13:29:31 -0500