Mike Carrell wrote:
The Y2K crises was averted by *very* intense activity by *many* programmers in *many* responsible industries to find and correct the old computer code that could have cuased some very nasty results. Those warnings *were* heeded by responsible management.
That's true, although as I said it was not "averted" so much as confronted and paid for -- paid for through the nose. As I said, responsible management took care of the problem in the 1980s during routine maintenance. Since programs have to be updated anyway, the extra work cost practically nothing. I recall that it took me only a few days to replace and test all of the date routines in a large software package. It was easy because I write modular, object-oriented code with only one set of routines for all data-and-time procedures. Nowadays all software is written in this style.
Some older programs in COBOL were disorganized, with date routines repeated in different ways, and many different methods of inputting and printing dates scattered around the code. And they used the infamous 2-digit-year dates, something I never did, even back in 1978. (I used Julian dates.) This probably would have cost a great deal of money to fix even in 1982, which I suppose is why it was neglected until 1998 by some companies, when it had to be done in a rush job. It was even more difficult in some cases because the source code was lost.
Faced with this problem, many corporations decided to throw away the COBOL programs and mainframe hardware, and start over from scratch with personal computer hardware. This was probably cheaper in the long run. They had to make the transition sooner or later. In that respect, the Y2K crisis did produce side benefits, by forcing the widespread adaption of modern hardware and software. This is analogous to saying that a hurricane sometimes helps city planners upgrade the sewers and power systems, because all the houses are knocked to pieces and bulldozed out of the way.
- Jed
