<<BM, yes, yes, and yes. There are some old legacy software programs, mostly COBOL and FORTRAN, with two digit dates. Even very few of these will fail at y2k because we programmers (yes I am an old time COBOL/FORTRAN programmer) were smart enough in the 1960's and 1970's to look beyond 2000. Most of these old programs with two digits span the period 1961 to 2060, that is to say that 61 to 99 are assumed to be 1961 to 1999 and 00 to 60 are assumed to be 2000 to 2060. Very few of the old programs will fail in year 2000, but look out for 2061!!!>>
Roger, as a fellow programmer, I disagree with you completely. There are many billions of lines of code written in COBOL alone that are still in use. Very little of it is y2k compliant and must either be changed or replaced. As to the method of storing the 2 digit year and then internally coding your way around the deficiency is only a partial solution. As you know, in order to correctly sort a file or read it index sequentially, dates are stored as yymmdd (or some variation thereof). The lack of at least one more digit to correctly index the file is the problem. Therefore, the only somewhat foolproof solution is to expand the date fields in file to (c)cyymmdd and then track and change where necessary these fields throughout the code.
I say "somwhat" foolproof because there is no magic bullet out there (just some software aids) and the solution to me the problem is akin to finding needles in haystacks without a counter to tell you when you've found them all, bringing up the possibility of problems as bugs are detected in the ensuing years.
No-one can accurately predict what will happen but the ignorance and complacency that man shows towards this problem is going to come back to haunt...
As for being smart to code your way around the problem back in the 60s/70s, you are obviously the exception, because there are many applications written well into the nineties that failed to take the problem into account.
Neil. |
