Usually, Monday is not a good day for me to actually get work done. Today was an exception! I even missed the shuttlebus (don't worry, there is still a T-bus) because I was starting to work on a Y2K fix for some FORTRAN77 code that was originally written in F66. I'm not sure how old it really dates back to, since the first documentation dates to the conversion (in 1986). And I was enjoying it. I'm a sick, sick puppy.
I spent most of the day writing a (perl) wrapper for this thing, since I need to run it many times. On a data set that spans 1999-2000. I tossed in all of the date-conversion stuff automatically into the perlscript, but didn't notice until later that the underlying FORTRAN will assume that I mean 1900. Silly code.
I really, really prefer FORTRAN to perl. This is rather pathetic. I was all gooey (not GUI, mind you) over mucking around with reading input files in FORTRAN, only minutes after I scowled harshly at my Camel Book for the pain that was involved in doing the same thing in Perl. (It was one of those rare cases where using pattern-matching was more of a pain than less flexible methods.)
Hmm. Actually, I don't know if I should fix this, since it has to be compatible with another, much larger, piece of legacy code, which is still missing it's Y2K fix. (And I'm not writing that, because that's someone else's project, and it would break all of the carefully constructed hacks in the perlscript wrappers for that code. It will be someone else's job to fix those, too. Heh.)
I think I can get away with, for now at least, only processing the 1999 data, and plastering big comments saying NOT Y2K COMPATIBLE over everything else. But I'm still going to print the code out now and see if I can figure out an elegant fix on the T ride home. 'Cause if I can, everyone will love me forever. (Or, at least, a small subset of satellite-tracking people will be mildly grateful.)