Last week I saw references to both COBOL and DOS (see this FCW piece and this piece from Slate); I think both in connection with unemployment insurance systems which are running on ancient software. I never did much programming with DOS (I was more into WordPerfect macos) but I did take courses in COBOL and did one application as a sideline to my regular job. The System/36 ran COBOL as did the mainframes in Kansas City.
I can understand why both private and public organizations still run COBOL. Every change of software runs the risk of creating new problems, so if you've got an application that runs without problems and supports the organization, there's little reason to switch to a newer language. That's particularly true if the organization is adding new programs or functions, so available people and work hours are needed to support the new.
All that said, the downside of keeping the old programs is you have to live with the old silos and the old thinking, forgoing chances to integrate, and likely forcing you to invent kludges or bridges on occasion. For example, with issuing the federal payments under the current program (CARES), I suspect Treasury had to write new programs to match ID's in IRS files against those in unemployment files.
No comments:
Post a Comment