TED, re serious beer bottle site. I thought it was very interesting. A few "languages" missing - are ADS/O, CSP and Easytrieve that obscure already? Yet there is an entry in "Bobo" (The author (below) created the language. "Bobo" is Spanish for idiot).
Most interesting to me was the line that JCL "can't be done: No looping constructs." Well, who says you have to solve the problem with a looping construct? (And how are all those solutions in "batch" type languages going to execute anyway without JCL?)
Then I started thinking about JCL "solutions" to this problem, and how they might look if they had been in production for 10 or 20 years.
In the grand MIS tradition of using a sledgehammer to crack a nut, there would be a PROC with a meaningless name like NTX592D0.
This PROC has 157 steps, including SORTs, IEBGENER's, IEFBR14's, etc. Over a hundred of the steps execute COBOL programs, most of which are intended to subtract one from the current "beer bottle counter". (A couple of them add four, and are sometimes followed by programs that subtract five.) Steps 42 and 83 read in files from outside sources, who reserve the right to not transmit, or to change file formats, without notice.
The "beer bottle counter" is passed from step to step as a parameter six times, as nine different flat files (mapped to thirteen different COPYBOOK's), seven VSAM files, two IMS segments and a DB2 table. (Some of the VSAM files created during the PROC do NOT contain the "beer bottle counter", and no longer have anything to do with the PROC, but no one knows this, and in any case no one would dare touch the PROC to remove them.)
Oh, and the Ops run book would have the name and phone number of someone who left the company 5 years ago as the contact in case the job goes down.
Don't laugh, folks. Multiply this by a few hundred times, and this is pretty much how you get your bank statement and phone bill each month.
Kevin |
