The following anecdotes have been submitted by COBOL programmers. We hope that you enjoy them.
I was hired a while back to port a system from a Macintosh running a database product called Omnis, to an IBM System 36. Those who know the background on a 36 know it is really a terminal controller guts with a few niceties tacked on to allow it to compute, so it is a terribly underpowered platform for a real multi-user system. Also, COBOL is not the usual language of choice on it, but rather RPG. However, the client wanted me personally to do the conversion, and I knew COBOL better than RPG, so I convinced them to buy the compiler.
I hacked together a program which I admit wasn';t everything it should have been (the multi-user record locking was overly sensitive to a particular user who would power off in mid-session). However, the stated goal was to reduce the daily report run from 5 hours (and growing steadily) to under one hour.
After moving all the data from the Mac, and firing up the first set of reports, the
printer started chattering after about 10 seconds. The user was sure the data results must
be wrong, and sure enough, they didn't match the run from the Mac based system. After
three days of manually checking, and retesting against the production data, the user
admitted that the COBOL program was accurate, the existing Mac based system had been
giving erroneous results for the last year, and that despite the fairly light hardware,
the COBOL software on the System 36 was able to support 10 users and still execute reports
in less than 1/1,000th the time the Mac based system had done.
I'd been at my new job for only a couple of months when we got a contract that
entailed, among other things, converting standard copybooks from RPG to COBOL. The client
had built a lot of copybooks over the years and there were many inconsistencies, but the
team I worked with was confident we could assemble an operation for successfully
converting most of them, most of the time.
I got bogged down by some details on packed decimal formats, especially with respect to
the subsequent output file(s), so I spent a lot of time sketching out quick diagrams of
packed and unpacked numbers to illustrate the process to others and to double-check
myself. Someone commented that the doodling looked like little boxes, and at this point
it's only fair to provide some personal background - I had accepted this job with less
than three weeks to get moved 400 miles to my new location: finding a mover, finding a new
apartment and getting ready was an incredible (and educational) experience, but I did it.
Only, my new apartment was a lot smaller than the brochures would have you believe, and
this became my favorite source of griping ("My apartment is so small!"
"How small is it, Suzanne?" "The cockroach is a hunchback!!").
I've concluded I'll get all my stuff unpacked, moved in and organized probably a week
before my lease is up.
Late one tiring Friday afternoon towards the end of the project, one of my co-workers asked me how I was doing with unpacking. I replied that the packed decimal format was no longer a problem. He looked at me strangely and said he meant the boxes, how was I doing with getting boxes unpacked. Still clueless (and starting to wonder why he was so fixated on this annoying detail) I said that the concept of boxes was a good one for illustrating packed and unpacked formats, wasn't it. Ever see Homer Simpson strangle Bart? When it finally dawned on me what the poor guy was trying to say (and it was nice of him to believe I actually have a life), we both laughed and I apologized for being stuck in a box.
flexus
A Company of Interex Software, Incorporated
P.O. Box 640, Bangor, PA 18013-0640, U.S.A.
Telephone: 610-588-9400
Telefax: 610-588-9475