PAAAAAR@CALSTATE.BITNET (06/02/89)
Some one asked about NonDP COBOL projects -
At my old University - Long Ago and Far Away
I taught myself COBOL by
(1) writing a program to analyse the behaviour of finite state machines
The FORTRAN 4 version never worked. Cobol took one week!
(2)designing a COBOL emulator for Knuth's MIX computer.
the following tools were developed using COBOL
COBOL Compiler using "Recursive Descent"
"Recursive descent" is only "recursive" when the syntax
has a non-tail recursion - and COBOL didn't.
PS - last time I read ANSI Specs of COBOL and FORTRAN neither
forbade recursion...
and its easy to develope clean code
for implementing recursive PERFORM for COBOL
(You have to stack both Entry and EXIT paragraph/section!)
COBOL Cross reference - 2 versions.
One with exceptionally evil structure because the programmer
had been working at a shop that only permitted ONE read
statement per file per program!
The above were in use for several years.
Decision Table to COBOL translator.
The COBOL secretary converted ALGOLish code into COBOL.
We also had a one or two competitions in the department (students and faculty)
to do various things - free choice of language - COBOL came off equal or better
than most languages we used - The compiled code tended to be smaller for small
programs for example, and the COBOLer had fewer bugs BUT this experiement is
not valid because the people doing the programming were all very different.
In my humble, but grey haired opinion, its bad craftspeople that blame their
tools. (present company excepted)
Dick Botting
CSU San Ber'do