The copy books have are no constraints, no DRI actions and if you read
a file with one DATA DIVISION from a copy book, it can cheerfully cut
up the bytes and fit them into that template without an error. My
favorite horror story from QDB was a hospital that hung an MRI mag
tape on a patient report application. The records in both files began
with the patient id and that was all that was validated. It ran.
I seem to remember that data dictionaries first appeared outside of
COBOL as part of the Structured Programming/ Structured Analysis
movement. Yourdon, Gane & Sarson, Warnier_Orr, etc. methods defined
a simple model that then grew over time when we got interactive
development tools.
Received on Wed May 02 2007 - 21:29:01 CEST