MF Cobol and data validation
We are in the process of converting from proprietary mainframe Cobol to
MicroFocus/Unix. In testing, we have noticed something a little
unsettling. There doesn't appear to be any field level data validation
being performed by executable programs. i.e. A file of all alphanumeric
characters can be read by a program that defines all the fields as numeric
-- you can even perform computation on the alphanumeric data.
In our old mainframe world, the data management layer of our operating
system was linked into our programs. This checked all data entering the
program to ensure that it fit the expected characteristics the programmer
was expecting. Alphanumeric data in a numeric field generated an error and
subsequent program dump.
I know that .int modules will perform this validation, but our MF sources
strongly urge that we do not use them for production.
Am I hallucinating? If not, how are others handling this problem?
| Something unknown, is doing we don't know what. -- Heisenberg
Uncertainty Principle|
|
|
| "Hey Yogi! What time is it?" "You mean now?" -- Yogi Berra
|