Re: INTEGER & VARCHAR() on I-specs

On 1/15/2016 1:29 PM, Jon Paris wrote:

" I'm wicked comfortable ...” - but that is the problem Buck. Maintainability should not rely on the skills of one person.

I couldn't agree more!

But (you knew it was coming) I am here, I'm already collecting a salary,
and it's cheaper to have me muck with it every so often than it would be
to have anyone completely rewrite it. Remembering that a rewrite would
not improve function at all; it would be treading water in order to make
it easier for someone to maintain. It /can't/ improve the functionality
because the entire process works this antique, batch way :-(

I stopped using the cycle when a) I found that I was fighting it more than it was helping and b) When it took 5 - 10 times longer to educate new programmers in how to modify complex cycle programs than if it had been written using “normal” techniques.

I am absolutely not advocating writing new stuff this way! Rather,
there is a certain real cost associated with rewriting a program, and at
the moment, that cost is far higher than the maintenance cost. When the
day comes that my outside parties finally step from the Stone Age into
the Bronze Age, I'll be the first one to plead for a rewrite. Because
at that point, the existing functionality will be obsolete, and with it
the programs that implement that obsolescence.

Can’t help but wonder if the OPs original program couldn’t have been rewritten in less time than it took to research and fix the VarChar problem.

Often, that isn't clear until I've actually taken one path or the other.