Re: Infinite Loops and Explicit Exits

From: Peter Lacey (lacey_at_mb.sympatico.ca)
Date: 10/31/04 

Date: Sun, 31 Oct 2004 14:47:55 -0600

Rick Smith wrote:
>

>
> I recognize the truth in what you say; but until WG4 takes
> an overt step to deprecate PERFORM ... THRU,
> multiple paragraph sections, and GO TO; such as by
> make these archaic, then what I have said remains valid
> for standard COBOL.
>
> When I discuss standard COBOL, my intent is noble;
> that is, I try to rise above the petty; but, in my own use
> of COBOL, I embrace the petty. Mr Plinston, you are
> preaching to the wrong person.

When I see posts like this, it seems to me that the purpose is to force
the elimination a style of programming (which is perfectly valid,
productive, and reliable) in favour of one which the author of the post
prefers. The ostensible reason is to eliminate errors which can be
caused by the misuse of the statements involved. Why not then eliminate
every verb which can be misused? Consider the COMPUTE verb: if somebody
accidentally uses a "+" instead of a "-", or enters 3.15 instead of 3.14
in his computation - dreadful possibly even catastrophic results can
occur. Get rid of the thing!!!

My example is facetious, of course, but my point remains valid. There
are very few verbs in the Cobol canon which can't be misused. Seems to
me that the argument ought to be applied everywhere.

The misuse of GOTO's only applies during the debugging and testing
phase, anyway. It isn't as if a GOTO will suddenly misbehave after
years of laying in wait. If a situation hasn't been anticipated by the
programmer after the same years and the program misfunctions as a
result, it's a design fault, not the fault of the verb in question.

I agree: Djikstra and his acolytes have done us a dubious service.

PL