Evaluating opinions - FilePro Error Messages

[John Esak]

>
> Now is the error wrong, misdirecting, or is it right.  It would have
> been nice if the error had said
> there was a conflict or duplicate use of the OPEN thing.

I think the error reporting did as you say "just about as good as can be
expected". Heck, it put you on the line with the duplicate... and you found
the problem.

>
> I would suggest a way to see all DECLARED VARIABLES, labels, Lookup
> aliases, Field Headings,
> Selection sets, array names, (whatever I forgot also) in one alphabetic
> list, like the declares are listed.  Include a
> code saying which thing it is.  That would make it easier to check for
> conflicts.

An excellent idea. I heartily agree. I would love to see that somewhere in
*cabe. I would also add an addendum to the problem you had above. The syntax
checker _should_ disclose and thereby prohibit _any_ use of what filePro
considers a "protected" word... certainly OPEN would fall into this
category... but what about processing written before OPEN existed?? and so
on. It is a tricky problem to deal with such a huge legacy of users,
processing, and new functionality. In any case, even just a list of words
that filePro uses might be nice if it were available from within *cabe...
and I mean put every conceivable (at-the-time-of-printing) possible word
that filePro uses... like "time", "end", "user", etc. Whether a built-in
edit or a function call, or whatever... list them ALL out somewhere. This by
itself would be pretty useful. However, I think Ken (et al) have done a
pretty good job in keeping these things from conflicting with just about any
processing you can write. So many things are legal that probably (really)
shouldn't be... the system is actually very forgiving in this regard. As
support for this statement, check out the "obfuscated" code I wrote for a
fun piece in one of the fPDJ mags. It is absolutely amazing how far wrong
you can write things and have them still act as they "should" ... within the
bounds of what logic and specified functionality dictate.

John