David has, I think, already made the necessary distinction. He had substantive, as opposed to format/syntax validation in mind. There is no objection to enforcing a syntactic requirement early that it will anyway be enforced later.
There is the question whether one should bother, and the answer to it is contextual. I sometimes, for example, write macro-definition mnotes explicating errors that the assembler would diagnose anyway but too summarily or using terminology that would be opaque to non-programmer users. In general, though, redundant checks are problematic, not just wasteful. They require that error-prone collateral changes be made in too many places when a constraint is relaxed. John Gilmore, Ashland, MA 01721 - USA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
