Re: Ada validation is virtually worthless

[dewar@cs.nyu.edu (Robert Dewar)]

Despite what Raj said, the DOC has been an extremely effective tool in
preventing extensions in Ada. Ada is the only language I know of where
the incidence of vendor extensions of the language is essentially nil.

Obviously there is no objective way to tell if a compiler has extensions,
but in practice the requirement to sign a legal document declarting that
you have no deliberate extensions is a very strong control over such
extensions, considering that nearly all other compilers for nearly all
other languages *do* have deliberate extensions of one kind or another.

I suspect Raj has not really bothered to investigate validation very
much, and has not for example, examined VSR's in detail. I usually find
that people making glib statements about validation like this don't know
very much about it!