Re: Assertions in Ada

[Brian Rogoff <bpr@shellx.best.com>]

On Sat, 23 Aug 1997, Ken Garlington wrote:
> 
> Yes, I think this definitely represents my feelings at the time.
> Although I don't think anyone argued _against_ including assertions,
> I know I didn't feel that it was something we desperately needed,
> particularly vs. pragma Reviewable, etc. (Obviously, if you've
> read my Ariane responses, you know I haven't changed my mind
> yet, either!)

Are the assertions that you're discussing GNAT style pragmas or the
more complete Eiffel/Sather set? I think the former are OK, and may
become a de-facto standard. While I understand the benefits of the 
latter, I'm still an agnostic about whether they belong in the language 
proper, or whether a more complete specification language is really 
better left as part of a formal development methodology like the 
B method. So for example if B tools generating Ada were available, the
effort of putting a powerful specification tools in the language would
have been wasted. I think the decision to do nothing was wise in more ways
than one. "When in doubt, leave it out".

-- Brian