Re: Language support for flexible handling of system-detected errors.

[Ingo Marks <adv@region-nord.de>]

Kevin Cline wrote:

> "Marin David Condic" <dont.bother.mcondic.auntie.spam@[acm.org> wrote in
> message news:<a8hqrv$r83$1@nh.pace.co.uk>...
>> What makes you think that there are any 100% error free Ada compilers out
>> there? :-) Its impossible to prove that there are no errors in a compiler
>> or any other sufficiently complex piece of software. And yes, I've
>> encountered errors in every Ada compiler I've ever used. It might be the
>> case that Ada compilers - having run through validation - at least fully
>> implement the language and do so in a consistent way.
> 
> I wouldn't count on that either.  Validation suites may be incomplete,
> particularly for a very large language like Ada.
> 
> I found bugs in two validated Ada-83 compilers within a month after
> I started using them.

Ok I agree with you. But isn't it better to use a "validated" Ada compiler 
(although it cannot be proved to be 100% error free - I know Turings proof 
;-) than to use C++ compilers which are proven not to be 100% error free?

The very existence of validation mechanisms extends my confidence in Ada.
They may be envitably incomplete but that's better than nothing.