> >Plus, those darn Ada compilers just don't work! The code
> >just never seems to compile the first time, like it does in
> >VC++. ;-)
> >Mike Berman
> Fellow Teamers,
> disregarding the smiley above, this is actually one discussion we ought to
> have. Tomorrow, I'll be heading in harms way for the future focus of R&D at
> my company.
> I know that there will be discussion about the use of Ada, and perhaps one
> argument will be: "In the early 80's, compiler vendors put Mega$ into their
> compilers and tools, and they had just about reached maturity in the late
> 80's. Now that the massive $ are out of Ada, there will never be mature
> compilers and tools." Then comes a few unverified rumours that ObjectAda
> (or any of the other compilers that I have not personally used recently)
> barfs on the third level of generic instantiation, or something.
> What's your HtG understanding of the state of the vendors and products?
> To what extent does CLAW, ObjectAda and the others match the Delphi and
> VC++ on the major platform (i.e. NT in the near future).
I think if you are concerned with correctness of compilers, you will
find that Ada compilers still outshine most C++ compilers, and will
continue to do so. There are several reasons. One is that C++
as a language standard has refused to settle down, and still hasn't
achieved standardization. Another is that validation and conformance
to standards is taken more seriously in the Ada compiler community
than in the C++ compiler community.
Certainly if you start talking about three levels of template
instantiation in C++, you are in uncharted territory.
There is certainly a transition issue now as systems are
moved from Ada 83 compilers to Ada 95 compilers. In general,
any large system is "trained" to the particular compiler used
to develop it. As you move a large system to a different compiler,
you will inevitably hit areas of weakness in the new compiler, and
have to retrain the system (through work-arounds), or the compiler
(through compiler support), for the two to "get along" as well
as the system did with its former compiler.
This is a fact of life of transition for any language and any large
system, in my experience.
> Daniel Wengelin <[log in to unmask]>
-Tucker Taft [log in to unmask]