Re: Is the "Ada mandate" being reconsidered?

[stt@henning.camb.inmet.com (Tucker Taft)]

Tom Robinson (robinson@gdesystems.com) wrote:
: ...
: I think that the Ada market is in transition.  

I think the same statement can be made about the whole compiler
industry, and especially the C++ market, by the way.

: ... As an Ada user the transition
: to Ada 95 seems like it might be a bit tricky.  For large projects it looks
: like there are going to be more decisions that I (as a buyer of Ada 
: technology) will need to make:

:   (1) Go with gnat.  Do my own maintenence of the "free" compiler.
:   (2) Go with gnat.  Pay someone *ACT?* to provide support.
:   (3) Buy a product from a vendor that does not have direct support
:       for fixing their own front end problems (AdaMagic).

Note that this approach to having multiple compiler vendors
use a common front end has a long history in the C, C++, and Fortran 
marketplace, and is even increasing in the C++ marketplace with the 
growing predominance of Edison Design Group's C++ front end.

The justification is simply that the more people who use the same
front end, the fewer bugs will be left in it for *you* to stumble
over.  In fact, the proliferation of Ada 83 front ends I believe
was one of the major problems with the Ada 83 marketplace.  Having
to maintain all of those front ends through the multiple ACVC releases
was extremely expensive for the collective Ada 83 client base.
Having to deal with the peculiarities of different front ends when
moving from target to target added to a user's porting costs.  And by
having multiple front ends, it was more difficult for all the vendors
to agree on common pragmas, attributes, etc.  

With the new, smaller number of front ends, more coordination 
is likely (e.g. the GNAT and the Intermetrics folks keep in 
close touch on pragmas and attributes, etc), and you can
more likely find another compiler based on the same front end on 
your next target.

As an anecdote, there are several users who have found that 
the biggest expense in porting from Ada 83 to Ada 95 is the inevitable
expense in porting from front-end to front-end, not in accommodating 83=>95 
language changes.  Having lots of front ends out there is not 
necessarily a "good thing" given limited overall resources...

:   (4) Buy a traditional type of Ada product from the few vendors that
:       still exist (Rational?).

Heaven forbid ;-).

: Tom Robinson

-Tucker Taft   stt@inmet.com   http://www.inmet.com/~stt/
Intermetrics, Inc.  Cambridge, MA  USA