Re: Do we need "Mission-Critical" software? Was: What to Do?

[Robert Dewar <robert_dewar@my-deja.com>]

In article <umq1yuixba0.fsf@maestro.clustra.com>,
  Ole-Hjalmar Kristensen <ohk@clustra.com> wrote:
> Btw., c2ada shows that the automatic generation of thin
> bindings is indeed feasible, but the c2ada implementation
> lacks a lot in terms of stability.

Yes, interestingly there seems to be very little demand for
working on this kind of tool. At one point Ada Core
Technologies had a fairly high priority project to adapt the
SGI binding generator (which is far more advanced and robust
than c2ada, and also handles C++) for general use (it currently
uses a proprietary C++ front end), but there simply has not
been any significant demand for this tool, so the priority
of the project got reduced, and many more important tasks
are ahead of it at this stage.

> It seems to me that your position is that lack of bindings to
> foreign libraries is not a problem, since you do not see any
> demand for tools. I hope you are right, but I do not believe
> so.


It is a problem, but I don't think it is a blocking problem to
the use of Ada in most cases.

Through the history of Ada we have had people saying

"If only we did X, then Ada would be more widely used"

But most of these claims have proved experimentally false, and
I think this is another such case. If someone produced a much
improved c2ada, I doubt this would suddenly make Ada popular.

Indeed, most of the small scale development of which N.Brunot
talks about is done in Windows environments, and here the
issue is not translation of C headers, it is automatic binding
to type libraries. I think the gnatcom approach is FAR more
interesting, valuable, and promising in this context, since
it allows for automatic reliable generation of thick Ada
bindings from language independent specifications (it is
quite fortunate that Visual Basic is so much more widely
used than C, since it means that Microsoft has to take a
language independent approach, and Ada is quite happy to
take a place at the table).


Sent via Deja.com
http://www.deja.com/