Re: Ada student homework ?

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

In article <s3_R5.442083$i5.7683711@news1.frmt1.sfba.home.com>,
  tmoran@acm.org wrote:
>
>   Is there?  Not "anyone can ...", or "... has been at least
> fiddling with GtkAda,", but "Is there a release sequence of
> GtkAda versions that have been developed and tested with
> multiple Ada compilers?"

No idea! We are not planning to make money selling GtkAda
for use with Rational etc compilers, so we are not in the
business of providing releases for other vendors compilers.

> CLAW users have full source code and there is certainly no
> more impediment to them making it work on some new fifth or
> sixth Ada compiler than to users making GtkAda work on a
> second or third or fourth existing or new Ada compiler.

Oh but there is! And you should not lose sight of it. There is
nothing to stop Rational or Aonix from selling and supporting
a version of GtkAda that works with their compilers. Far from
preventing it, we are happy to encourage it, and as I mentioned
we have responded to a few minor inputs from Rational on some
minor incompatibilities that were easily corrected.

With CLAW on the other hand, unless I misunderstand the license,
it would be a copyright and license violation for a third
party to produce this fifth port and sell it themselves. If
the CLAW license DOES allow free creation, distribution, and
sale-for-profit of deriviative versions, then it is indeed
Free Software in the normal sense, and it is a bad example
for my comparison.

> In fact, if a new Ada compiler does not show any bugs or
> weirdnesses that need workarounds, porting CLAW may need no
> changes at all!  It is true that we have taken it
> upon ourselves to do that work for CLAW

Presumably because you intend to make money selling it for
these other compilers, seeing as you have exclusive rights
to do so.

> I'm happy to hear that ACT has "been happy to cooperate with
> them [Rational] on this [porting GtkAda]."

As I said, we have no objection at all to other vendors using
parts of our technology that are complimentary. Indeed there
are quite a few users using mixtures of GNAT and Rational
technology, why not?

> To return to t previous topic, I trust we agree that
> GtkAda's portability across OSes sometimes, but not always,
> exceeds its disadvantages in missing functionality under
> Windows and its portability problems across Ada compiler
> vendors.

It's not a "disadvantage" of GtkAda that it does not take
advantage of some special Windows features, it is a critical
part of its design that it be portable across systems.

As to portability problems, I have no knowledge that there are
more problems in GtkAda than in CLAW. I know that you have
encountered quite a few portability problems with CLAW which
required new versions of compilers etc that fixed bugs (I
know we have fixed a couple of bugs that CLAW showed up,
since it uses quite complex features of Ada). In the case
of GtkAda, so far, we have not encountered any significant
portability problems, and those we have encountered were
fixed by trivial changes in the source. For example, GNAT
was allowing pragma Preelaborate in a package body, and there
were a couple of such instances in the GtkAda sources -- so
we simply made trivial changes to the sources (and fixed the
GNAT bug that allowed these wrongly placed pragmas, a bug
which interestingly had never been noticed before :-)
Certainly it would be our intention to make any such
necessary corrections if and when they prove to be
needed. One advantage of trying to keep the sources as
simple as possible, and therefore easily portable across
compilers is that it makes it easier to use GtkAda with
various different versions of GNAT.


Sent via Deja.com http://www.deja.com/
Before you buy.