Re: Proving Correctness (was Java Portability)

["nicolas" <n.brunot@cadwin.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 5627 bytes --]

"Pascal Obry" <p.obry@wanadoo.fr> a �crit dans le message news:
ulmkxuy80.fsf@wanadoo.fr...
> Sorry but I'm very upset by your mail. If nobody does the job the job wont
> gets done that it !
I'll start to reply your last remark.
There is absolutely no attack about this XML library. It's great to have it
for students and Gnat users.

But we were talking about standard and portable libraries designed to be
independant of compilers.
This is what Ada is designed for. This is about what Ada users are giving
lessons to everybody.
George gave me this XML library as an example to show that this type of
library is available.

This is not the case.
We don't use Gnat because Gnat doesn't satisfy our requirements. That's our
problem.
That doesn't mean Gnat is a bad product, no product is perfect, and no
product can satisfy everybody.
Gnat, especially public version, is great for students, for anybody who
wants to try Ada, and certainly for professionals who made Gnat choice.
A lot of people are using others compilers than Gnat, a lot of people work
with Windows.
You definitely cannot take this XML library example to show that standard
portable libraries are available for Ada users.

> No. Just grab the library and changes it to be whatever-compiler
> compatible. You have the source, use them...

No, I shouldn't have to do that,
Anyway, Gnat.spitbol is not available for other compilers, and I don't have
to spend my time retrieving it and trying to adapt it to my compiler.
It's ok for Ada fans ready to spend more time on their favourite language
than the time they would spend on a language they don't like.
This is not ok for people considering languages as tools, who choose the
most convenient one for their needs.
It's a nonsense to say to Ada programmers
"I made my own specific library, adapt it yourself if you don't want to use
my compiler"
This is exactly what people who choose Ada don't want to hear, and they are
right about that.

> Right. So ask you vendor to adapt an Open Source component or sell an
> equivalent component.

All the thread is about the fact that this is not an option and goes against
Ada promotion
Ada is designed to write code as much as possible independant from platforms
and compilers.
An XML library has nothing specific to a platform or a compiler.
You can develop one with code totally independant of the compiler, and there
is no reason to do otherwise.

You may ask why we didn't develop that ourseleves :  it is for very simple
reasons.
- We are not paid for that
- Java does it much better.
- When we are not working, we are more interested in others things in life
than playing with computers

> I don't see a problem with Aonix, Rational taking the XML library, adapt
it to
> their compiler and release the result as an Open Source add-on for their
> compiler.

It's clear to me that a standard library shouldn't have to be adapted to
each compiler.
We have more than 1 million lines of Ada code and are very careful to have
them compiler independant.
That's the least you can expect from an Ada library.
Ada language is designed to be platform and compiler independant.
If Ada users starts to release such basic libraries specific to one compiler
or one platform, the only thing we do is to prove to others languages users
that we are unable to apply what we are trying to promote.

> > Sax relies on Gnat specific packages.
> So why ? (see above). XML has been done by an Ada fan working for ACT.
That's
> not a big surprise. I don't understand why you are so upset ?

I am not upset, Georges told me I should be satisfied with this XML library.
I am not satisfied because we don't use Gnat for our software releases, and
because libraries tied to one compiler without any valid reason, go against
Ada promotion.
The only consequence is that I don't use it.
No big deal ...

> And BTW, it is no fair to be upset by something that has been done. Why
are
> you not upset by the fact that these components have not been ported by
you
> Ada vendor for their compilers ?

Once again I'm not upset about anything
I strongly believe that the lack of standard available library you can use
'as it is' without any modification, whatever compiler you use, goes against
Ada rationale ...

> Possible, but not something that should be done by ACT for obvious reason
:)

I don't care who should do it .... I'm a user and like any other user I
choose the best tool available for my needs.
If ACT does everything to force Ada world to be dependant of Gnat compiler,
it is ACT's decision.
But I think it could kill Ada, and therefore ACT ...

> Let me add that we have tried in AWS to be compiler independant (see
> AWS.OS_Lib) but since 2 or 3 releases this is not true anymore. I don't
have
> the time to try on all available compilers, the GNAT library is so nice
that I
> just can't bypass it (GNAT.Regexp for example). So yes now AWS is mostly
> working with GNAT... But I'd love to see others porting some stuff to be
able
> to avoid this dependences :)

I think it's a dead-end if everybody does his own job specific for his own
tools, and wait for others to adapt it for others tools.
This shows all the limitations of the process.
We are giving to others languages users, all what they need, to be convinced
that Ada users are themselves unable to stick to what they promote.
If you can't bypass GNAT.Regexp, obviously GNAT.Regexp should be in the set
of standard libraries, it shouldn't be named GNAT, and should be available
as it is for any compiler.
Once again Ada users give strong lessons to the rest of the world, and do
exactly the opposite in practice ...