Re: Anybody in US using ADA ? New language competition? (long)

["Marin David Condic" <mcondic.auntie.spam@acm.org>]

I'd agree that you wouldn't want to try to get it to work on either only 1
compiler or on all compilers. Its not useful unless it really is portable to
some extent and obviously, not all compilers would be appropriate. (Would an
embedded compiler be able to utilize a GUI?) You'd have to pick a set of
compilers (and volunteers) who would (dare I use the word?) "Validate" that
the toolset worked in that environment. No small job, but I'd expect that it
wouldn't have to be *perfect* (bug-free) - just compilable and executable
for some major functionality. A bit more than a smoke test and significantly
less than a formal qualification test.

"Vendor Neutral" might be a bit harder. For some things (like a database?)
it would probably work, but for others (like a GUI?) it might be competing
against what a vendor already has. If you work to the least common
denominator (things that no vendor will object to as a competing product)
you probably will have a very small library. Not sure how to handle that one
other than to have SIGAda distribute a disk full of stuff and if it gains
acceptance as "Conventional Ada" then the vendors will ultimately have to
follow.

MDC
--
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jast.mil/

Send Replies To: m c o n d i c @ a c m . o r g

    "I'd trade it all for just a little more"
        --  Charles Montgomery Burns, [4F10]
======================================================================

Randy Brukardt <randy@rrsoftware.com> wrote in message
news:v1p2fu1h297j04@corp.supernews.com...
>
> I'm sure the vendors are in favor of someone doing that. But it will
> need to be vendor-neutral -- it does no good if most of the code only
> compiles with GNAT. Then vendors will say -- yet another hobbyest
> library for GNAT -- and ignore it.
>
> I'll reiterate that our experience with Claw is that GNAT is no better
> or worse than other Ada compilers for bugs. It allows some illegal code,
> won't compile some legal code, etc. In order to have something useful to
> end users, it really has to be the case that you can pick up the
> packages off of the SigAda web site and use them, no matter what Ada
> compiler you are using. For that to be the case, they have to have been
> compiled successfully with at least 3 different Ada compilers.
>
> Of course, some compilers may be just too buggy or limited to work with
> some software. I suspect just making the code work on GNAT, ObjectAda,
> and Rational Apex would be enough. You may not want to be restricted to
> the least common denominator code (like we did with Claw).
>
> Anyway, RRS would distribute such a library if it made sense, and making
> it work did not take up too many resources. After all, given:
>     10 bugs reported by customers;
>     10 bugs in new ACATS tests; and
>     10 bugs in a pseudo-standard library
> it's fairly likely that work on the library will be later in the list.
> :-)
>
>               Randy Brukardt
>
>
>