Re: GNAT license rational (was [ANNOUNCE] XML/Ada 0.5 released)

["Vincent Marciante" <marciant_remove@li.net>]

"Ted Dennison" <dennison@telepath.com> wrote in message
news:FuUH6.6682$SZ5.545769@www.newsranger.com...
> In article <3AEE0E5C.2ED1@li.net>, Vincent Marciante says...
> >What I'm really wondering is whether or not having
> >the exception in general, infrastructure-type Ada
> >code that is released is at all appropriate.  I
> >thought that knowing the exact rationale for the
> >GNAT would have been helpful.
>
> Well, I think I *could* take a stab at answering this. But, lest I bring
down
> the "wrath of Dewar" upon myself yet again over some minor point, perhaps
it
> would be better if I just pointed you to some old posts that were made 5
years
> ago when this matter was first being discussed.
>

<thread references snipped>

In 1995 Robert Dewar wrote:

http://groups.google.com/groups?q=+%22One+important+thing+to+remember+here%2
2+group:comp.lang.ads_maxm=5&as_maxy=2001&rnum=1&ic=1&selm=dewar.807152618%

in which he says:

>One important thing to remember here is that the whole idea of the LGPL
>is precisely to solve the problems of distributing proprietary programs
>built with a GCC compiler that incorporate runtime material. This is the
>only reason for existence of the LGPL!
>
>Whether the current wording is right to achieve this needs some
examination.
>It appears that there are approaches that are workable and completely
>consistent with the current wording. For example, if you put all the
>runtime library routines in a shared library and distribute executables
>that work with this shared library, which is certainly feasible in some
>systems, this seems one approach.
>
>On the other hand, this may not be adequate, in which case we need to
>reexamine the LGPL wording to make sure that the problems are solved
>for GNAT.
>
>We will address this issue in the documentation for the forthcoming
>version 3.0 of GNAT, and make sure that any problems are ironed out.

Which gives me the impression that he would think
that it is not appropriate to have the exception in
general, infrastructure-type Ada code whose usage
is completely up to the choice of a developer (usage
is not automatic due to choosing using some program,
as is the case with a program like GNAT that might
cause otherwise GPLd code to be utilized by the
developer's own code.)

Vincent Marciante

(To reply directly, please remove the underbar and
 the word that follows it from my address.)