Re: [ANNOUNCE] XML/Ada 0.5 released

[Emmanuel Briot <briot@gnat.com>]

[Posted on behalf of Robert Dewar]



First, a general note, the direct posts you saw from me in this
thread are not posts at all, but the results of some very strange
cc mechanism in what were intended to be personal replies to David
Botton. I apologize for these, they were definitely worded as private
mail to Dave, and not for general consumption (in particular I
certainly would not have posted the comment about RR, and I want
to make it clear that I appreciate RR's efforts in the Ada area,
indeed they are a technical partner of ours, and we point customers
in their direction, but we certainly do not regard them as a competitive
force, quite the contrary, we regard them as partners).

None of these posts should be regarded as official comments from ACT, they
are simply off hand comments to David. I will refrain from answering
any of David's email in the future to prevent this strange occurrence.
It never occured to me that a cc could be used in this way to create
a direct post that looks like it was deliberately posted as an
article.

  > I am glad to see you back, Mr Dewar, I hope it's not just a temporary
  > thing.

As I explained, any posts that appear directly from me are entirely
accidental, and most certainly that phenomenon is temporary. I do not
intend to post my personal email here, and once again, apologize for
doing so (email is a tricky area, with all sorts of traps like this :-)

I do watch some threads, and will post indirectly from time to time
when it is appropriate to do so.

-----------------------------------------

Second, a couple of specific comments (these ARE intended as posts).
In general from now on, only posts coming indirectly from someone
else at ACT are intentional posts :-)

  Florian said

  > But what will you do if someone contributes code under the GPL?  Or
  > won't you accept such contributions?  (Maybe that's a highly
  > theoretical issue at this point, but you never know...)

I must say I am completely puzzled by this comment. All the code in
question IS under the GPL, so certainly people can contribute GPL'ed
code. You have to explain what on earth you mean here :-) Furthermore
people have contributed a lot of GPL'ed code in the past, as well as
GMGPL'ed code where appropriate, so there is nothing theoretical about
such a possibility.

  Ted said

  > Since a copy was cc-ed to me, I too was wondering where it was. I'm glad to see
  > it finally appear, as it has some good info in it (despite the personal
  > attacks).

Please do not regard it as a personal attack if I correct plain inaccurate
information, and if I suggest more careful checking up on information. It
really doesn't help if incorrect stuff is posted. if you are relying on
your memory, check first :-)

  > Now perhaps you can see why I got so defensive all of a sudden. I'll leave it up
  > to everyone else's judgement whether what I said was reasonably close to the
  > truth, based on the old messages I referenced. What started that thread that
  > contained the RMS suggestion to use a modified GPL is unfortunately lost to the
  > mists of time, but my memory of it was that someone piped up about their lawyers
  > refusing to allow use of GNAT due to the GPL/LGPL issue. 

No, nothing is lost in this mists of time, and the facts are quite clear, 
despite Ted's memory to the contrary. Here are the historical facts (if
you like, go back and check the document trail, you will find this all
documented).

1. The original GNAT contract required all software to be released under
the GPL or LGPL and the copyright assigned to the FSF. The idea was to
release the runtime under the LGPL. This was a direct response to my
suggestion of what the contract should say.

2. On examination, we became concerned that the LGPL was not the right vehicle
because of two factors

  a) the annoyance of distributing objects
  b) the issue of generics

So I created the GMGPL very early on, and we used it from the very beginning
for all GNAT sources (neither the GPL nor the LGPL was ever used for any
GNAT runtime sources). We then discussed with Stallman to ask if he had
any objection to the change.

He was confused at first, and thought we were suggesting using the *GPL* for
the runtime library, and he encouraged us to look to the C model and use a 
modified non-restrictive version of the GPL, but that was what we were already
doing in any case.

So once again, no, there was no one who "piped up about their lawyer
refusing to allow use of GNAT" that had any influence on this decision
which was made long before any lawyer had a chance to pipe up :-)

Sure, there were lawyers who piped up, and we went through many occasions
on which we had to demonstrate to lawyers that our GMGPL license for the 
runtime meant that they could use the system without concerns. Now days,
we provide a formal license agreement to our customers that clarifies all
issues. In the case of non-customers, there really is no clear legal license
agreement, and indeed it is a bit unclear in court what the status would be
for users of the public version (in other words, courts would have to decide
whether the fact that you had published a statement that something was
distributed under the GPL constituted receiving specific permission for
copying, we hope it would, but we do not know till it is litigated).
For customers, the situation is like dealing with any other company,
indeed it is like dealing with Microsoft, except our license is a bit
more liberal than theirs :-)

(by the way, I comment on Microsoft's latest outburst that they have a lot
of chutzpah to complain that they don't like the license we choose to give
because it means they cannot abscond with *our* intellectual property :-)

But back to piping up lawyers ... as I said in my previous message, we knew
perfectly well that using the GPL for everything would have caused trouble
(not the least of which is that it would have been non-responsive to the
DoD contract which required the LGPL to be used for the runtime :-)
Neither the decision to use the LGPL, nor the subsequent decision to
replace it with the GMGPL had anything to do with external lawyers. Indeed
at that time, we knew more about how GPL licensing works than virtually
all IPR lawyers (now of course Microsoft, Redhat, WRS, IBM etc have large
armies of lawyers who have studied these licenses VERY carefully).

(another side comment on Microsoft is that of course when they try to spread
FUD about open source software in general, and the GPL in particular, what
they are really trying to do is to stop people using the stuff. It must be
very frustrating for Microsoft that not only is there important software
that they don't own it, but they can't even buy it, no matter how much
money they have :-)