GtkAda License Question

GtkAda License Question

[M E Leypold]

(Oh no, not a licensing question again).

In the new GNAT-2006 release I found GtkAda labeled as
GTKAda-GPL. Since, if that were so, it would hinder the use of GTkAda
under GMPL-conditions with gccada, I started (as a first step) to look
into the licensing of the older GTKAda 2.4.1 and now I'm confused:

 - Most (if not all) of the files which are later linked to the
   executable seem to carry the GMPL linking exception.

 - Neither COPYING nor README mention the linking exception.

Does anybody here have any idea how the licensing situation of GtkAda
actually is? Are the copyright header with linking exception are only
remnants of (happy :-) days long past or is GtkAda actually licensed
with a linking exceaption (in version 2.4.1) and it just has been
forgotten in the README to mention that? And what about the new
versions from the CVS (i.e. 2.8.0)?

Regards -- Markus

Re: GtkAda License Question

[Ludovic Brenta]

M E Leypold writes:
> Does anybody here have any idea how the licensing situation of GtkAda
> actually is? Are the copyright header with linking exception are only
> remnants of (happy :-) days long past or is GtkAda actually licensed
> with a linking exceaption (in version 2.4.1) and it just has been
> forgotten in the README to mention that? And what about the new
> versions from the CVS (i.e. 2.8.0)?

The only authoritative source is AdaCore, so when in doubt, ask them.
Robert Dewar clearly said during FOSDEM that the headers, the COPYING
and the README files have no legal force whatsoever; users are always
required by law to ascertain the licensing terms with the licensor. The
presence of the GMGPL in the headers allows you to presume that these
licensing terms are the ones to use; IANAL so I don't know if this
argument can convince a court of law.

Now, I think it is reasonable to presume that:
- GtkAda GPL edition (both version 2.4.1 and version "gps-3.1.3" are
included) is pure GPL
- GtkAda 2.4.0 from libre is GMGPL
- GtkAda from CVS is GMGPL (the headers have the exception.)

In Debian, all libraries are GMGPL, except ASIS 2006 which will soon
arrive in Etch under pure GPL. Use Debian. Debian is good for your
mental health :)

-- 
Ludovic Brenta.

Re: GtkAda License Question

[M E Leypold]

"Ludovic Brenta" <ludovic@ludovic-brenta.org> writes:

> M E Leypold writes:
> > Does anybody here have any idea how the licensing situation of GtkAda
> > actually is? Are the copyright header with linking exception are only
> > remnants of (happy :-) days long past or is GtkAda actually licensed
> > with a linking exceaption (in version 2.4.1) and it just has been
> > forgotten in the README to mention that? And what about the new
> > versions from the CVS (i.e. 2.8.0)?
> 
> The only authoritative source is AdaCore, so when in doubt, ask them.
> Robert Dewar clearly said during FOSDEM that the headers, the COPYING
> and the README files have no legal force whatsoever; users are always
> required by law to ascertain the licensing terms with the licensor. The

Which is obvious nonsense, since that would render all GPL licensing
as practiced presently invalid. Also I'd like to ask "which law"? US
law? International copyright law? 

But I can see why Robert Dewar would prefer that reading of the
situation. That would also explain why there is no independent other
site offering current libre-Downloads (like GtkAda, GNAT GPL).

> presence of the GMGPL in the headers allows you to presume that these
> licensing terms are the ones to use; IANAL so I don't know if this
> argument can convince a court of law.

Interesting. Probably all license files in my software package are not
valid then ...

> Now, I think it is reasonable to presume that:

> - GtkAda GPL edition (both version 2.4.1 and version "gps-3.1.3" are
> included) is pure GPL

> - GtkAda 2.4.0 from libre is GMGPL

> - GtkAda from CVS is GMGPL (the headers have the exception.)

Exactly this is where the problem starts. The CVS has GtkAda 2.8.0,
but I won't build it if it is GPL-only. If the binary in the GPL GNAT
bundle is based on identical sources, why would it be under GPL? If
the sources are different, then either the CVS source builds better
than the source on the GPL GNAT bundle or the other way round. Both
situations would incite further questions.

> In Debian, all libraries are GMGPL, except ASIS 2006 which will soon
> arrive in Etch under pure GPL. Use Debian. Debian is good for your
> mental health :)

But only as long as the the upstream licensing doesn't change -- oh so
suddenly -- to GPL. Then all us small developers of graphical gadgets
have either to shell out > $15000 / year just for maintaining our
smallish projects or we pull maintenance very suddenly.

What joy.

Regards -- Markus

Re: GtkAda License Question

[Ludovic Brenta]

M E Leypold writes:
>> - GtkAda GPL edition (both version 2.4.1 and version "gps-3.1.3" are
>> included) is pure GPL
>
>> - GtkAda 2.4.0 from libre is GMGPL
>
>> - GtkAda from CVS is GMGPL (the headers have the exception.)
>
> Exactly this is where the problem starts. The CVS has GtkAda 2.8.0,
> but I won't build it if it is GPL-only. If the binary in the GPL GNAT
> bundle is based on identical sources, why would it be under GPL? If
> the sources are different, then either the CVS source builds better
> than the source on the GPL GNAT bundle or the other way round. Both
> situations would incite further questions.

When producing a GPL version, AdaCore simply remove the "exception"
paragraph from every file using a script before packaging the sources.
No magic, and no difference in the actual software.

>> In Debian, all libraries are GMGPL, except ASIS 2006 which will soon
>> arrive in Etch under pure GPL. Use Debian. Debian is good for your
>> mental health :)
>
> But only as long as the the upstream licensing doesn't change -- oh so
> suddenly -- to GPL. Then all us small developers of graphical gadgets
> have either to shell out > $15000 / year just for maintaining our
> smallish projects or we pull maintenance very suddenly.
>
> What joy.

That would be a good reason for a fork, then.  If sufficiently many
people are interested in the maintenance of these tools, it is a
realistic proposition.  But let's not think of a fork before it is
actually necessary.

-- 
Ludovic Brenta.

Re: GtkAda License Question

[Jean-Pierre Rosen]

M E Leypold a �crit :
>> The only authoritative source is AdaCore, so when in doubt, ask them.
>> Robert Dewar clearly said during FOSDEM that the headers, the COPYING
>> and the README files have no legal force whatsoever; users are always
>> required by law to ascertain the licensing terms with the licensor. The
> 
> Which is obvious nonsense, since that would render all GPL licensing
> as practiced presently invalid. Also I'd like to ask "which law"? US
> law? International copyright law? 
> 
Hmmm, before making such a statement, you should know that Robert Dewar 
*is* a lawyer, i.e. he is a registered expert to the court about matters 
of software copyright.

-- 
---------------------------------------------------------
            J-P. Rosen (rosen@adalog.fr)
Visit Adalog's web site at http://www.adalog.fr

Re: GtkAda License Question

[M E Leypold]

Jean-Pierre Rosen <rosen@adalog.fr> writes:

> M E Leypold a �crit :
> >> The only authoritative source is AdaCore, so when in doubt, ask them.
> >> Robert Dewar clearly said during FOSDEM that the headers, the COPYING
> >> and the README files have no legal force whatsoever; users are always
> >> required by law to ascertain the licensing terms with the licensor. The
> > Which is obvious nonsense, since that would render all GPL licensing
> > as practiced presently invalid. Also I'd like to ask "which law"? US
> > law? International copyright law?

> Hmmm, before making such a statement, you should know that Robert
> Dewar *is* a lawyer, i.e. he is a registered expert to the court about
> matters of software copyright.

So all current practice of distributing GPL software is flawed insofar
as the COPYYING and README files don't establish a binding agreement
between the receiver and the author of the software? So the receiver
has actually find the author (or copyright holder) to ascertain the
licensing terms and if he finds him, it might be that the copyright
holder decides that today there are different licensing terms for this
applicant?

I seriously hope that is not what Robert Dewar meant. Indeed I think
he has been quoted out of context (thanks to Ludovic anyway for the
quote!) and one has to see how things pan out with ACT when asking for
clarification on that issues.

Let me add, that the process AT&T vs. BSDI actually seemed to attach
some significance to the headers and even if the law has changed since
(it has, at least concerning the defaults for files w/o headers) I
seriously doubt they suddenly on't have any significance at all.

Regards -- Markus

Re: GtkAda License Question

[Jean-Pierre Rosen]

M E Leypold a �crit :
> So all current practice of distributing GPL software is flawed insofar
> as the COPYYING and README files don't establish a binding agreement
> between the receiver and the author of the software? So the receiver
> has actually find the author (or copyright holder) to ascertain the
> licensing terms and if he finds him, it might be that the copyright
> holder decides that today there are different licensing terms for this
> applicant?
> 
COPYING and README are not *proofs*, that's what he meant. Anybody can 
take propriatory software and redistribute it with a COPYING file (or 
change the headers); this won't make the software free. What Robert said 
is that if you want legal assurance that the software is free, you need 
a signed letter from the author.
-- 
---------------------------------------------------------
            J-P. Rosen (rosen@adalog.fr)
Visit Adalog's web site at http://www.adalog.fr

Re: GtkAda License Question

[Dmitry A. Kazakov]

On Wed, 21 Jun 2006 19:11:52 +0200, Jean-Pierre Rosen wrote:

> M E Leypold a �crit :
>> So all current practice of distributing GPL software is flawed insofar
>> as the COPYYING and README files don't establish a binding agreement
>> between the receiver and the author of the software? So the receiver
>> has actually find the author (or copyright holder) to ascertain the
>> licensing terms and if he finds him, it might be that the copyright
>> holder decides that today there are different licensing terms for this
>> applicant?
>> 
> COPYING and README are not *proofs*, that's what he meant. Anybody can 
> take propriatory software and redistribute it with a COPYING file (or 
> change the headers); this won't make the software free. What Robert said 
> is that if you want legal assurance that the software is free, you need 
> a signed letter from the author.

In my admittedly ignorant eyes, that looks like a different case. After
all, anyone can forge a signature on the letter. But the question is, let
Markus correct me, whether the COPYING _as distributed by the author_ is a
proof in the court. For example, may the author withdraw authentic COPYING.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: GtkAda License Question

[Georg Bauhaus]

On Wed, 2006-06-21 at 20:23 +0200, Dmitry A. Kazakov wrote:


>  But the question is, let
> Markus correct me, whether the COPYING _as distributed by the author_ is a
> proof in the court. For example, may the author withdraw authentic COPYING.

This might shift the question to what distribution means and implies.
I think that usually the law is about contracts. There are culturally
different ways of establishing a contract, there is good
faith etc., but these are legal matters ... IANAL

Re: GtkAda License Question

[Georg Bauhaus]

On Tue, 2006-06-20 at 10:46 +0200, Jean-Pierre Rosen wrote:
> M E Leypold a écrit :
> >> users are always
> >> required by law to ascertain the licensing terms with the licensor. The
> > 
> > Which is obvious nonsense, since that would render all GPL licensing
> > as practiced presently invalid. Also I'd like to ask "which law"? US
> > law? International copyright law? 
> > 
> Hmmm, before making such a statement, you should know that Robert Dewar 
> *is* a lawyer, i.e. he is a registered expert to the court about matters 
> of software copyright.

Also notice that every time you install just about any software,
then in Europe at least,
you must explicitly (and knowingly) confirm that you agree to the
licensing terms.
I believe it is a requirement of EU contractual procedures that
you must check corresponding boxes. (Which again is probably not
enough for such and such, etc...)
(Opposed to this requirement, I remember arguments that the
sentence "by opening this box you agree to ... contained therein"
is irrelevant.)
IANAL.


Georg 

Re: GtkAda License Question

[Frank J. Lhota]

Georg Bauhaus wrote:
> Also notice that every time you install just about any software,
> then in Europe at least,
> you must explicitly (and knowingly) confirm that you agree to the
> licensing terms.
> I believe it is a requirement of EU contractual procedures that
> you must check corresponding boxes. (Which again is probably not
> enough for such and such, etc...)
> (Opposed to this requirement, I remember arguments that the
> sentence "by opening this box you agree to ... contained therein"
> is irrelevant.)
> IANAL.

This is also true in the U.S.: the installation of every non-free 
software package seems to require the acceptance of an End-User License 
Agreement (EULA).

Re: GtkAda License Question

[M E Leypold]

Georg Bauhaus <bauhaus@futureapps.de> writes:

> On Tue, 2006-06-20 at 10:46 +0200, Jean-Pierre Rosen wrote:
> > M E Leypold a �crit :
> > >> users are always
> > >> required by law to ascertain the licensing terms with the licensor. The
> > > 
> > > Which is obvious nonsense, since that would render all GPL licensing
> > > as practiced presently invalid. Also I'd like to ask "which law"? US
> > > law? International copyright law? 
> > > 
> > Hmmm, before making such a statement, you should know that Robert Dewar 
> > *is* a lawyer, i.e. he is a registered expert to the court about matters 
> > of software copyright.
> 
> Also notice that every time you install just about any software,
> then in Europe at least,
> you must explicitly (and knowingly) confirm that you agree to the
> licensing terms.

I'd like to offer 2 observations here (but IANAL):

 (1) There were various companies which didn't think the GPL binding
     for exactly that reason (they hadn't clicked any boxes ...,
     meaning, they hadn't agreed explicitly to anything). Their vie
     didn't hold up in the courts. (It was somthing about firewalling,
     Harald Welte was the owner of the code in question).

 (2) According to the FSF the GPL is no licensing _contract_ but a
     "grant of rights". The other side doesn't need to agree, since
     the default would be "no rights". (Of course Dan Bernstein of
     QMail fame disagrees here).

     This has no impact on Robert Dewar's quote. He holds that
     whatever rights are granted with the file COPYING or in the file
     headers, it's not binding and there are no rights given, you
     "have to check with the licensor". I doubt that.

 (3) If we stay with the contractual model. In German law (which is no
     case law like American or UK law), there is the concept of
     "konkludentes Handeln" which I would roughly translate as
     "implied consent". I have not seen it applied to the GPL
     situation (yet), but the idea is, that a contract manifests
     itself even in simple transactions at the supermarket checkout,
     since whereas both sides probably don't say anything like "I
     agree that I take the money and you get the goods" the
     circumstances and the actions speak for themselves: One gives
     moneay, the other takes it and that is what you do in a super
     market: buying goods.

     Applying that to the usual way of distributing GPL software: Just
     unpacking the sources, reading COPYING and then using the source
     to generate an executable would be "implied consent".

     "Well, Mr. Judge, I read, that they thought that the code should
     be used under the GPL, yes, but I thought it somehow would not
     apply to me or was just any old COPYING files". Rather
     believable.

> I believe it is a requirement of EU contractual procedures that
> you must check corresponding boxes. (Which again is probably not
> enough for such and such, etc...)

I don't think so. There have been some attempts to make it so (EU law
is such a nice argument if you want to twist national law into
something else ...), but mostly with respect to default warranties etc
which can only be restricgted by explicit agreement. That interacts
with the GPL but it's NOT generall required for the user to check
boxes in Europe to get rights under the GPL. I live there.

But IANAL: Don't construe that as legal advice. I'm not a lawyer.

> (Opposed to this requirement, I remember arguments that the
> sentence "by opening this box you agree to ... contained therein"
> is irrelevant.)

That is "licensing". And since you bought the box, there is already a
contract, which cannot be restricted again by a text in the box. Other
situation.


> IANAL.

IANAL2 (I'm not a lawer too). Considering how f***ed up by legal pits
and traps the internet has become, this propably should become the new
standard greeting in forums and usenet. "IANAL!". "IANAL2!". 

Regards -- Markus

Re: GtkAda License Question

[M E Leypold]

Excuse all the typos in my last post. -- Regards -- Markus

Re: GtkAda License Question

[Björn Persson]

Ludovic Brenta wrote:
> Robert Dewar clearly said during FOSDEM that the headers, the COPYING
> and the README files have no legal force whatsoever; users are always
> required by law to ascertain the licensing terms with the licensor.

Hmmmmm? Then the typical "you don't even have the right to sneeze" EULA
wouldn't have any legal force either. Am I supposed to travel to Redmond
and get Bill's hand-written signature on a contract each time I buy a
computer with Windows pre-installed?

-- 
Bj�rn Persson                              PGP key A88682FD
                    omb jor ers @sv ge.
                    r o.b n.p son eri nu

Re: GtkAda License Question

[gshapovalov]

Björn Persson wrote:
> Hmmmmm? Then the typical "you don't even have the right to sneeze" EULA
> wouldn't have any legal force either.

Well, accordingly to what I have read on the subject they do not. In
fact it usually is worded like this: EULA(s) will not hold up in the
court, should the case be raised. And I seem to remember the claims
that this was indeed already demonstrated on some occasions. However I
am skimpy on the details, sorry.
And, as usual - do not take anything I am saying about legalities
seriously - check everything you need to rely on yourself..

George

Re: GtkAda License Question

[gshapovalov]

Ludovic Brenta wrote:
> Now, I think it is reasonable to presume that:
No, it is not, you *should* contact AdaCore and ask, see below.

> - GtkAda GPL edition (both version 2.4.1 and version "gps-3.1.3" are
> included) is pure GPL
> - GtkAda 2.4.0 from libre is GMGPL
No, it is not. When repackaging it for Gentoo I actually raised the
issue with AdaCore (exactly because of that contradiction of headers
having the GM excerpt but claiming pure GPL in some file, *not*
COPYING, which was missing). The answer was that it is pure GPL, as
everything distributed off libre2 site. Although, if you got it from
the original libre site (is it still functional? It seems to redirect
me to libre2 nowadays) you might still have the GMGPL version, although
I doubt there is any actual difference. They don't seem to care much
about packaging stuff cleanly..

Of course you should not take my words for anything and coact them
yourself. You are held legally responsible for any action taken on my
advice ;).
The added benefit of more people contacting them is that this may
finally force them to update the documentation to reflect the present
situation more clearly..

> - GtkAda from CVS is GMGPL (the headers have the exception.)
Again and for the same reason, I would not presume that, or in fact
anything. With them it seems it is obligatory to ask..

> In Debian, all libraries are GMGPL, except ASIS 2006 which will soon
> arrive in Etch under pure GPL. Use Debian. Debian is good for your
> mental health :)
Um, in the view of the above I am not sure you can distribute them as
GMGPL ;).

George