RE: [Axiom-legal] Licensing Aldor (was: GPL vs. modified BSD)

[C Y]

--- Bill Page <address@hidden> wrote:

> > I can readily accept that the legal aspects of this are not
> > simple to work through.
> 
> Could you give an example of such a "legal aspect"?

Well, I don't know the company policies, contracts, or other issues
specific to NAG so anything I say would be no more than speculation. 
However, things I can see as being possibly relevant would be business
agreements licensing source to include in the compiler, agreements in
the past not to release some of the code, trade secret concerns,
liability concerns, and probably others I'm forgetting.  Axiom itself
would suggest the possibility that business agreements with IBM might
be relevant.

Are any of these a risk?  We don't know - insufficient information. 
But before proceeding I can see it as being quite reasonable that such
issues need to be thoroughly checked.  Remember, there is no revenue
stream here to help pay for any sort of legal activity.  IF lawyers are
being consulted (who knows?) then a time frame of a year wouldn't
surprise me in the least, particularly when the issue at hand is not
critical to the revenue stream.
 
> > Mike Dewar's update that there is still work in progress
> > is very encouraging to me.
> 
> Yes, to me too. But the delay is really deadly. In my opinion we
> (the Axiom project) would have been much better off if we had
> known two or three years ago whether or not Aldor was going to
> be part of Axiom or not.

Oh, no argument there.

> If the answer had been "no", we would be
> struggling but at least work would now be well underway to improve
> SPAD and to write new algebra code. If the answer had been "yes"
> then the work to re-write the Axiom algebra library would likely
> be near completion and many people would be involved in writing
> new algebra code.

Well, we would be further along anyway.  I think a truly literate
re-write of the Axiom algebra library (e.g. make each pamphlet into a
publication worthy summary/paper/history as well as code) would be the
work of decades, but that might be just me overestimating the effort
involved (even if it is that long of a project, I still view it as
worth doing).  I still think a lot of that work is probably doable even
without knowing the final language (paper writing can precede code, and
perhaps it should anyway in order for the author to have the knowledge
to write the code well) but the uncertainty over project direction does
tend to have a harmful impact.

> But because there has been no decision for so
> long, the result is that nearly everything concerning the algebra
> is "on hold" and otherwise enthusiastic people (such as those
> developing Sage right now) don't want to touch Axiom because of
> the overwhelming amount of legacy code and uncertainty about
> Aldor - the only interesting part to many people with a computer
> science orientation. (Gaby being a fortunate exception :-)

I think the legacy code is its own problem, Aldor or not.  There is
lots of really smart code and design ideas in Axiom - the problem is
recognizing it and separating that from things like that hideous lisp
packages nightmare.

> I think the apparent preoccupation with licenses and legal opinions
> of people working in open source is only in response to challenges
> to this freedom by commercial proprietary interests.

Sure, but the legal system is the ultimate trump card - they are the
only ones who can say "stop doing that" and back it up.  Naturally such
power is treated with respect.
 
> > Morally, re-licensing the code willy-nilly would be about the
> > same as taking GPL code and incorporating it into a commercial
> > product.
> 
> There is nothing legal about using GPL code in a commercial product.
> Of course it carries the obligation that the source code of the
> commercial product must be open source, but contrary to most
> "corporate expectations", open source code has not proven to be
> a detriment - quite the contrary. "Commercial" should not be
> confused with proprietary or "closed". For example consider Red
> Hat Enterprise Linux, Sun's StarOffice (aka. OpenOffice) and IBM's
> webserver products based on Apache. This is likely to be the norm
> in the future rather than the exception.

My mistake - make it closed source and commercial.  The situation I had
in mind was not commercial open source software, but software that
wishes to remain closed incorporating GPL code and selling it without
releasing back the changes.  This has been done before - I think PearPC
had some problems with that in the past (CherryOS?), to name one
example.

> > Even if the original authors had said that might be acceptable
> > under the right conditions, doing it without their explicit
> > approval would be a Bad Idea.  Aldor.org and NAG MUST be
> > allowed to do as they please with their work, the same as
> > we are allowed to do what we please with our own.
> 
> Well, yes copyright law does require the permission of copyright
> holders. Violating the wishes of the copyright holders could
> leave one open to a law suit but Tim has already explained why
> such action is very unlikely.

I think releasing code that has not been released under any license
would provoke a MUCH stronger response than subtle issues about Aldor
Public License vs. GPL.  Nor would I fault such a response - while
questions about APL vs. GPL are in good faith, releasing unreleased
code is an unambiguous affront to the copyright holder.

> > Any violation of that trust and respect is very dangerous to the
> > community, to say nothing of relations with NAG (who after all
> > didn't have to do any of this in the first place.)
> 
> There is a risk I suppose that, since NAG retains the copyright on
> Axiom (see Axiom license), that they might choose to re-license
> Axiom at some point in the future. But that would not retroactively
> affect any versions of the Axiom that had already been released
> under the existing license.

Well, look at it this way then - if the Aldor source were to become
available under circumstances where it was not clear NAG approved, no
one would develop it or use it.  Nor would they release it through
legitimate avenues, effectively killing ANY chance for a truly free
Aldor compiler short of a rewrite.  ReactOS had some problems with the
means people were using, and had to do a full source tree audit.

> > > > Tim is correct that only lawyers can render really useful
> > > > opinions on these issues, but that doesn't mean we don't need
> > > > to make a good faith effort to respect the license to the best
> > > > of our abilities.
> 
> Which license? As far as I can tell, what I was proposing in terms
> of releasing Aldor source under the Aldor Public License would
> conform perfectly with that license. Releasing Aldor under GPL
> would be more restrictive but compatible with the Aldor Public
> License.

Debatable.  Anyway, stated intent to release code is not the same as
actually doing it - the assumption cannot and should not be made that
it can be released.

As Mike has already mentioned, if people weren't interested in getting
this done a simple statement that it wouldn't be done would keep their
inboxes neater.  There may be reasons for the delay that they can't
talk about legally, and given that they have released Axiom to us in
the past I think we owe them the benefit of the doubt.  Yes the delay
is harmful but there are things we can do in the meantime - writing the
academic papers needed to describe the existing mathematical domains in
Axiom doesn't necessarily need to wait on the language decision, at
least to start.

Cheers,
CY


 
____________________________________________________________________________________
Sponsored Link

Compare mortgage rates for today. 
Get up to 5 free quotes. 
Www2.nextag.com