Subject: Re: GPLv3 draft
From: simo <>
Date: Tue, 11 Apr 2006 09:27:55 -0400

On Tue, 2006-04-11 at 13:29 +0900, Stephen J. Turnbull wrote:
> >>>>> "simo" == simo  <> writes:
>     simo> I think that you are muddying waters here :-)
> Only in the sense that I know I misunderstood, and want to understand
> better.  Therefore I ask questions that "those who think they know"
> don't bother to ask.  Yet, the waters _are_ muddy, by the very fact
> that this thread has occurred---I'm not the only one in it.

I understood that, do you see the smile :-)

> My aim is to get you to stop "giving me fish" and instead "teach me
> _how_ to fish."  I do not deduce from that fact that you consistently
> expressed yourself in logical fallacies that you are incapable of
> teaching me what I wish to know.

I wish I had enough knowledge of all matters to be a master, I am not
and I humbly submit only what is evidently off track.

> Specifically, on what grounds does one conclude that the terms of two
> licenses are "the same"?  In particular, if two licenses have the same
> meaning under current law, but one is explicit and the other implicit,
> does strong copyleft permit distributing code licensed under the one
> under the wording of the other?  More generally, how robust does the
> conclusion that "the terms are the same" need to be to changes in
> legal environment to satisfy the GPL requirement?  Also, which terms
> simply don't matter (eg, my apparently bogus argument that the
> requirement to distribute a specific text, namely the GPLv3, would be
> incompatible with GPLv2, even if other terms were legally identical)?
> Why don't they matter?  Do I have any protection or recourse if the
> licensor declares that her understanding of the terms is _not_
> identical (regardless of what the author of the license template might
> say)?  Is "same terms" a transitive relation?
> Without an answer to such questions, I consider it unacceptably risky,
> not to mention morally presumptious, to redistribute material I
> received under a copyleft license under any wording but that of the
> original license.

Well, Law is _sometimes_ like a program, but more often it is a matter
of interpretation, and the interpreter is not always the same, I'd say
it is never the say. Writing law is like using defensive programming to
write programs that need to run on different compilers and hardware and
OSs, that all work slightly differently than others, only with a degree
of uncertainty a lot more ample as it is written and interpreted by
humans and not machines.

This means that you either get a sentence or some advice from a lawyer
that has enough experience to forecast the most probable outcome.

> Don't you?

I do, and the most safe thing is to ask the authors of the software how
they interpret the 2 licenses and if they give me permission to combine
2 works (unless licenses have a very clear history of compatibility).
Because, in the end, it is the authors that have the power to sue you,
nobody else.

> ObRef FSB: if that argument is strategically correct, then businesses
> will need to make choices with very-long-term implications as soon as
> the GPLv3 is published.  None of the choices may be very appealing;
> GPLv2 limits availability of the large body of GPLv3 code to your
> downstream, GPLv3 may involve new restrictions or grants you do not
> wish to make, and GPLv2 or later opens your downstream up to a fork if
> some users prefer v2 while others prefer v3.

That's true of any license.
But consider that the FSF will for sure re-license most of the GNU
software under the GPLv3 as soon as it is official. If you have
dependencies on that you will need to make a decision, which can easily
be to stop updating the GNU component you use and stick with the latest
one available under the GPLv2.
The fork is always a possibility, and I do  not think projects will
break on a license change, they will decide which license to keep their
project imho. 

> Let's not forget that IBM, one of the biggest companies contributing
> to open source, is also one of the world's best at manipulation of
> intellectual property, and that the Linux kernel, that poster child
> for open source, is reluctant to move away from GPLv2.  Such forks are
> a very real possibility.

I don't see how, the Linux kernel seem to be GPLv2 only, this means it
is an all or nothing switch, sure you will find companies that can try
keeping it under GPLv2, but how long can they do it ? Sooner or later,
new hardware, drivers, kernel features will make you do the switch.
If that good or bad I do not know, but I would not care that much about
forks, they happen all the time, most of the time they are a good thing,
even when they cause some wows around.

> It is important to understand GPLv3 and its context.  You may laugh if
> you like, but there are hackers involved in FSBs who have consulted me
> about these issues.  They understand that they ultimately need to talk
> to lawyers, but they do not boss the lawyers.  They need to advocate
> policy to those who do.  They want to gain some understanding of the
> social, economic, and legal consequences in order to decide what to
> advocate.  They know my biases, and for precisely that reason they
> trust my advice to be informative.  I'd like to deserve that trust.

I do not laugh at all! I've told you once, and tell you again, that I
appreciated many comments from you on this list in the past.
The fact is that in this case I have been interested and involved some
more, and I think you got some things the wrong way, but this is my
opinion only. More important, I thinks it is really good to question the
GPLv3, the drafting process is what it is for, questioning it to make it
better and more acceptable (up to a point). Only, I would like a bit
more constructive approach, and a wish to feed back the findings, or
questions, so that it can be really improved even if only in improving
the documentation that explain how it works.

>     simo> Don't think so, to solve [the problem of interpretation
>     simo> varying across jurisdictions] you should really draw a set
>     simo> of licenses each in the native jurisdiction language and
>     simo> with precise reference to this jurisdiction laws.
> As far as I can tell, the FSF doesn't consider that a good strategy.

I agree.

> Instead they prefer to make adjustments and translations only when
> necessary.  The GPL doesn't even grant permission to make translations
> of itself into other languages for *any* purpose, let alone for use as
> a license.  That language has not changed at all in draft GPLv3.

No, and will not change, I do not know if you know any other language, I
do, and I assure you that not allowing translations is a _very_ good
thing, I saw monsters in my own native tongue. Sometimes even lawyers
didn't get it right. Writing a translation is really difficult, and it
is better to reserve it when there are no other ways.
The Berne convention and later treaties make the GPL valid in most
countries. Local laws may twist it here and there, but there is nothing
you can do about that.