[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Falcon P.L. license (ITP:Bug#460591)

Josselin Mouette wrote:
> On jeu, 2008-03-27 at 18:58 -0700, Sean Kellogg wrote:
>>> No one can patent the grammar that you wrote, so this is completely
>>> useless. The only point of these clauses seem to claim the copyright on
>>> scripts using the language.
>> Huh? Why can't someone patent langauge grammar/syntax?
> I should have written “no one *else* can pattent the grammar that you
> wrote”.

Unluckily, this is far from being true.

Any patentable idea can be patented by anyone, and the patent is the
preferential title to establish authorship rights on the idea. In other
words, Mr. A can have a (patentable) idea and start making use of it,
Mr. B can patent it and then Mr. A has this options:

1) quit the idea and go fishing for a living.

2) FIRST, act against THE PATENT, trying to invalidate it as it wasn't
based on an original idea; it's A that must demonstrate he was right.
SECOND, when (if) he wins the trial that will make Mr. B patent to be
removed, act against Mr. B and ask for damage refounds.

Funny enough, in scenario 2 Mr. B may file another patent on the idea,
changing it enough to be recognized as having some novelty content by
clerks that usually have just a vague idea of what a novelty in the
field may really be. This can go on forever; if you just take some
minutes to google a bit, you'd find some interesting cases (some similar
scenario are even used as common examples by Mr. Stallman, when he talks
about "de-grade", and about the compression algorithms that GNU had to

A recent U.S. law recognizes to "prior invention" some limited rights
(i.e. that of continuing using inventions as if the patent wasn't
filed), but it's a quite limited guarantee, and still you'd have to
prove the invention was yours while i.e. being suited by the patent
filer. Also, U.S. is not the rest of the world, and there are some
countries (Italy to say one I am paricularly concerned with) in which if
you don't patent first, you're going to have an hard time.

The "unpatentability" claim in this licenses can't prevent Mr. B to
patent an idea protected by the license, but it has the net effect to
make Mr. A able to sue Mr. B for damage directly; the damage would
descend from having broken the license terms, which, in case of having
patented an idea that can be seen as "the work" or "a derivative work",
would be quite easy to prove in a trial.


Since I am here with mail client open; I would like to redefine the
thing I said this night about FPLL being more *restrictive* than LGPL
for some (but not all) applications.

The reason that made me to start all this was exactly the fact that
Falcon engine is quite invasive, and it also allows to be quite
"invaded". The most common way  an application can use the Falcon engine
consist in extending and modifying the Falcon::VMachine class; also, non
trivial embedding applications would have to inject code in the engine,
and/or in the standard modules that are provided under the same license
terms of the engine. Not to talk about 3d party modules, which must
include relevant inline code and perform massive integration with the
engine. This is the practical reason why all the major scripting
languages have a "proprietary license"; I investigated them, but found
impossible to apply them as they are usually application (or
"foundation") specific. Moreover, Falcon engine requires and allows an
higher level of integration than those usually provided by the scripting
engines I have analysed.

Applying LGPL as-is to Falcon, without complex and carefully suited
exceptions, would have the net effect to enforce LGPL to every thing
that goes with Falcon: Falcon (script) applications, applications using
Falcon as a scripting engine and 3d party extension modules. The fact
that LGPL *defines* a "free zone" of applications that need not to
comply to it doesn't mean that those applications may exist in practice.
In case of Falcon, that "free" area would be an empty set. This is why I
wrote the license as such and this is why I claim that FPLL is *less*
restrictive than LGPL/Apache2. They define a set of applications where
they are not applied, and so, a set of completely free applications (for
whose FPLL would require just a notice to be placed somewhere), but
after careful studies I realized that in case of software as Falcon
engine that area would have been void. Or at least, users may have been
rightfully worried about their applications to fall in a covered
category, which may have scared away some business developers in some
scenario I was willing to include.

That wasn't acceptable, as I have already commercial, closed source
applications that run the Falcon engine.

The fact that every single programming language, when using LGPL or GPL,
is forced to provide an exception is a clear sign of the deficiency of
LGPL in this area (read, *excessive* strictness), and this even for
languages that provide closed (non embeddable) engines or whose engines
require a level of integration far less intrusive than Falcon's.

That's why I needed FPLL; otherwise, now, we would be discussing of LGPL
with very funny exceptions here, and frankly I can't see why that should
be considered "better". Tons of exceptions, different from case to case,
doesn't make a better legal scenario for the open source community than
a small set of better suited licenses.

Nevertheless, I am dual licensing Falcon with GPLv3. If you want to get
more restrictions, I am not stopping you... :-)

Giancarlo Niccolai.

Reply to: