Re: Patents again

[David Kastrup]

Abdullah Ramazanoglu <abdullah@ramazanoglu.tr> writes:

> begin  David Kastrup <dak@gnu.org> dedi ki:
>> 
>> That is like saying by calling a plumber you automatically pay him.
>> This isn't so: you still have to do the payment yourself, and the
>> plumber can't just pilfer the amount from a wallet lying around.
>
> Would it solve the problem if a new clause in a future "GPL v.3"
> explicitly stated that any patented code is unenforceable against
> any GPL project?

No.  Only the patent holder can grant a licence.

> A patent holder accepting such a GPL v.3, would *explicitly* have
> granted a non-restrictive patent license for all the GPL codebase, I
> think.

No.  You are always talking about "codebase" and similar wishywashy
terms.  Licences apply to individual copies.  Only for the copies a
patent holder distributes does he have to obey the GPL.  And only the
original copyright holder can hold him to it, not the recipient.

>>> But it doesn't end here. What if they indirectly and from multiple
>>> untraceable sources sneak patented code into GPL codebase? It should
>>> be easy to do, at least I can readily think of a couple of ways.
>> 
>> It does not matter how the code got there.  The patent holders have
>> a right to payment if their code gets used without a licence,
>> regardless of whether they put it there themselves or not.  You can
>> sue them to stop distributing code under the GPL without a
>> non-restrictive patent licence.  But you have to sue them: there is
>> no automatism involved.  In particular, the party receiving the
>> code has no base to sue for a patent licence, only the party that
>> has the copyright on the original GPLed code can sue for
>> compliance.
>
> However, if the patented code got there indirectly, then patent
> holders would be on pure legal ground, and you wouldn't even be able
> to sue them.

I repeat: it does not matter how the patented code "got there".  You
can force the patent holder (and everybody else) to retract any copies
he had been distributing without complying to the GPL, by including
non-licenced patented code.  Regardless of _who_ distributed the stuff
with unlicenced patents in it, you can force him to stop that.  Only
the patent holder has the additional option of continuing to
distribute _if_ he licences the patent for that use.

> As for the case where it got there directly, I'm surely not a
> lawyer, because I simply cannot understand how a GPL-bound patent
> owner could enforce their patent on any GPL program.  By
> distributing their patented code under GPL, they're explicitly
> stating that anyone can use it, modify it, and redistribute it under
> the same conditions, freely in both senses.

No, they aren't doing anything of that sort.  That's only in case when
they are _complying_ with the GPL.  But when they are merely
_claiming_ compliance but not licencing the patents, then the
recipients may not redistribute the code further, just like it was not
allowed to be distributed in the first place.

> They're *explicitly* giving the permission and freedom for their
> patented code.

Only to their customers.

> How then, can they ever go to court and claim that I cannot use
> their code without paying to them for their patent?

If I am a direct customer of theirs and buy stuff from them they claim
I can use, then they can't sue me.  But if their licence did not
extend to non-customers and I use the stuff nevertheless, for example
by taking code from the program and putting it in to a completely
different program, they can sue me.  If they sue me, the copyright
holder of the GPLed original can in contrast sue them for breach of
GPL, but there is nothing I can do without a valid licence from them.

> It's beyond me.

If your neighbor steals apples from the grocery store, this does not
give you leave to steal pears from him.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum