Re: [fielding@apache.org: Review of proposed Apache License,
Steve Langasek said:
> On Mon, Nov 17, 2003 at 03:48:12PM -0500, Joe Moore wrote:
http://nagoya.apache.org/eyebrowse/ReadMsg?listName=license@apache.org&msgNo=24>>> > Thanks. I think the new S5 looks like this:
>
>> > 5. Reciprocity. If You institute patent litigation against any
>> > entity (including a cross-claim or counterclaim in a lawsuit)
>> > alleging that a Contribution and/or the Work, without
>> > modification (other than modifications that are
>> > Contribution(s)), constitutes direct or contributory patent
>> > infringement, then any patent licenses granted to You under
>> > this License for that Contribution or such Work shall terminate
>> > as of the date such litigation is filed.
>
>> > That's certainly better. It still has a problem in the following
>> > scenario:
>
>> > 1. I start using Apache.
>
>> > 2. I develop a new process -- let's say an encryption algorithm,
>> > like
>> > RSA -- and patent it.
>
>> > 3. Somebody contributes an implementation of my algorithm to Apache.
>> > This somebody has patents on critical parts of Apache.
>
>> > Now I'm screwed: I can't sue Apache for illegally using my work
>> > without my permission, or I'll lose my license to their code.
>
>> > What this amounts to is a non-Free patent license, since it is
>> > revocable by an unrelated lawsuit.
>
>> No, that's much worse.
>
>> 1. I use Apache.
>> 2. I develop a new process, and patent it.
>> 3. Somebody reads my patent, and implements it in Apache.
>> 4. A competitor of mine (otherwise unrelated to Apache) runs Apache
>> and gains access to the method in my patent.
>> 5. I can not sue the competitor for patent infringement ("against any
>> entity
>> ... alleging that ... the Work ... constitutes direct or contributory
>> patent infringement") or I lose all patent grants under the Apache
>> license. 6. I can not sue the person who contributed the patented
>> method (for patent claims), for the same reason.
>
> AFAIK, #5 does not hold. You needn't allege that Apache violates your
> patent in order to argue in court that your competitor's work does
> violate your patent, regardless of where they say they got the idea.
> (It would simply not be relevant to the legal proceedings unless you
> intended to sue over Apache itself.)
How can I say that my competitor is violating my patent (by running the
contributed Apache code) without alleging that Apache contains the code?
Note that I did not suppose that my competitor did anything other than use
stock Apache, which happens to have my patented method implemented (not by
me) in it.
> It's still an improvement anyway, since only the patent license is
> revoked. I think the clause as written above is compatible with the
> DFSG, given my understanding of software patents as a legal abortion
> that should be regarded as non-existent by default.
All patent licenses relevant to "the Work" and granted by this license are
revoked.
>> 7. In addition, I can not, in _any_ lawsuit, allege that some part of
>> unmodified Apache has patent-covered code.
>
> ... without losing your rights to any patents covering Apache. But
> since software patents should be nullified en masse, and most free
> software seems to be covered by patents that we have no license to in
> any case, I don't see how this should be considered non-free.
I agree that most software patents should be nullified en masse (there are a
few novel inventions in software, but most software patents are crap). I
also agree that most software is covered by at least one crap software
patent that we have no license to.
However, a license is non-free if using the software incurs a significant
cost to the user. The proposed Apache license imposes a very significant
cost to the user. That cost is... they must give up the right to sue over
any patent that is in Apache.
>
>> This revised clause does much more to destroy the usefulness of
>> software patents. If Apache were licensed under this proposed
>> license, any software patent could be subverted (if the patent holder
>> uses Apache).
>
> A very nice side effect indeed!
A very nice side effect, but one that makes the Apache license more costly
than a "copyright on all modifications to this software must be assigned to
the original authors" license.
The Apache license is not the place to try to abolish software patents.
--Joe
Reply to: