www-legal-discuss mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From robert burrell donkin <rdon...@apache.org>
Subject Re: IBM's license for WS-Security - Take #2
Date Tue, 05 Jul 2005 17:35:45 GMT

On Mon, 2005-07-04 at 10:48 +0100, Ben Laurie wrote:
> robert burrell donkin wrote:
> >>>If your goal is to ensure that your licensees have all necessary patent 
> >>>and copyright rights to create whatever derivative works they want 
> >>>from Apache's source code, you are bound to fail.  That is an 
> >>>impossible task.
> >>
> >>Yes, I think that has been the goal, and if it is not, then I think a 
> >>clarification is needed from the Board of how the patent issues are supposed

> >>to be dealt with both for the relevant projects as well as for all the users

> >>out there, who think that ASF code base has no known patent issues attached 
> >>to it.
> > 
> > i cannot see how the ASF could possibly offer a patent guarantee for
> > derivative works. there is no limit on the patents which a derivative
> > work may infringe. the best that could be offered is a promise that all
> > derivative works would be entitled to use any patents owned by
> > contributors that necessarily infringe the original library. 
> You mean that the library necessarily infringes 


> (not sure I understand 
> the use of this word "necessarily", btw. though I've noticed it is popular).

it's so popular that it's even in the ASL 2.0 :)

IANAL but i think the word describes a distinction between those patents
which are infringed by the nature of the code and those which are
infringed only when it is combined into a derivative work.

> > i don't quite grasp the necessity of the connection between derivative
> > works and the issue of code with known patent encumbrances. IMHO what
> > most users think of as 'having no known patent issues attached' is that
> > the ASF has licenses for any patents know to be necessary to use the
> > software. 
> The point is that we are trying to produce code that is useful to people 
> downstream, including those that modify or add to it. We don't want 
> patent grants to evaporate simply because they changed a line of code 
> somewhere.


IMHO it is not unreasonable (though) for a contributor to ask that
derivative works that infringe patents that the original did not must
obtain licenses for those but that licenses granted for the original are
not lost just because it is combined into some large work.

say i owned three patents: 
(a) a general software patent (boo hiss evil) about logging (say);
(b) a whizzy bit of electronics hooked up to smart control program used
to log information about aircraft;
(c) a unrelated patent about boats.

i contribute code to the ASF which embodies (a) in project foobar. i do
not think it unreasonable that derivative works of foobar that infringe
patents (b) and (c) are unprotected by the license granted for the use
of (a) in foobar. the necessarily bit comes in with (b): though foobar
may be used as part of the control code, (b) is not necessarily
infringed by foobar and use of foobar in the control code would not
allow (b) to be automatically granted by the foobar license. i would
hope that any derivative works of foobar would inherit the same license
granted for (a) to foobar.

(i suspect that we're all in consensus here but fumbling around to
understand the language...)

- robert

View raw message