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

Re: Not-so-mass bug filing for the patented IDEA algorithm

On Tue, 10 Apr 2007 14:17:13 +0200
md@Linux.IT (Marco d'Itri) wrote:

> On Apr 10, Florian Weimer <fw@deneb.enyo.de> wrote:
> > I plan to file a couple of bugs (not too many, probably a dozen) on
> > packages which contain implementations of the patented IDEA
> > algorithm -- because the presence of that code makes them
> > non-free.  As far as I know, no program in Debian actually uses
> > this code, it's just inherited from upstream libraries.

> Is there actually a patent violation if the code is not used?

Distributing the code could still be an issue - the problem with
patents in general is that you won't find out for real until it gets
before a court.

IANAL, but being open source, a patent lawyer would probably try to
claim that distributing the code ALLOWS the infringement of the patent
as if that makes Debian complicit in the infringement. Whether the code
actually does include an implementation of the patented algorithm
hasn't been confirmed here - it could just be a linkage.

Which are the offending libraries? Is this mass-bug-filing intended to
be against the applications that link against the libraries or just the
offending libraries themselves? If the applications don't use that code,
I don't see that the maintainer of that application can do a lot about
the 'bug'. Why do the upstream libraries contain an implementation of
the algorithm in the first place? Wouldn't it be better to have dfsg
versions of such libraries? (i.e. I'd support the bug filing against
the specific libraries for this purpose.) If the code isn't used, it
should be trivial to create a patch that simply comments it out or
removes it entirely. Or to prevent a SONAME bump, replace the function
definition with a no-op/error.


Neil Williams

Attachment: pgpz9JEmrgpLt.pgp
Description: PGP signature

Reply to: