On Thu, 21 Nov 2002, ddutt@xxxxxxxxx wrote:
> > Either writing a plugin for a Cisco protocol is a controlled use of
> > the patent, or it's not. If it's not, then Cisco's request doesn't
> > make sense.
>
> It is not the writing of the decoder of a protocol that is the problem.
> It is
> its release under GPL and its effect on patent invalidation that is the
> problem.
>
This is clearly a lawyer problem, not a GPL problem.
Either the plugin is covered by the patent or not. If it is not,
then there is no way the GPL can invalidate the patent, and
there is no need to change the license. If the plugin is covered
by the patent, then no one except Cisco or a patent licensee is
allowed to write any plugin, under any license at all, that
decodes Cisco-patented protocols. The fact that the GPL mentions
the word "patent" isn't really relevant, since the logic applies
to all licenses.
OTOH, if there are other, non-Cisco FUD related reasons to
change the license, that's a different matter. Something to
consider : Having the GPL on a successful application like Ethereal
puts a great deal of pressure on vendors to donate source code.
This results in a higher-quality product for users, and makes
it easier for Ethereal developers to evolve the code without
stranding users. Changing the license removes that pressure, and
results in binary-only plugins and all the pain that implies.
As was mentioned earlier in the thread, it's still possible
(maybe, see patent isssues above) to release an open source plugin
if there is a binary plugin available, but why not keep up the
pressure to have source code released in the first place?