Subject: Re: termless copyright and patents
From: <>
Date: Thu, 5 Oct 2006 11:58:34 +0900

Thomas Lord writes:

 > In my proposal, any skilled programmer is likely to think of
 > copying the string comparison algorithm to MySQL if they want to
 > speed MySQL up.

OK, I see where you're going now.  I think it seriously sucks, and
from your point of view, at that.  See my reply to simo.

 > The problem with your approach is that first, programmers may
 > not, in general, copy Alice's code to libsection11 or,

In general, they may copy it.  That is precisely what is promised by
the covenant in section 11.  Modified versions are at some risk,
because they might infringe claims not embodied in the code conveyed
by Alice.

 > second, if they have done so, people may not use that GPL library
 > in all of the usual ways.  That would be a failure of the GPL.

No, because the GPL already fails in precisely that way with respect
to third party patents, and that cannot be fixed by a covenant.

If your proposal were to be enacted, IBM et al could immediately
freeze their GPLvTL activities that might be covered by the proposal.
They would create a joint venture, and transfer those claims that they
were willing to put into GPL software under GPLv3 Section 11 to that
corporation.  Any work that they do on GPLvTL ("version Tom Lord")
software would be carefully camouflaged as "for hire".

Now what?  Is it really worth the extra complexity you'll need to deal
with that?  Do you really want to impose extra costs on people who
have a responsibility to their stakeholders to maintain the value of
their assets?  Not to mention to do stuff that the free software
community will have to object to as sleazy?  (Sure, people object to
enforcement of patents at all, but that's nothing new.)

And that assumes that IBM is willing to do any work on GPLvTL stuff at
all after being stabbed in the back that way.  Note that they can
easily stick to GPLv2, and if desired add the IC covenant or the GPLv3
Sec. 11 covenant, for all their current activities.