Microsoft/Novell Partnership

John Macdonald john-Z7w/En0MP3xWk0Htik3J/w at public.gmane.org
Mon Nov 6 06:14:53 UTC 2006


On Sun, Nov 05, 2006 at 09:17:26AM +0000, Peter P. wrote:
> John Macdonald <john at ...> writes:
> 
> > But, if code covered by GPL can be claimed to carry out a
> > process that is covered by a patent, copying the code is not
> > an issue, and permission from the author is irrelevant (unless
> > the author also is enforcing the patent).  The GPL can be used
> > to force the owner of the patent to stop copying the GPL code,
> > but if the patent owner doesn't care about distributing GPL
> > code that threat is especially hollow.
> 
> This paragraph is slightly confusing ;-) Could you please rephrase it a little ?

If a company sends you a legal notice that you are infringing on
their patent, the notice will describe the conceptual process and
results you are carrying out and achieving.  It will not matter
whether you were allowed to copy the code - it is duplicating
the patented process without a license that is the violation.
So, whether your code is covered by GPL, derived from the original
Bell Labs release code, or whatever does not affect that you are
breaching the patent.  (It might help you in your course of
proving that the patent is invalid; but in the mean time you may
received an injunction that forces you to stop selling your
product for a while.)

> > expensive court orders that do things like prevent sales -
> > even if you eventually win the challenge, not being able to
> > operate your business for the duration can be devastating.
> > And, maybe, a few of those patents will fail to be challenged
> > successfully.  That leads to trying to code around a concept,
> > which is harder than just rewriting the code independently.
> 
> SINCE SOFTWARE PATENTS ARE NOT RECOGNIZED OUTSIDE THE US AND NEVER WILL BE WITH
> G*DS GRACE NOBODY NEEDS TO CODE AROUND A CONCEPT. ALL THEY NEED TO DO IS CHARGE
> MORE FOR THE US VERSION WHICH LINKS AGAINST THE LIBRARY WITH PATENTED CODE,
> DISTRIBUTED AS THE 'US VERSION' EXACTLY LIKE THE US INSISTS THAT CERTAIN TYPES
> OF CODE BE DISTRIBUTED AS CRIPPLEWARE 'NON US VERSIONS'. THIS IS A
> CVS/RCS/VERSION CONTROL ISSUE AFTER ALL. </shouting>

See below.  The U.S. is important.

> > Or it leads to having a U.S. fork and a rest-of-the-world fork.
> > That would be painful and hurt Linux badly.
> 
> Most software that is worth some money already has that fork built in. US
> edition. NON-US edition, everything is there. Crypto code, patented algorythms
> and other types of code forking nightmares have plagued Linux and BSD
> distributions ORIGINATING in the US since their inception. BSD was on the COCOM
> no-supply of the west list before the fall of the iron curtain, just as an
> example. The US is the nexus of these code splits. No other country attempted to
> control glorified hashes and caesar codes as 'weapons' or tried to put engineers
> who were suspected of discussing it in jail. Additionally the US the the home of
> the DMCA, which means that USERS ARE FORBIDDEN TO DISASSEMBLE THE CODE SUPPLIED
> TO THEM, INCLUDING FOR THE PURPOSE OF DETERMINATION OF THE ALGORYTHM BEING USED,
> INCLUDING IF IT IS SUSPECTED THAT THE ALGORYTHM IS A PATENTED ONE.

Wrong.  The issue over crypto code was that U.S. law made
illegaly to ship it out of the country without going through a
prohibitably expensive per-customer approval process.  There was
no problem with importing crypto code into the U.S., though,
so no code fork was required.  It just meant that only non-U.S.
developers would work on the code, that the repositories had
to be outside the U.S., that U.S. repositories had to limit
foreign dowloading (impossible, but they  were allowed to just
make an attempt).  This was just a small piece of the system.
A U.S.  based distribution like Red Hat simply didn't include
crypto, because it was too hard for them to get around the
export barrier.  (Hugh will likely chime in here on details
that I've got slightly wrong; or maybe Henry - they were both
involved heavily in the crypto work for a while. Henry more
than Hugh, but I haven't seen Henry posting here recently.)

A patent would mean there were chunks of functionality that
could not be deliver to U.S. users of any distribution.  If the
component involved was sufficiently critical, people would
spend time working around the patent; but if the work-around
was too costly (either making things too slow, or reducing
functionality significantly), there would be a fork for the
rest of the world's users who would not want to pay the cost
for a legal issue that didn't affect them.  That leaves lots of
developers spending time developing two versions of the code -
taking more than twice as much time to do that portion of the
development.  This would leave MS laughing in the background
the way they did back in the days when there were lots of
Unix vendors who did a better job of fighting each other than
in fighting MS (or the way we've been able to laugh about MS
being forced to make special distributions for regions that
flex their anti-trust laws sufficiently).

-- 
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists





More information about the Legacy mailing list