🚀 go-pugleaf

In article <kr3h53INN3p3@network.ucsd.edu> mbk@lyapunov.ucsd.edu (Matt Kennel) writes:
>I have a suggestion to "improve" the GPL that I don't think anybody will
>really contest:
>
>An addition saying specifically something like:
>
>	"The FSF disclaims all potential patent rights for any algorithms
>         implemented in this code.  Furthermore it wishes the release of
>	 this code to be considered as 'publication' for the purpose of
>	 establishing existence of 'prior art' on algorithms we invent
>         or use in our code."
>
>Just to make everything completely explicit.

I'm not sure this is aggressive enough for the Brave New World of
software patents.

If the GATT treaty goes through and it then becomes impossible to ever
get rid of software patents, this may not be such a good idea.  It might
soon happen that the only way FSF could stay in business would be to
have patents and to enter into cross-licensing agreements with IBM and
other big software firms.  gcc could then use compiler tricks that IBM
or MIPS or someone has patents on in exchange for their ability to use
tricks that first appeared in FSF code.

If FSF decided to follow this strategy, it would look for several
things in gcc2 that weren't in gcc and file for the broadest possible
patents on them.  The patents would be licensed free of charge to anyone
in one of the two ways described below:

1.  Any freely redistributable (GPL-style, X-style, or public domain)
program could use FSF patents freely.

2.  Any organization producing proprietary code could use FSF patents
by signing a cross-licensing agreement: e.g. IBM could use all of FSF's
patents if FSF can use all of IBM's patents that apply to software.

Judging by some IBM statements that RMS has quoted saying that they
see the chief benefit of their patents as being the ability to crosslicense
other patents, they'd probably be happy to sign such an agreement; other,
more aggressive organizations (Intel, say) would be less willing.

The clock is ticking if this strategy is to be applied: to patent
something that gcc2 uses, the application must occur within one year
of gcc2's publication.

If this strategy were followed, and the LPF nightmare comes true, where
software patents paralyze development, a software author might then have
an "out" -- donate the program to the FSF, which would have crosslicensing
agreements with almost everyone.

Just a suggestion.

--
Joe Buck	jbuck@ohm.berkeley.edu