Subject: Re: Is free software innovative ?
From: Ben_Tilly@trepp.com
Date: Tue, 12 Dec 2000 09:14:04 -0500


david@fetter.org wrote:
> Bernard Lang wrote:
>
> > Hi,
>
> > a common argument that I see against free software (and used as a
> > justification to ignore free software when discussing software
> > patenting) is that free software is not innovative.
>
> I have a similar concern, and it's not because I have an anti-free
> software agenda.  This concern is legitimate, and won't vanish with a
> Raymondesque hand-wave.

I don't because I regard most of the "innovation" in commercial
software as being fluff marketing.

> > Innovative means : containing stuff that could have been patented,
> > according to a rather strict application of the novel and inventive
> > rule.  (not the USPTO garbage criteria)
>
> OK.

Whose strict application of the rule?

> >    examples I have in mind  (I may be wrong ... correct me
> >       - if not innovative or inventive
> >       - if facts are wrong
> >       - or anything else
>
> >     add more facts, if you want, keeping it short

A big one you missed, the web.  The original web server/browser were
developed and released free from CERN.  Several ideas involved would
have been patented had it been done by a commercial company.  (Which
would have killed the concept, but that is life.

[...]
> > it has to contain original material ...  being just nice, useful,
> > well programmed is not enough.  for example, I have no idea whether
> > Linux contains new technology that could have been considered
> > patentable.

I have heard that ReiserFS contains patentable stuff in it.

> The kernel module stuff could be.  Then again, a novel use of knives
> and forks could be under US patent law :P
>
> The most recent Free piece of software I could add to this that's
> innovative would be Perl, developed with CIA funding.  Emacs was
> developed with lots and lots of state funding at MIT.  TCP, UDP and IP
> were done with government grants start to finish.

Perl was developed with CIA funding?  That is an interesting and
IMO dubious claim!  Larry Wall did work at JPL for a bit, but AFAIK
the bulk of Perl was done on his own time with his own energy.

Anyways Perl's RE engine contains many ideas which could have been
patented.  Very specifically differentiating .* from .*? was a step
which had not been obvious in advance, as judged by the fact that
nobody did it for many years even though it was easy to do and very
useful once done.  Further evidence is provided by the fact that
everyone else is now advertising "Perl compatible regular
expressions".

I leave it to your imagination how much taking out such patents
would have dampened the popularity of Perl...

And there has been lots of recent innovative free software.  For a
random instance, look at Ogg Vorbis.

  http://www.oiskas.com/ogg-vorbis/faq.htm

The specific algorithms it uses are easily patentable.  In fact the
various competing formats (which range from proprietary to almost
free) are patented.  The entire point of Ogg Vorbis is to not be
patented.

Cheers,
Ben