Re: Software Patents: The Lowest Common Denominator

[Alan Horkan <horkana maths tcd ie>]

On Sun, 17 Jul 2005, Jeff Waugh wrote:

> Date: Sun, 17 Jul 2005 13:41:36 +1000
> From: Jeff Waugh <jdub perkypants org>
> To: foundation-list gnome org
> Subject: Re: Software Patents: The Lowest Common Denominator
>
> <quote who="Alan Horkan">
>
> > In an effort to let developers do what they do best and get on with
> > developing software [2] so I would like to propose a standard build flag
> > which could be used to fence off code that could be effected by software
> > patents.
>
> A few mostly rhetorical questions:
>
>   Are you violating the patent by implementing it with code,

My limited understanding is there is nothing preventing the implementation
in code of a patented idea as this can be viewed as an acedemic excercise.
(And the example of Freetype seems to suggest there is some merit to
this.)

> or by building it into a binary (and then distributing it)? I would be
> hard pressed to agree if you suggested the latter.

I am not a lawyer, you will of course need to ask one which is why this is
such a damned nasty problem.  I believe there is a problem is you
distribute binaries with intent, if you know in advance you are
distributing implementations of patented ideas (but you seem to recognise
this yourself).

It is nearly better for developers not to touch patents with a bargepole
for this reason, trying to identify patents and avoid them only makes it
worse for developers.

>   Violating a patent with knowledge and intent is generally punished more
>   ferociously than doing so unwittingly. In some jurisdictions, it amounts
>   to triple damages. A "--unencumbered" build flag would clearly demonstrate
>   knowledge and intent in the implementation.

A proper legal opinion is needed which is why I brought this to the
Foundation list rather than to one of the development lists.

I am not sure exactly how this might work.  Perhaps we would have a
policy advising developers to distance themselves from software
patents.  Perhaps it would be a matter for the foundation lawyers to
recommend using a build flag if the legal status of code is challenged by
customers/auditors/competitors.

It was my hope that the foundation could seek the necessary legal advice
and come up with practical solutions for developers to protect themselves.

>   Would the absence of this build flag deny our ability to efficiently draw
>   lines thicker than one pixel on the screen [1], or just the ones we're
>   *really* worried about (please provide definition for real worry)?

As I mentioned in my previous mail I've noticed comments in bugzilla about
software patents, particularly with regard to Nautilus and 'Spring Mounted
folders' and Apple allegedly hold a patent somewhere in that area.
(Initially I was going to check but I'm not a lawyer and even if I read
the patent I don't think it would necessarily make things any less
ambiguous for me.)

I think I have more chance of learning to read Monoglian than developers
have of safely and correctly interpreting and avoiding software patents
and there needs to be some way to allow developers to protect themselves
from the software patent minefield other than entirely avoiding whole
areas of software.

We know software patents are already a problem.
How can we start defending ourselves?

Sincerely

- Alan H