[Patents] Re: Bruce Perens: Software Patents vs Free Software

[orre@nada.kth.se (Roland Orre)]

On Fri, 4 May 2001, Edward Avis wrote:
> That's why patents are limited to twenty years - it's considered a good
> balance between incentives and freedom.  A two hundred year patent span
> might do even more to 'encourage new and non-obvious technology', but
> the price wouldn't be worth it.  It would be a poor 'patent bargain'.

As the the logical price per unit of used software compared to
production goes to zero for software, the logical protection
time, as the software development speeds up would need to go
down to zero as well.

To be efficient a software patent has to be rather wide. The one
protected by the patent will thus feel safe for quite some time.
(20 years is an almost infinitely long time) and this will thus
not be an incentive to enhance the development within this area
(as they had to spend so much money on it anyway...) and they
will be rather free from competitors due to the monopolistic effect
of the patent.

With a copyright protection scheme it is different. The protection
is more specific and the protection will therefore be efficient
a much shorter time, before competitiors have developed competing
alternative solutions. In this case the long protection time
of 50 years will not harm, the alternative solution is usually
better and there will thus be an incentive for the company
being protected to enhance their own solution as well.
This will thus be a more fair competition. You can not steal
a solution due to copyrights,  you have to develop your own,
you have to be creative, but the patent could disallow you
to work around it because it strikes too wide.

With this I will also say that there is no determinable
"obviousness level" for a software patent solution either.
The "obviousness level" will be dynamic as well, being
a function of the current state of the art and the collected
prior art. To handle such increasing "obviousness" levels
which patents would need, I have hard to see as anything
like a human can handle.

> Consider the ends of the patent system, not the means.

Yes, in the long run, when we have started to somewhat
master the "programming language" of the physics, which
is hard to say when but my guess is within 30-50 years
then copyright would do well for "hardware" as well.

Of course, many people may consider the current
"silicon compilers" and high level CAD tools to somewhat
give the hardware some properties of software already now.
Far from it, these are just like game playing when we start
master the technical principles needed for "matter compilers".
Then  the current industrial concept will be outdated also for
material sciences.
The patent system has probably died automatically before that...
But the copyright system I see to be still be valid.

So, from this point of view, it would be completely irrational
to first replace copyright with patents, which will slow down
the software development and increase the costs, and possibly
also delay the more advanced technical developments for a long
time, until finally copyright will do for both.

        Roland Orre