[Gllug] Software patents

[Richard Jones]

On Wed, Apr 23, 2003 at 02:14:51PM +0000, Chris Bell wrote:
>    There is a lot of discussion at the moment about software patents. The
> patent system was designed to encourage research and development, and allow
> the patent holder to have the first bite of the cherry, until the patent
> rights expired, so I thought that anyone trying to obtain a patent on almost
> anything had to prove that it was a result of original work, and not
> something already in the public domain.

No, you can apply for a patent for anything you like. The proof stage
comes later (often years later).

First you write your patent in a standard format, post it to the
patent office, and subject to the most cursory of checks they give it
a patent number. Then, up to a year later, you send your "claims". The
claims are what makes the patent, and those are what matter in court.

The formula for claims is quite interesting. Generally claim 1 starts
by claiming that you've invented The World, then claim 2 narrows that
down a bit to saying you only invented Life on Earth, and you keep
narrowing down until you reach something like claim 20 which is what
you're actually patenting.  (It's a wierd system, but then it was
invented by lawyers).

While you can write the application yourself, it's always a good idea
to get a lawyer to write the claims.

So far you haven't spent much money (< 20K). Now you need to open your
bank account because lawyers are going to take upwards of 100K to

(a) apply for your patent in every country in the world
(b) do a search for prior art

The search isn't necessary, but it kind of means that you've done the
groundwork to prove that your patent is novel. It's a stick to hit
potential infringers with.

In the extraordinarily rare case where a patent claim comes to court
(I recall reading that something like < 1% of all patents are
challenged in courts), then prior art becomes an issue.

>    Anyone could object to a patent application if they could prove that it
> was not something new, or already in the public domain, so worries about
> commercial companies taking out patents on existing open source projects
> should be groundless. It may be less clear if development is progressing
> within different groups, where one group could show that they are first to
> reach a satisfactory conclusion.

Yes, but this doesn't matter if you don't have money. If you get hit
with a suit for infringement, then you'd better either have lots of
money or a lawyer who is capable and willing to work "pro bono" (for
free).

This is why I advocate a kind of insurance scheme for free software
programmers, kind of like the GMC for doctors. Everyone pays in a
small amount per month, and in return we have a large organisation
prepared to take on the infringement / RIAA / MPAA / DMCA / DeCSS /
daft techno law cases.

>    I understand that it is rather different in USA because patent offices
> receive a large payment for each patent granted, and do not examine claims
> very closely, leaving battles over patent ownership to the courts so that
> the only winners are in the legal profession.

The UK patent office doesn't appear to be very thorough about checking
patents. In fact I don't believe they check them at all on the initial
application.

>    Many patents for the special video effects which are now commonplace are
> owned by Quantel, they did a lot of development work when almost everything
> was done in expensive hardware, and had a policy of applying for patents to
> cover anything they could devise, although they seem to claim that their
> patents cover the effects themselves and not the means to generate them.
> This means that they are happy to claim infringement of patent rights if
> someone tries to produce a similar effect using look-up tables, or software
> using clever mathematics.

They can claim that all they like, but that's not how patents work.

Of course, 90% of the battle with patents is having the biggest portfolio/
army of lawyers.

Rich (who has written two software patents ... blacklist me!)

-- 
Richard Jones, Red Hat Inc. (London) and Merjis Ltd. http://www.merjis.com/
http://www.annexia.org/ Freshmeat projects: http://freshmeat.net/users/rwmj
PTHRLIB is a library for writing small, efficient and fast servers in C.
HTTP, CGI, DBI, lightweight threads: http://www.annexia.org/freeware/pthrlib/


-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug