[clug] Re: A most interesting read, most interesting

[Michael Cohen]

On Sun, Dec 31, 2006 at 03:09:46PM +1100, Jacinta Richardson wrote:
> I think the issue here is one of language.  Peter is concerned about 
> _ideas_ while you and Martijn are focussing on the expression of those 
> ideas in _code_. Let's just say that I have an amazing new idea.  Perhaps 
>  it's a method in which internet shopping can be made easier because I've 
> come up with an algorithm which uses some fantastic algorithm to predict 
> all the stuff that the consumer was thinking of buying plus a few things 
> that they could be persuaded to buy.

In a utopian world this would really work like this - a person could wake up
one morning and dream up a wonderful new invension that will revolutionize the
world as we know it. In reality, however, this can not really happen. Every new
idea is based on some previous "prior art". To take your example, your algorithm
would probably be based on some other algorithm much improved or whatever.  If
you were to take a patent on your new algorithm, how much credit should you
give to your prior art? is it fair to claim that your idea came from a vacuum?

Whats worse is that in order to take a patent you only really need to think
about the idea - not actually make it. I can envision myself flying to work on
a jet pack - but I dont think I can make one. Does this mean I should get a jet
pack patent? and if I were awarded one would it be fair to all the people who
are currently working on actually making a jet pack?

The problem with the patent system is that the utopian concept of "an idea" is
simply not true any more. People dont invent in a vacuum any more. Instead
patents are used in industry today as a means to stop your competition from
inventing the same thing. They are nothing more than weapons in the war that is
business. If you want to get a product to market you need to find those people
who might hold patents which by chance might cover the area you are trying to
develop (they dont need to have a real product at all, just an idea about a
product) and then bribe them to not sue you. There are companies out there
which have no other revenue source than to collect and enforce patents - and
that is such a lucrative source that most largish companies collect vast number
of patents - to use in defence or offence.

For example if a large company holds a large number of patents and can use that
fact as bargain chips. They can negotiate deals with other large companies to
allow them to develop ideas which might be covered by the oppositions patent in
exchange with allowing the opposition to develop ideas owned by them.

Although the patent system appears at first to support small inventors by
affording them protection from big players, it does exactly the opposite. If a
small company (which might only have 1 or 2 patents) tries to actually produce
their ideas, they might find that they are infringing on patents held by larger
companies. (To take your example, you try to make money from your new ecommerce
algorithm, but find it needs to run on a GUI which is patented by MS with a
dialog box look and feel which is patented by MS also and you might ship a
system running the NTFS filesystem which is patented by MS too. So all of a
sudden you are breaching at least 3 MS patents. MS then turn to you and say -
your idea is pretty good, we like it - give it to us for $10. You say, its
worth way more than that, no way. They say, its not worth anything if it never
gets to market - you cant make a product without their blessing because they
trump you 3 patents to 1. You lose and give it away - game over).

One of the most important challenges in engineering is the difference between
knowing you can do something, and actually doing it. Every high school kid
knows how to do lots of things - from making a battery to building a power
station. Very few are able to take it to a product stage. Patents create a kind
of casino where you can gamble on the next big thing, put your money down, and
hope that some big company is trying to develop the same thing - then you might
win the jackpot...

Traditionally patent offices were very careful about the patents they issued to
ensure that there was no prior art, and the innovation was really innovative.
These days its difficult to do, because the scope of science is so wide, and
the patent office needs to cut back and streamline its operation. The current
view seems to be to award patents first, then let the parties fight it out in
court. There have been patents issues on really stupid things, but contesting
it in court is expensive and dangerous so many people just prefer to pay
licensing instead - even if the patent should have never been issued. This is
particularly bad in the area of software patents where it affect just about
everyone e.g. the gif patent, the mp3 patent, the FAT file system patent, and
the list just keeps going. Its pretty difficult to produce any software product
today without having to license some form of patent. Software patents
absolutely kill software innovation and dont make sense to anyone.

Open source is about leveling the playing field - if you are able to publish
your idea widely enough and soon enough, you can stop patents from being taken
on it. Then everyone wins (including you), because everyone gets an equal
chance to implement it. 

Michael.