[Fsfe-ie] Minutes 30-40

Aidan Delaney adelaney at cs.may.ie
Tue Jun 1 08:49:57 CEST 2004


Hey,
Here's minutes 30-40 (spell checked), I'll have 40-45 for tomorrow.

[RMS] So...sometimes you'll be able to avoid patents, sometimes a
standard means it doesn't matter even if you find a better way to do the
job, it's still useless coz it's not the standard.  And sometimes there
are patents so broad that avoiding it is hopeless, the entire field of
public key encryption was off-limits in the U.S. because of a very broad
patent until 1997.  We just couldn't do it, most people couldn't do it.

So that's the option of trying to avoid patents.  The next option it to
try and get a licence to use the idea.  Now that's not necessarily
possible.  The patent holder doesn't have to offer you a licence, the
patent holder can say "I will allow no competitors, I'm shutting you
down", and that does happen.  I once got a letter from somebody whose
family business was making casino games using computers, and he was
being threatened by a competitor who had a patent, and the competitor
wasn't...didn't...wouldn't offer him a licence, just said "you gotta
stop, you gotta close your business".  When I looked at this patent,
claim one said "A network containing more than one computer where, each
computer implements more than one different game, and each computer can
show more than one game session at a time", that's all.  I'm sure that
there were various computer labs, back in the 1980s, in universities,
where they set up a network of workstations that had window systems and
installed more than one multi-player game.  And if so they were doing
exactly what that patent described and they probably didn't even realise
that there was an idea there and this is one interesting point, the idea
of a software idea patent forbids you from using might be something you
did without even conscious awareness of it as-such.  but anyway, the
patent office tends to think that if somebody else did something once,
or did one thing and this other guy does it multiple times or several
different things, that that's an invention.  But actually in computer
science it's a basic principle that anytime you do something once that
you can do it n times.  That if you have one thing, you can have a
choice between n different things, and if you've done...and if you do
things you can do them in parallel.  Anyway, although this patent was
utterly absurd, we couldn't help him.  He had no money to consider going
to court, so he just had to shutdown.  

However, most patent holders will offer you a chance to get a patent
licence, but they charge a lot of money for it.  For instance the patent
holder of the natural order recalculation patent was asking for 5% of
the gross sales of every spreadsheet in the U.S.  Now, maybe you could
afford to sign one patent licence like that, but what really happens,
since...but you know...suppose that another patent holder comes and
another and another, what happens when the twentieth patent holder asks
you for 5%?  And then another one comes along, well this is really just
a joke because people in business told me that two or three such
licences would be enough to drag your business into failure.  

However there are some software developers for whom it's easy...it
usually, usually easy and painless to get a patent licence.  And those
are the global mega-corporations.  Because they have lots of patents and
they and they use those to force everybody else to cross licence with
them.  I.B.M. wrote and article in Think Magazine issue number 5 1990,
about I.B.M.s patent portfolio.  It said that IBM got two types of
benefit from it's 9000 active U.S. patents.  Benefit one was collecting
money.  Benefit two was getting access to the patents of others through
cross-licencing and the article said that the value of benefit two was
perhaps an order of magnitude greater than, in other words, ten times as
big as benefit one.  Now this is interesting because, you see, the
patent system is like a time consuming lottery.  If you are a small
business what happens to you from the patent system is, to a large
extent, affected by randomness.  But IBM is so big that IBM can measure
the average for us.  IBM can measure how big the benefits are and how
big the harm is.  And IBM's conclusion, stated in this article, was the
harm is ten times a big as the benefit.  Now you might think that that's
impossible, it's like saying "well the children are above average,
right?", it's a zero sum thing, how could the harm on the average exceed
the benefit?  Well the answer is, that harm dosn't really happen to IBM
because IBM practises cross-licencing and avoids it.  So there's this
big harm and a certain amount of good, but IBM makes the harm go away so
all it gets is the good.  But IBM measured for us the size of the harm
that didn't happen, {sniggers from audience} and that measurement is
useful because while IBM has 9000 patents and they can make everybody
cross licence, if you've got a small business you aren't going to get
9000 patents and you aren't going to be able to make most people
cross-licence with you.  Now many companies say "We've gotta get some
patents to defend ourselves with".  Well the problem is each patent
points in a particular direction, it's like a landmine buried in a
certain spot.  You know if..if I've got a patent pointing there and
there, and somebody over there points a patent at me, my patents don't
do me any good at all.  IBM with its 9000 U.S. patents, it has patents
pointing everywhere, so no matter where somebody is, if he attacks IBM,
IBM can almost certainly ahh...threaten to shoot back.  And say "lets
uhh.. lets sigh an cross-licence so that neither of us gets hurt.  Now
interestingly, these many companies that say to their employees "Get us
some patents we've gotta have something to defend ourselves with", while
they like to say it's for defence, they don't offer to sign a contract
with the employees committing themselves to using these patents only for
defence. So, a few years later when the company has been bought out or
has new management and they got this patent pointing over there well,
they might just decide to use it opportunistically for aggression, and
nothing can stop them then.  

Now this practice of cross-licencing is very important to understand
because this is why the mega-corporations want software idea patents,
because they know that they will cross-licence with each other and they
will form an exclusive club, and the rest of us will be down at the
bottom of a cliff wondering how we're ever going to get up there.  It
gives for a measure of dominance over everything.  And that's why they
want it.  It's also important to recognise that this reveals the
falsehood in one common myth about the patent system. The myth of the
starving genius.  This myth starts with a scenario, which is implausible
in  number of ways, unlikely.  It says, imaging there is a brilliant
developer of whatever, who's been working for years to come up with a
better way to do whatever and now it's working and he's going to start a
business.  And because his idea is so good his business is going to be a
tremendous, it's going to be a great success.  Except for just one
thing, which is that the big companies will come and compete with him
and take all his customers away and then his business will fail and
he'll starve. {laughter from audience} Well, this scenario is
implausible whatever (is to happen {slightly garbled}).  First of all in
high-tech fields, people don't usually have ideas in isolation.  Once in
a while yes, but usually they have ideas because they are working in the
field with other people and working on solving real problems.  The next
thing is, that, if this person is going to start a business well, it's
probably going to fail anyway no matter what. 
-- 
Aidan Delaney 	email: adelaney at cs.may.ie
		web:   http://www.cs.may.ie/~adelaney
		gpg:   http://www.cs.may.ie/~adelaney/public_key.asc
		
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://mail.fsfeurope.org/pipermail/fsfe-ie/attachments/20040601/ac76f815/attachment.pgp 


More information about the FSFE-IE mailing list