Subject: Re: Examples needed against Soft Patents
From: "Stephen J. Turnbull" <>
Date: Thu, 20 Jan 2005 13:46:16 +0900

>>>>> "Taran" == Taran Rampersad <> writes:

    Taran> Stephen J. Turnbull wrote:

    >>>>>>> "Taran" == Taran Rampersad <> writes:

    Taran> Now, consider the Trade Secret:

    >> AKA the "ASP loophole" in the GPL.  People are well aware of
    >> this aspect.  But trade secrets can't force free software to go
    >> _backward_.

    Taran> Well, I didn't say that it was something anyone in Free
    Taran> Software ever said what you are saying. So I do not
    Taran> understand where part of your response came from. I wasn't
    Taran> even talking about Stallman other than a reference to him
    Taran> in the history of the matter.

I simply meant to address your claim that FLOSS people talk about
copyright and patents, but neglect trade secrets.  My point is that
FLOSS advocates are aware of trade secrets, they're just not
particularly of importance to the movement in my opinion, and I
advance that as the reason why they aren't discussed so much.

    Taran> My point was that Trade Secrets have been an important part
    Taran> of proprietary software, and are no longer as much so.

This isn't due to FLOSS.  This is due to the fact that most functions
can be implemented by non-geniuses now, just by looking at the
behavior.  And on the other hand it has become well known just how
easy it is to get software patents.

I also think you underestimate the degree to which companies still
can rely on simple secrecy and obfuscation.

    Taran> Even if I can tell you what a black box does, that doesn't
    Taran> make what happens in the black box a trade secret. The
    Taran> trade secret is *how* something is done.

But that's my point; for most software, _what_ is done is sufficient
to give even a mediocre programmer enough clue as to _how_ to do it.
As Martin McLuhan told us 40 years ago, the medium _is_ the message.

That's why SourceForge has only 42 distinct projects, each of which is
implemented 1500 times---and 500 times more on Savannah.  ;-)

    Taran> Patents slay trade secrets in this way because the cost of
    Taran> entry to compete may also include the cost of using a
    Taran> patent - financial or otherwise.

And the cost of entering in the presence of a trade secret is the cost
of redeveloping the technique from scratch.  Everything has its costs.

    Taran> Working forward through a patent - not by historical
    Taran> significance, but how it is used - can hamper competition
    Taran> because the entry to compete has another added barrier,
    Taran> guarded by lawyers with sharp pens.

Not _can_ hamper competition.  _Does_ hamper competition _by design_.
A patent that prevents FLOSS developers from distributing their
independently developed programs is doing _exactly_ what it is
designed to do.

If this causes other developers to go find their own new ideas, thus
doubling the rate of _forward_ progress, it's a win all around (except
for the alleged injury to natural rights).  Whether it actually works
that way is an empirical question.

    Taran> But with a patent, they have to pay a license fee to even
    Taran> use the same process. Plus supporting governmental
    Taran> institutions that handle the legalities and logistics of
    Taran> the framework. All because the first baker got a patent on
    Taran> a process.

Yes.  This is a good thing, if you abstract from the transactions
costs.  The problem with software patents (and probably patents in
general) is that they impose huge transactions costs for mediocre
benefit in terms of market-mediated reuse.  (Again I'm ignoring the
alleged injury to natural rights.)  Wholesale, indiscriminate
cross-licensing of whole portfolios is evidence of that.

    Taran> Trade Secret: Getting rich over a period of time because of
    Taran> consistently thinking of better ways to do things.

This is clearly unbalanced.  A better way of doing things is a
_durable good_, an _asset_.  Someone who has published the first
public key algorithm has given a gift that keeps on giving _forever_.
Shouldn't that person receive income from that work as long as others
find it useful?

    Taran> Patent: Getting rich by having only one good idea and
    Taran> getting the framework in place.

Yup.  The reward matches the contribution.  Just _what_ is wrong with
that?  Remember, every dollar paid in this model is paid voluntarily;
not buying is always an option.

The problem with patents is not that they restrain competition; that's
what they're for!  The problem with patents is that for various
reasons they simply don't produce the benefits in practice that naive
theory hopes they might.  (And the issue of freedom, but I'm
restricting myself to the economics here.)

    Taran> Are there any *individual* holders of software patents? 
    Taran> Flesh and blood versus legal entities? I doubt it, and that
    Taran> lack could be seen as evidence of how patents stifle
    Taran> competition.

I'm sure there are many.  True, the RSA software patent expired.
However, I know of several individuals with business method patent
applications in progress.  Of course they're all incorporated, but
that's a device to limit liability---they're all one-to-three person
firms.  I bet there are several people on this list who hold software
patents (I'm equally they're all freely licensed :-).

    Taran> No, but I'm fairly certain that GNU expedited what we see
    Taran> here today. If companies were comfortable with their trade
    Taran> secrets as they were, we probably wouldn't be where we are
    Taran> now. GNU sort of added fuel to the fire.

Sure, but the fact that GNU is open source, or that it harbors all
these radical free software advocates, is irrelevant, I think.  What
GNU really signifies is the dramatic drop in the cost of going from a
specification to a fairly competitive product.  GNU simply made it
impossible for small proprietary businesses to get rich on average.
(You can have a high income by working _very_ hard, but being first in
gives you very little advantage after a few months.)  But if it hadn't
been for GNU, small and medium size proprietary businesses would have
competed the profit out of much the same market segments as GNU has
done.  Just a little more slowly, and they all would have made more

    Taran> Not that it really matters too much. The idea was to mix
    Taran> the contexts; I do believe that FLOS has influenced the way
    Taran> proprietary companies look at patents, copyrights,
    Taran> trademarks and - trade secrets.

I'm sure it has, but I doubt that it's a major player compared to the
decreasing cost of replicating external functionality in new software.

Whether it has had a major i