Subject: Re: GIF/LZW patent
From: <stephen@xemacs.org>
Date: Thu, 28 Sep 2006 11:49:58 +0900

Thomas Lord writes:

 > No, but nice try.   It's a little hard to convey why it doesn't work but
 > I'll try.

You miss Larry's whole point.  Patents create friction around the
innovation.  Copyright creates friction around the whole work, most of
which is *not* innovative in the same sense.

It is the balance between the incentives created by monopoly rents due
to friction and the costs of the friction that we're interested in (if
we're talking about the economics arguments for/against the policy).
In a world of perfect information, the friction of patents is pinpoint
and therefore small, while that of copyrights is diffuse and large.

The real problem with Larry's argument is that to make the friction of
copyright go away, you just close your eyes.  Ignorance *is* an
excuse.  With patent, however, ignorance is no excuse.

Nor is there perfect information, which leads to the "chilling
effect."  Nobody is completely ignorant of patents: they know patents
exist, but they don't know which techniques are involved.  That means
that like a quantum wavefunction, the friction to *each* patent is
everywhere in the software world.  This starts to look like copyright
again from the point of view of incentive effects.

But maybe it's worse, due to the difference in treatment of
inadvertant infringement due to ignorance.

 >    SCM uses "three bit tags in object references"* and so eight-byte
 >    aligns objects, "all of which are represented as 8-byte objects"* so
 >    cons-pairs are very compact and the main allocator is simplified.
 >    Within objects, are "additional variable-width tag bit sequences"*
 >    and so dispatch in a graph-code based 'eval' is simple and pretty
 >    quick.  The result is that there's enough tag space to use a bunch
 >    of types as opcodes and so 'eval' does "graph-rewriting to incrementally
 >    compile input forms"*.   This works out to a simple enough `eval'
 >    that its code fits easily within a typical L1 cache which gives
 >    it some interesting performance characteristics.
 > 
 > Every one of the ""* terms indicates a technical choice that was well
 > understood more than two decades ago.    The combination of these
 > little choices "harmonizes" well -- for example, the way the allocator
 > and 'eval' are both simplified by the tagging choices is one of the
 > beauties of this particular implementation.    Thousands of additional
 > little choices add to the harmony.  
 > 
 > And that's the problem.   With a patent, one could *try* to characterize
 > the chord.   SCM's author was the first one to put all of these elements
 > in just exactly this way.    But there's no way to formulate a useful
 > claim.

Looks to me like you just did.

You talk about thousands of choices, and I'm sure you're right.  With
copyright, SCM, Inc., gets to monopolize all of them in one fell swoop
without any cost to them.  But the goal is not to give SCM a monopoly
so they get rich; the goal is to give them a monopoly for
contributing.

So you just wrote one claim for them, and it took effort.  They're
going to have to abstract those thousands of interrelated choices into
some number of claims, possibly in different patent applications.  At
some point they're going to say, "oh, hell, just take it" (the "Friedl
License" ;-) for the rest.

So not only do we know they belief that "this is an important
contribution", they also have to tell us in some detail *what* they
think is important.  We can either work around their claims, or we can
license just the ones we need, and there's *no negotiation about scope
involved* (in theory), it's defined by the patent actually awarded.

With copyright, sure you can say "we want to license from line 528 to
10016" but this is going to be much more complex to negotiate.