Subject: Re: patent trolls and X-licensors
From: "Ben Tilly" <btilly@gmail.com>
Date: Fri, 9 Jun 2006 17:02:12 -0700

On 6/9/06, Stephen J. Turnbull <turnbull@sk.tsukuba.ac.jp> wrote:
> >>>>> "Ben" == Ben Tilly <btilly@gmail.com> writes:
>
>     Ben> I do, I do, and in that case I expect to have to explain
>     Ben> myself in more detail. :-)
>
> I've greatly enjoyed and benefited from those explanations in the
> past.  Thank you very much!

:-)

[...]
>     Ben> [Patented techniques] can't get implemented in free software.
>
>     >> That's a definitional problem.
>
>     Ben> What do you mean by a definitional problem?
>
> By definition, free software allows anyone to use it.  If you don't
> own a patent whose claims are embodied in your software, you can't
> apply a (truly) free license to it.  The GPL itself says that if you
> know about a patent you must arrange for it to be licensed for use in
> all circumstances where the software can be used (ie, use and use of
> derivatives), or you may not redistribute.

Ah yes, absolutely true.  It is what I was thinking, but I would never
have called it "definitional".

> BTW, IMO, that means that all GPL-licensed software, except that
> distributed by copyright holders and IBM, is at minimum irresponsible
> and probably fradulent.  At least if you take the "anything is
> patentable and probably has been patented" rhetoric seriously.

Even that distributed by copyright holders is open to question since
the copyright holder has no idea what patents are out there.  However
as a practical matter, I'm a firm believer that some things only work
because people know what laws they need to ignore.

>     Ben> If I've patented my idea, and you know about it, you're going
>     Ben> to be less inclined to put it into a standard unless I'm
>     Ben> willing to agree to terms that mean that people can actually
>     Ben> use my patent.
>
> Eh, I think you're cheating a bit.  There's no point in creating a
> standard at all if people can't actually use it!  Stripped of the
> hyperbole, what exactly are you claiming?

I was trying to state a logically trivial statement.  You caught me. :-)

My point is exactly that.  People who are trying to create standards
are generally not trying or sympathetic to people who are trying to
hijack that standard with a patent.  However there is no shortage of
patent holders who want to successfully hijack that standard for their
own benefit.  (eg: RAMBUS.)

The point if therefore obvious.  But not devoid of content because
this is a game that does go on all of the time.

>     Ben> Even if I say that I I'm willing to license under reasonable
>     Ben> and non-discriminatory terms (RAND), a lot of people are
>     Ben> going to complain.  One of the big reasons why they will
>     Ben> complain is because licensing fees mean that people can't
>     Ben> implement the patent in free software.
>
> If you mean that they can't do it because that would amount to a
> prohibited royalty on the software, see "definitional problem".

Exactly.

> If you mean that they can't do it because people object to paying for
> value received, those people are beneath my consideration.  Well, not
> completely; let me gloss that a bit.  [...]
>
> If you mean that they can't do it because there are no social
> mechanisms in place for collecting revenue on free software and paying
> the patent-holder, well, what are we waiting for?  Let's get cracking
> on designing those systems and comparing the costs and benefits of the
> various alternatives.  One of which is abolishing patents; another of
> which, I must remind you, is abolishing free software as we know it.

The kinds of barriers that this would introduce run exactly counter to
the ideals of the open source philosophy.  Which means that they also
run exactly counter to the theory of value embodied in open source.
Good luck reconciling those two.

In fact I strongly suspect that "abolishing free software as we know
it" is the only way to reconcile those two while raising those
barriers.

> For example, how about socializing only the *intellectual property
> rights*.  Anybody who gets value from contemplating and developing the
> intellectual side of software is allowed to do so freely, but control
> over economic rights stemming from actual use could be privatized.
> There would be a fair use exemption for embodying patents for the
> purposes of working out those new ideas.[1]

/me whispers "transaction costs"

I'll explain that comment in a second.

> I'm sure everybody on this list is horrified by the very idea, but how
> do you think non-developers would react?  "That's how it works for
> *me*, anyway" seems likely to be the most common reaction, except for
> people involved in P2P "sharing" of consumer goods (goods that even
> rms won't say are inappropriate objects of copyright), who will be
> horrified by the idea of actually paying money for something they're
> used to getting for free.

Non-developers won't understand the issue.  Until the CTO is
approached by a patent holder demanding an audit of the Linux servers
in use that the CTO didn't know about.  Afterwords most non-developers
won't understand the issue.  They also won't understand why IT got
that much worse.

What this would do is introduce a huge legal risk for using free
software.  With proprietary software, a patent holder can go to the
distributer and go after them for licensing fees.  One negotiation,
fairly limited transaction costs.  With free software you're pushing
the patent holder to go after customers one by one.  This is a
significant transaction cost per customers, for both sides.  If patent
holders actually started doing this, you'd get a substantial legal
risk just for using free software.  Once legal understood this, the
result would be an end to the use of free software.

Now I'm not saying that this conclusion would be wrong.  There is no
question that free software and patents are on a fairly inevitable
collision course.  At some point one or the other will have to give.
(I believe that the current state where there are lots of patents that
conflict with free software but few do anything about it is unstable.)
 However my theory of how value is generated by software says that
free software is more important than patents.  Therefore I'm against
free software losing to patent law.

[...]
>     Ben> The [MP3] standard became widely adopted and THEN they
>     Ben> started enforcing the patents.  It is exactly examples like
>     Ben> that which make people leery of patents in standards.
>
> Really, anybody who's surprised by that timing should think again (or
> more likely, should start thinking for the first time).  Enforcing
> patents is not cheap; it's only going to be done when the revenue
> stream being threatened is mature, and defense is more profitable than
> growth.

The reasons for the timing are economically obvious.  My point was not
that the timing is surprising, it is that the success of the patented
MP3 format can't be taken as evidence that people don't try to avoid
patents.

[...]
> There's also the problem of the boycott of proprietary software
> proclaimed by the GNU Project, which is a very big factor in Linux
> users' consumption habits.  I wouldn't be surprised if you start to
> see growth in patented software for Linux as the population of Linux
> users with more money than brains and/or free software scruples grows.

For me, personally, the issue is what Ubuntu can readily support.  And
that isn't an abstract boycott.  That is practical experience saying,
"Stuff works better if I stick to what they can support, and what they
can support is only stuff that is not encumbered by nasty patents."
Considering that Ubuntu is one of the fastest growing user desktops in
the Linux world, I would expect to see problems continue going
forward.  (Yes, there are ways to enable a ton of audio and video
applications, largely by downloading from European sites that ignore
the patents which afflect the USA.  However I've found that when you
do that, stuff tends to break after a couple more updates...)

[...]
>     Ben> Odd, I'd thought that you'd be biased because you were a fan
>     Ben> of XEmacs and that whole Emacs vs XEmacs split.  Which
>     Ben> happened a while ago.
>
> Well, I have to admit that I think Emacs would be a lot better today
> if rms had let Lucid install its improvements into the main line, and
> *then* got to work on the things that were important to him.  But I
> didn't think it really mattered; there was plenty of room in Emacs
> space for a bunch of on-the-edge types playing with Mule and eye
> candy, and another bunch doing the deadly serious world's-best-editor
> shtick.

I think that Emacs would be better today if it had lexical scoping...

> Unfortunately, today the Emacs community simply is uninterested in
> doing any of the things needed to keep it relevant to the IDE market
> (outside of its historical niche, of course).  That's important
> because Emacs is too big to be "just an editor."  Emacs is still the
> best text editor available, but it no longer clearly justifies the
> effort of learning Emacs, unless you're willing to buy into some of
> the major applications (eg, MUAs) at the same time.  Not everybody is,
> and I sure can't blame them.

I'm not.  I use vi because I find it easier on my fingers.

[...]
> And there's no excuse for any of that.  Lisp *can* compete with Python
> (at least for veteran Lispers), but Emacs Lisp cannot---nobody in
> their right mind writes batch scripts in Emacs Lisp (except for
> maintaining Emacs).  Lucid Emacs and Energize *could* have given Emacs
> a 10 year head start on Eclipse, but seems to me we're about 5 years
> *behind*.  In 1995 the only browser that could properly handle the
> Japanese web was w3.el on a Mule-patched Emacs, but today w3.el can't
> handle the modern web at all----lynx does better.
>
> You bet I think it would all be different if a company (or consortium)
> with paying customers had been funding the Emacs mainline for the last
> 15 years.  But until Hrvoje left, I didn't really think it much
> mattered.  The various Emacsen would survive, and prosper.  But Hrvoje
> forced me to take a serious look at what the Emacs community was
> doing, and not doing.  I see no real way to avoid the conclusion that
> Emacs has connived at making itself quite irrelevant to the great
> majority of programmers (other than a reasonable option as an editor,
> and many would not even concede that), and current policies are going
> to make the gaps described above grow rapidly.

I don't think that a company with paying customers would make the
difference that you're hoping.  For a somewhat equivalent product I'd
look at something like VisualWorks Smalltalk.  Both were first class
programming environments.  Both have very loyal communities.  Of the
two, I'd say that Emacs had and continues to have the larger impact.

Actually that comparison is unduly fair to the proprietary product
since I chose one that has continued to survive to the present.  I
don't think that a proprietary version of Emacs would have been likely
to survive so long.

When you're talking about the evolution of a programming environment
over 15 years, well 15 years is a very long time in the software
world.  A commercial backer can't be depended on to last that long.
Things change.  If I'm still working using a Perl/Apache/Linux/Oracle
combination in 15 years then I'll be a dinosaur and will expect to be
treated as such.  (Actually I'm already a bit of a dinosaur to use
that combination.)

> Those policies are mostly rooted in GNU's political goals, and the
> enormous conservatism of Emacs hackers.
>
>     Ben> However I'll note that I have yet to hear of a real case of
>     Ben> someone who needed to solve a technical problem choosing to
>     Ben> solve it by doing a patent search looking for a previous
>     Ben> solution.
>
>     >> Don't you think that's the saddest thing you've written today?
>
>     Ben> Not particularly.  It is the wrong tool for the job.
>
> That's what I consider sad!

Once again, I fail to see what you see as sad about this situation.

>     Ben> What distinguishes a good business from a great one is
>     Ben> execution.  Which depends on a ton of small details.  For
>     Ben> instance one recent subtle interface tweak increased our
>     Ben> profit margin by over 20%.  You could spend weeks analyzing
>     Ben> our site and not even notice the change that is responsible.
>     Ben> We could face a direct competitor without that tweak and bid
>     Ben> up their cost of advertising to the point where we make money
>     Ben> and they don't.
>
> And if you patented it, you could give it to them (at a price), and
> both of you keep your costs down.  But you clearly don't plan to give
> it to anybody, or even post an URL here and ask us to look for it
> ourselves.  Right?

Sorry, you walked right into  this one. :-)

It really isn't patentable.  I'll tell you what it was.  We changed
the subject line of an email.  This is not a patentable change.
Neither the old nor the new subject lines are trademarkable.  They are
both too short for copyright to matter.  The insight that lead us to
change the subject line was not particularly new or unobvious.

We would have no interest in licensing this change to a competitor -
because after all what would keep them from getting our good idea and
finding a tweak that improves business another 10%?  Then *we'd* be at
a huge disadvantage!  (Well not really, because we have tons of other
significant yet almost invisible tweaks on our site that they
wouldn't...)

Hopefully that clarifies my point.  You could take every single
potentially patentable idea that we have, and still lose money hand
over fist.  That tells me that the ideas that we could patent are NOT
what is valuable about our business.  Conversely if we needed that
patent protection to remain on top against normal competition, then I
submit that we're doing a bad job and society would be better off if
someone else ate our lunch.

[...]
> Granted, in computer science it's harder to disentangle by field, but
> it seems to me that the social effects are just as important.  If you
> start splitting computer science departments into IT groups in
> engineering schools and business schools, and computer science
> departments proper, in whatever division your university keeps math
> and physics, I suspect the basic researchers in CS proper would feel
> only slightly more pressure to patent than I do.

And the gap between basic research and applied gets wider...

[...]
> Everywhere I look (not just in computers), I see losses from "tweaks"
> and "best practices" that never get propagated because they're either
> trade secrets or simply because nobody's looking for them.  I may be
> biased; "unused tweak" losses are rampant in university teaching, and
> absolutely huge in Japanese universities.

Unused tweak in university teaching.  Here's a simple one.  Assign
cumulative homework sets.

That is on any given day have the easiest third of your homework be
about today's material, then middle third be about stuff from the last
week, and the hardest third be from any point in the class.  This
forces people to constantly review the course, which puts a lot more
of it into long-term memory - very helpful in a cumulative subject
like, say, mathematics.

I've suggested this to multiple professors.  Several tried it and like
the results.  I've never heard of anyone doing it who I haven't
suggested it to...

> I think they're big enough to justify some strong forms of
> intellectual property.  Not the patent system current in the U.S., but
> it might look more like that than like no-patents-at-all.

But that's natural.  A "best practice" that you know about and others
don't is a competitive advantage.  Take the example I offered
involving my employer.  Under what possible licensing scheme would
work to license what we've learned about what a subject line change
can do?  It is clearly worth a lot to us.  It would be virtually
impossible to tell if someone else stole our idea or came up with it
independently.  (Particularly not if they did it better than us.)  It
is hard for any given company to find out whether the idea is worth
anything at all to them.  It is probably worth either a lot less or a
lot more than it is for us.  And we stand to lose a lot if someone
improves on it and won't let us use the improvement.

I submit that in no sane IP system would we license the idea.

[...]
> My point is that free software all too often results in "solutions"
> that are unusable unless you have more brains than money, and probably
> more time on your hands than both money and brains put together.
> IMHO, society should be arranging the rules for these markets to
> benefit those who have more money than they have brains, and more
> brains than they have time, and not so much of any of the above in the
> first place.

My theory is that by trying to make that alternate rearrangement you
run the risk of killing the goose that lays the golden egg.  But
you're welcome to try, and free software businesses are the ones who
allow people to exchange money for a reduction in the amount of brains
and time they need to implement the solution.

> Footnotes:
> [1]  For me it would be merely a return to my roots.  When I first
> started working with publically distributed source, that's actually
> what I thought "free software" meant.  I had no trouble classifying
> "no commercial use" licenses like Ghostscript's Aladdin license as
> "free".  They're actually "free"-er than what I would have demanded
> then.

I remember when I felt that way.

Then I got a real job.  And learned to appreciate the value of having
peons be able to sneak useful stuff in "under the radar" in commercial
organizations.  And now I like it when stuff can be used for free
commercially. :-)

Cheers,
Ben