Subject: Re: "incentive void" (was Re: A different patent covenant...)
From: "Ben Tilly" <>
Date: Thu, 5 Oct 2006 11:39:17 -0700

On 10/5/06, <> wrote:
> Ben Tilly writes:
>  > Please stop trying to guess why I thought so and actually **ASK** if
>  > you care.
> My bad.
> Given that you refuse to claim qualification to judge, my curiosity is
> *really* piqued.  Why *did* you think that the Cameron technology is
> "oh-so-obvious and has been forever"?  (Please note that those words
> are the syntactic referent of your word "so".  Presumably you didn't
> quite mean that, but I'll let you disambiguate.)

That's like asking if I've stopped beating my wife.

For the record, I never thought that, and never said I thought that.
I challenge you to go back and find where you think I said that.  I
believe you will find that you were attributing someone else's
comments to me.

>  > No, I don't understand why I'd have to concede anything in fairness.
>  > I'm not redefining prior art.  I'm saying that if your implementation
>  > depends in an essential way on technologies that have been on the
>  > market less than 10 years, then you won't be granted the patent at
>  > all.  It is an additional bar that patents would have to pass.
> You'e using the same basic concept of "relevant existing technology",
> defining it broadly when a broad definition makes it easy to support
> denial, and defining it narrowly when a narrow definition makes it
> easy to support denial.  That's true whether or not you use the same
> term, and any patent advocate with a room-temperature IQ will call you
> on it.

I'm wondering where you're getting this from.  Incidentally I have to
say that I'd be inclined to say that a patent advocate who was calling
me on this probably has a room-temperature IQ. :-P

AFAICS the question, "Could someone have built this 10 years ago?" is
far less subjective than the question, "Is this obvious to an
experienced practitioner?"  There is far less room for debate and
opinion, and far more for concrete facts.

Furthermore I have to point out that you're re-writing what I was
saying in a very unfair way.  I did *not* say anything nearly as vague
as, "relevant existing technology".  I've phrased it differently each
time I've said it, but it has always boiled down to, "Has this
invention been possible for at least a decade?"

>  > The justification for this bar is that with a new technology,
>  > obviousness is too difficult for examiners and judges to fairly
>  > judge.
> That's not a justification for *this* bar, that's a justification for
> finding a better way to judge "obviousness".  *This* bar is only
> indirectly connected to obviousness.

People have been trying to judge obviousness for a very long time,
aren't doing very well, and we aren't making significant improvements.
 In fact, judging from the complaint level, we're going the wrong way.
 Isn't it worthwhile to consider another approach?

> The connection you draw is plausible, but I have to wonder if ten
> years is the appropriate interval in a field like software where the
> whole stock of software turns over on average 2.5 times in that
> period.  (That number is by memory, originally from the U.S. Commerce
> Department and the former Japanese Ministry of International Trade and
> Industry.)

Similarly one can ask whether 20 years is the appropriate length of a
patent grant in a field like software when software turns over that
quickly, and companies often do not survive their t-shirts.

The two arguments are connected in my mind.  I picked 10 years as half
the length of a patent grant.  I felt that if the idea hasn't been
produced in that time, then there is a reasonable chance that it is
sufficiently obvious that it won't be produced in the next 20 years,
and therefore it makes sense for government to provide incentives to
encourage its production.  If the length of a patent grant was
shorter, then a shorter period than 10 years would sense.  I would
even happy with a version of my proposal that allowed patents on
things that had been invented more recently than 10 years, but which
would reduce the length of validity on those patents.

Whichever version you pick would make the patent system be less likely
to cause chilling effects on new areas of innovation.  Yet you'd still
provide incentives to come up with ideas that otherwise people would
not come up with.

> It's also not obvious to me that obviousness is monotonic in time.
> Specifically, I would guess that any given dependent concept becomes
> progressively easier to see for a few years until the "conventional
> wisdom" congeals, after which it become progressively harder to see.
> (I'm not sure whether this has anything to say about the value of your
> idea for regulating patents.)

That is a good point.

However let me try another line of reasoning.  Patents are monopolies
created by the government to provide an additional incentive to find
good ideas that the market might not find on its own.  In other words
it is a government intervention to correct a potential market failure.

My suggested change follows the principle of, "Let's not have
government intervene until the market has demonstrated that it has
failed."  If the market is healthy and bubbling with good ideas
without government intervention, then there is no need for government
to intervene.  Government intervention (with both its positives and
its negatives) would only start to matter when the market hasn't come
up with an idea in a long time.

>  > As you can see this criteria is entirely separate from prior art.
> Sorry, I don't see it.  It uses the same basic concept.

Prior art says that you can't get a monopoly for inventing this idea
because someone else invented it before you.  My idea says that you
can't get a monopoly because it is likely that someone else would have
independently invented it.

>  > > There is no shame in just wanting patents to go away.  That's an
>  > > honest, and perhaps the most correct, point of view.  Why not stick to
>  > > that?
>  >
>  > Because there are patents that I think are good.
>  >
>  > What the patent system is supposed to do is create incentives for
>  > ideas which are obvious once you see them, but which otherwise might
>  > not come to market.  An example of a patent that I like is
>  >
> Of course, by the ten-year criterion it doesn't apply to blue and
> white flashes (Nakamura didn't invent the blue LED until 1993, and
> didn't become "widely available" until the late 1990s).  How does your
> criterion deal with that?  (That's a real question.)

I see no issue.  A patent covers whatever it says it covers, and not
just your original invention.

Suppose you get the patent on flashing shoes.  Technology improves
such that you can make prettier flashes.  You still have the patent on
flashing shoes, so you are the only one who is allowed to make flashes
of either kind.

In short, a patent holder is allowed to sell improved versions of his
invention.  Everyone else needs his permission to sell any versions of
his invention, even if those versions are better than anything he
could have produced at the time the patent was issued.

> I have to wonder whether it applies to any other color, either.  Were
> any LEDs of sufficient brightness to produce "flashes" when embedded
> in plastic of the right strength etc to serve as a component of shoe
> sole available in 1988?[1]

That is a good question.  I know that the first commercially viable
visible LEDs were invented in 1962 by Nick Holonyak Jr, and I know
that a decade later M. George Craford invented the far brighter
yellow, red, and red-orange LEDs.  So the LED technology was certainly
there.  I do not know for sure whether they had sufficiently good
plastics, etc to technically build the shoe in 1988, but I think they

BTW this patent brings to mind another patent that is vaguely similar.
 See  I think it is
clear that materials were sufficient for building that long before it
was invented.

>  > You'll note that I came up with a suggestion for a non-adversarial
>  > procedure that I think significantly improves the status quo in this
>  > regard.
> Sure.  But "non-adversarial" is not an advantage AFAICS, it's an
> independent property that a process might have.

We have to disagree on this one.

Adversarial processes have a number of issues, mostly centered around
the cost of having adversaries confront each other.

For instance with the current patent system suppose you're approached
by a patent holder who is asking for a license from you for a mere
$100,000/year.  You look at their patent, decide it is BS, and think
that you have a 70% chance of beating them in court.  But if you take
them to court and lose it will cost you millions.  Even though you
think their patent is BS, you're probably going to fork over the

This is not a hypothetical example.  I've seen estimates that over 70%
of software patents would not hold up in court.  But the invalid ones
are still worthwhile because nobody wants to pay the cost of
litigating them.  And it is common practice for patent trolls to go
for a lot of small fees from lots of companies knowing that even
though they'd probably lose, it isn't going to be worthwhile for any
individual company to call them on it.

This is not to say that adversarial processes are to be avoided.  They
have their place.  But they also have shortcomings.

> In fact, my feeling is that adversarial processes have a lot of
> advantages here, as long as they're public.  (By obvious analogy to
> academia, that being the innovation process I know best.)

In academia the process works well.  In part because it doesn't cost
as much to write a paper criticizing someone else's research as it
does to take someone to court.

> Footnotes:
> [1]  And I know of "prior art" for claim 1 that dates at least to my
> college days, making it pre-1979.  Cf. the selection "Skinny" from
> Alex Bevan's Springboard album:
> :-)

I don't have sound enabled on my laptop, nor can I find lyrics. Would
you mind explaining the reference?