Re: [Patents] US patent reform strategy inaugurated by myths & lies

[Stefano Costa <stefano.costa@icn.siemens.it>]

Erik Josefsson wrote:

> Your concern about limiting the patent system is shared by most
> economists. On the recent meeting in Paris at IFRI also the
law-scholars
> said a limitation to the patent system is both desireable and
necessary
> (the lawyer from IBM was an exception).

My proposal for that derives from the fact that product life span
is getting
shorter, so the 20 years were a (relatevely) long and right timing
20
years ago, but now are too limiting. In proportion, 5 years now are
like 20 then.

> According to the Swedish Patent Office, the Swedish Court
> of Patent Appeals and the Swedish Supreme Administrative
> Court, this apparatus is not an invention because it does not
solve a
> technical problem (the problem how to organize a queue is solved
with a
> calculation rule).

I don't want to go in specific problems, but only on general
issues.
I know by personal experience that whatever rule you may set in,
inevitably some problems occurs. Saying that, a shorter life span
will mitigate problems like different interpretations.
Keep in mind that this kind of problems occurs for *any* law, not
only patents, because non-formal language (like english) mixed
with deliberate "grey area" statements, will pop up here and there.
Another  reason to avoid to say "Rule A will make unfair treatment
of patent P". Of course it's important to keep track of problems,
so
I don't say that we don't discuss them, simply I say that this kind
of problem is inherent with the system.
I wish to emphasize that this kind of problems occurs on what you
call "good" patents, that is, hardware-related. It is less frequent
than SW-oriented ones of course, but patent litigation is a big
issue even in HW.

> I suggest that denying patent protection of physical realisations
of
> calculation rules would be a possible way to achieve the desired
goal.

The infamous interpretation problem as above is everywhere, even
with
your proposal.
Problem: a rigourous definition of what is hardware and software,
and the fact that the border is blurring.
Example: CPUs, FPGA. Nowadays CPUs are plenty of microcode,
i.e., each asm instruction is decoded in smaller units, that is,
each
asm instruction is a microprogram. The translation is done totally
in HW, but the code is "program". Now, a particular technical
implementation of microprograms is/should/would be patentable?
With your definition no, but is full HW indeed. Also FPGAs have
the same problem: physical devices, but with both HW/SW role.
So my final consideration is "Keep it easy, make it shorter".
Because it's a dead end riddled with an enourmous list of
exceptions/
rules, counter rules, that is with no end, and thus useless.

> The swedish interpretation of the law limits the patent system,
the EPO
> interpretation does not.
>
> Which one would you like to endorse?

Both, if you have read above.

Stefano