MR. LAURIE:  Thank you, Mr. Secretary.

I'd  like  to  address  the  issue  of  patent   protection   for
software_rped inventions generally rather than the specific ques-
tions raised in the Notice.

The views expressed are personal, they're my own, and they  don't
necessarily  represent  the position of either my law firm or any
particular client.

By way of introduction, I worked in Silicon Valley for 33  years,
initially  as  a  programmer and software designer and later as a
patent lawyer focusing on computer technology.  I currently teach
a  course  at Stanford Law School entitled "Intellectual Property
Protection for Information Technologies".

I have previously served  on  advisory  panels  to  the  National
Research  Council  and the National Academy of Science in connec-
tion with software protection, and I was the only  patent  lawyer
on  the  Advisory Panel to the Office of Technology Assessment in
its recent study "Intellectual Property Protection for Software".

When one listens carefully to the impassioned  arguments  against
quote  "patenting software" unquote, it becomes apparent that the
arguments and the basic intellectual propetpolicy positions which
underlie them can classified into three categories:

First position:  patents are bad.  Second position: software  pa-
tents are bad.  Third position:  bad software patents are bad.

The first position is most often heard in the halls  of  academia
and  raises  fundamental  social and economic issues which go far
beyond the scope of the present inquiry.

The second position is often advanced by some, though not by  any
measure  all,  of  the software companies that emerged and flour-
ished during the early and mid '8Os as a  direct  result  of  the
commercial  introduction of the microprocessor at a time when the
industry generally and incorrectly believed  that  software_based
inventions were unpatentable as a class.

The third position is the one most widely held today and for  ob-
vious reasons the easiest to defend.

In the heat of the debate over software patents,  the  boundaries
between the second and third positions tend to blur, but I submit
it is critically important to address them separately.

I xectfully submit that the second position  is  legally  unsound
and  that the third while correct does not represent an insoluble
problem.  In support of my thesis, I offer the following:

Premise one:  U.S. patent law does not protect software.  Rather,
it  protects  processes  and  machines that are quote "within the
technological arts" unquote.  In one of the first cases  to  con-
sider  the question of whether inventions involving computer pro-
grams constituted patentable subject matter,  Judge  Giles  Rich,
who  was  one  of  the principal architects of the current patent
statute, introduced  the  phrase,  "technological  arts"  as  the
modern  equivalent  of the constitutional term "useful arts," and
is therefore defining the outer boundaries of patentable  subject
matter,  both under the Constitution and under Section 101 of the
Patent Act.  That case was, In re Musgrave, decided in 1970.

Over the more that two decades since Musgrave, Judge Rich's  for-
mulation  has  remained  unchallenged by any subsequent decision,
although unfor|tely it has been ignored by many. Thus any process
that  is  not  sufficiently  applied  the physical environment in
which it operates to qualify as being quote "within  the  techno-
logical arts" unquote, constitutes unpatentable subject matter.

The critical distinction then is between applied  technology  and
abstract  ideas.  Examples of the latter include; laws of nature,
scientific principles, methods of doing business, printed  matter
and unapplied mathematical relationships.

Premise two:  Computer  implemented  solutions  to  technological
problems in the form of processes and/or machines typically exist
along a design spectrum, ranging from pure hardware, that is ran-
dom logic, to pure software, that is an externally_loaded comput-
er program running on a general purpose digital computer.

Intermediate points along the spectrum involve designs which  may
be  described as special purpose computers and which combine ele-
ments of hardware and software in varying proportions, using ran-
dom  logic,  array  logic,  such  as  PLAs  PALs,  microcode  and
firmware,  firmware  being  fixed  programs  stored  in  internal
read_only memory.

The particular point along the design  spectrum  that  represents
the  optimum  solution  to  a  given  problem  is determined by a
variety of factors, such as cost, speed, size, flexibility and so
on.   Moreover,  the optimum design point moves over time as com-
peting implementation technologies evolve at different rates. For
example,  in  the  mid '70s, complex video game functionality was
implemented entirely in random logic.  After the arrival  of  the
microprocessor,  the  very  same functionality was realized using

Finally, technologies such as logic synthesis are becoming avail-
able, by which a software solution can be quote "translated," un-
quote into an equivalent hardware solution, and vice  versa.   It
should  be self evident that as a matter of legal policy, the law
should not promote artificial distinctions  that  the  technology
does not recognize.

And I should point out that Mr. Lippe's company is in the businef
making  a  product  which  in  effect  translates  software  into
hardware.  Another example which I think illustrates the point is
the  technology of neural nets, which was originally created as a
pure hardware solution and has evolved now into a software  tech-

Premise three:  The fact that a particular solution  can  be  ex-
pressed  mathematically  or  is  a  series  of logical operations
should be irrelevant to the patentability of the solution.

In 1972, based on what many commentators believe to be an errone-
ous  interpretation  of its prior decisions involving laws of na-
ture and scientific principles, the U.S. Supreme Court announced,
in  Benson v. Gottschalk that a patent claim describing a process
which, quote, wholly preempts a mathematical algorithm is nonsta-
tutory;  that is, does not define patentable subject matter under
Section 101 of the Patent Act.

The result of this formulation has been over two decades of  con-
fusion  and  inconsistency in the case law involving the patenta-
bility of software_implemented prses.  The fact is that  mathemat-
ics  is  a  language,  albeit  a very precise one, and like other
languages can be used to describe concepts and relationships that
are  technologically  applied as well as those of a more abstract
nature that are not so applied.

As noted by Professor Chisholm in an article called The  Patenta-
bility  of Algorithms, the real issue is probably not one of sub-
ject matter under Section  101,  but  rather  one  of  indefinite
claiming of the invention under Section 112.  Under the constitu-
tional standard within the technological arts, it is the  subject
matter  of  the invention and not the language chosen to describe
it that should determine the presence or  absence  of  patentable
subject matter.

Premise Four:  Even if a particular software equipment and  solu-
tion  represents  patentable  subject matter, in order to justify
the exclusionary benefits conferred by a  patent,  it  must  also
pass the test of novelty and nonobviousness over the prior arts.

And Commissioner, you have pointed this out to severf the  speak-
ers, that there is significant difference between the patentabil-
ity of software as a class and the patentability of any  particu-
lar software invention.

This is the key factor that interrelates the second and third po-
sitions,  i.e.,  software patents are bad versus bad software pa-
tents are bad; that is, even if a software  implemented  solution
is  sufficiently technologically applied to pass muster under the
statutory subject matter test, in order  to  quality  for  patent
protection,  the  solution must also be novel and nonobvious to a
person of ordinary skill in the art.

It is submitted that given the objective to be  accomplished  and
accepted principles of software design, the great majority of the
software written today would not pass  the  nonobviousness  test.
Thus,  the  effectiveness of a patent system in a particular area
of technology is directly related to the degree to which the exa-
mining authority __ in this case the Patent & Trademark Office __
has access to the most relevant prior art.  To theent that  there
are  or  can  be  created  mechanisms  through which the Patent &
Trademark Office can access the widest body  of  software_related
prior art, the system will work.

A number of such mechanisms  have  been  discussed  during  these
hearings,  and  they  include PTO access to the growing number of
commercial and public databases of software  technology,  private
sector  assistance  in  supplementing  the PTO internal database,
early publication of patent applications coupled with third party
submission of prior art.

The important point is that the problem of bad  software  patents
is mechanical and not inherent.  That is, over time it can be en-
gineered away or at least reduced to a commercially tolerable er-
ror rate.

Finally, Premise Five:  A very heavy burden of persuasion  should
be placed on anyone who advocates that a particular kind of tech-
nology should be exempted from the normal operation of the patent

In 1980, in the Chakrabarty case, the U.S. Supreme  Court  inter-
preted  the  patent  copyright clause of the U.S. Constitution to
require that the scope of patentable subject matter should be  as
broad as possible __ anything under the sun that is made by man.

Those who maintain that software based invention  should  be  ex-
cluded  as  a class from patent protection argue that software is
different.  It's different, they argue, in terms of its essential
character  __  it's  logical.   It's  different  in  terms of the
creative process by which it comes into being  __  it's  authored
rather than engineered.  Or it's different in terms of the under-
lying economic model governing its production,  distribution  and
life cycle.

These differences have been discussed and debated  at  gatherings
of   distinguished   software  developers,  computer  scientists,
economists and legal scholars and practitioners under  the  spon-
sorship  of  a number of governmental agencies, including the Na-
tional Research Council, the National Academy of Science, the Of-
fice of Technology Assessment, the U.S. Congress and the U.S. Pa-
tent & Trademark Office.

Despite thet that the positions on both sides have been eloquent-
ly  expressed,  the results are inconclusive.  The primary reason
is  that  there  is  no  hard  data  available  to  support   the
anti_software  patent  position, and the evidence is anecdotal at
best.  Clearly, software is different, but is it different enough
from  all  other technologies to justify a special exemption from
the normal operation of the patent laws.

Given the unavailability for reliable data on the societal  costs
and benefits of patenting software_implemented technology, we are
presented with a situation where important policy decisions  must
be  based  on fundamental legal principles. In such a setting, we
must conclude that those who  would  withhold  patent  protection
from  technologically_applied processes and machines, that happen
to be implemented partially or wholly in software, have failed to
satisfy  the burden that the Constitution, the Supreme Court, and
sound legal policy have placed upon it.

Thank you.

COMMISSIONER LEHMAN:  Thank you very much an excellent statement,
Mr. Laurie.

There was something I was going to ask, and I may have to  follow
up  now on it because it slipped my mind.  But I think that was a
good description of __ you parsed out the problem very well.

MR. LAURIE:  If I could address a point that came  up  yesterday,
relating to the role of competition in intellectual property law,
and where the competition is more appropriately  addressed  under
the  anti_trust laws or under the intellectual property laws, I'd
like to say that I think that there is, there are many places  in
intellectual property law where competition plays a role, and the
patent law of misuse is an example, and as shown by the Seiko  v.
(Accolade)  case in the Ninth Circuit, in the copyright law under
fair use, competition plays a very important role.

Thank you.


Next I'd like to ask Lee Patch, the Deputy General Counsel of Sun
Microsystems to come forward.

Back to The San Jose Index

Forward to Lee Patch