MR.  TROESCH:   Good  morning,  Mr.  Commissioner,  distinguished
Panel.  My  name  is  Hans  Troesch.  I'm here speaking on my own
behalf.  My partners have reminded me of that.

Many years ago I earned a Masters Degree in Computer  Science  at
the  University  of Michigan, and for close to ten years, regard-
less of my various and more fancy job  titles,  I  consideLmyself
principally  to  be  a computer programmer. Today I'm an attorney
and a member of the patent bar.  I practice patent, copyright and
trade  secrets law, as we already mentioned, with the law firm of
Fish and Richardson.

I'm here today because I would like to offer my own  views  on  a
few  of the questions that the Patent Office has invited the pub-
lic to address at these hearings. As a preliminary matter, I must
confess  my  own deep concerns about the present fate of software
inventions in the Patent Office.  I believe that the logical, al-
most musical nature of software technology provides unique oppor-
tunities for advocacy and for confusion in a system that is based
on  a  more  structural,  may  I  say more sculptural view of the
world, but that is a topic for another day.

Today I will merely state my hope and belief that the Patent  Of-
fice  will rise to the challenge of finding and keeping qualified
examiners, securing access to the vast body  of  software_related
prior  art  that  is  not  of  record  in  the Patent Pce, and of
developing delimiting doctrines of novelty, obviousness  and  en-
ablement in ways appropriate to the peculiarly flexible genius of
software technology.

I would like to address Question 1 at this point  in  the  Office
Notice,  and to state my view that a computer program, that is to
say, a set of instructions that is executable on a  computer,  to
achieve a result, should be considered a machine within the mean-
ing of Section 101 of the Patent Statute, and should therefore be
eligible  for patent protection, without resort to the additional
and often redundant limitations to computer processors, read_only
memories or data input_output devices.

On the issue of eligibility for patent  protection,  I  dare  say
such a change in the form of the law would not greatly expand the
scope of protection available to inventors, at least not to those
inventors  who  can afford the kind of legal talent testifying at
these hearings.

Those of us who know what we are doing can get  computer  program
machines  covered.  The  pro.......T we have to go through may be
painful to watch, may be expensive, but we can do it.   For  that
reason I would promote my suggestion principally as one that will
improve the quality of the analysis of software_related ventures,
and the doctrine under which those inventions are examined.

On the issue of infringement we would have to be a bit more  sub-
tle. If we allow claims to be made to computer programs per se we
must be careful not to create a risk of  infringement  by  tradi-
tional print media and their successors in electronic publishing.
The publication for human readers, whether or not on  paper,  for
the  patent_protected  computer  program, should not by itself be
any kind of infringement of the patent.

I would like to turn now to one of the specific questions  raised
for  today's hearing. What aspect of a mathematical algorithm im-
plemented on the general_purpose computer should or should not be
protectable  through the patent system?  I believe that a machine
made up of computer program  instructions  thatXfully  transforms
data  or  information should be protectable under the patent laws
in all its novel and nonobvious aspects.  Given the importance of
information  processing  to our economy, it would be perverse for
us to continue to deny direct protection to a technology that  is
so important to our information processing prosperity.

This leads me unavoidably to the question of what  is  novel  and
unobvious  in  a  computer  program.  I believe that our greatest
challenge lies in these  two  questions,  regardless  of  how  we
answer the question previously posed.

For myself, I would not consider novel and nonobvious  merely  to
transpose  to a computer something previously known to be done by
hand, or in one person's head, or collectively by a group of peo-
ple. But if the method for transforming information is truly new,
then the doctrines that limit or preclude protection  solely  be-
cause the method is a computer program seems unwarranted.

It has been suggested that allowing mathematical algorithms to be
protected  would.......nature  from the public domain and give an
unwarranted universal scope of protection  to  a  patented  tech-
nique.  Personally  I find those rationales peculiar.  Taking the
computer programmer's informal definition that an algorithm is  a
predetermined  set  of  steps  to  perform a function, and that a
mathematical algorithm is one that operates on  mathematical  ob-
jects,  such  as  numbers, triangles, continuously differentiable
functions, then granting protection for a new, previously_unknown
and  nonobvious  set  of steps withdraws nothing from the public.
And if the patent reaches over a  broad  range  of  applications,
that  would  merely correspond to the broad usefulness of the new

In any other technology this would be grounds  for  praising  the
inventor,  not for denying protection. I would submit to you that
if someone were to object to a patent on the  transistor  on  the
ground  that it would have too many uses, you would find that ob-
jection incomprehensible.

One final point that might be kept in mi........`efore whence  it
comes  to  an  alarm  about  the  potential  breadth of claims to
mathematical algorithms; a  naked  mathematical  algorithm  claim
would  seem  to be the ultimate engineering claims, and therefore
particularly susceptible to being rejected  or  invalidated,  be-
cause  any prior art that shows the algorithm steps being applied
in any context would invalidate the claim. Personally I would  be
surprised  if any patent practitioner would ever rely solely on a
naked algorithm to protect his client's interests.

The Patent Office also poses the variant of this question, limit-
ing  it  to the implementation of the algorithm to a special pur-
pose rather than a general purpose computer. If  the  problem  is
bad  patents, this does not seem to be a solution.  If one begins
with a computer program that should be unpatentable because it is
not  new,  or because it is obvious, one should not in my view be
able to achieve patentability merely by attaching to the  program
the  input_output devices that are conventional for the processdt
the computer performs.

In other words, one should not be able to save an old or  obvious
bread_baking program merely by attaching a digital thermometer to
it. Conversely, if the program is new and not obvious,  then  the
conventional  addition  of  necessary computer hardware and other
devices is redundant to the claim, at least insofar  as  patenta-
bility  is  concerned.   Such additional limitations would not in
fact limit the scope of protection available to the inventor, un-
less  parenthetically  the claims are poorly drafted.  But such a
redundant edition of apparatus to the program claim does create a
potentially  substantial distraction for the patent examiner who,
in the terms of my example, in searching the art of digital ther-
mometers, may completely miss the point about bread_baking.

Finally, I would like to say a word about whether we  should  re-
place  patents  with  a  new  form  of  protection  for  computer
software. My one_word answer is no.  Patent law can deliver pred-
ictability,  definiteness  and  uniformity.   hr copyright law we
cannot protect your  ideas,  at  least  not  without  doing  some
violence  to traditional copyright principles, and we are subject
to forum shopping in thirteen circuit courts  of  appeal.   Under
trade  secrets  law, we can protect our ideas, but are subject to
the law as developed in any of fifty different state  courts  and
their Federal counterparts.  We're never quite sure what the pro-
tected ideas are and are at risk of having someone rediscover  or
reverse_engineer our inventions out from under us.  But under the
patent laws we can get warning about what is protected  expressed
with reasonable clarity and applied with national uniformity.  It
would be unfortunate if such a sound concept were to be  crippled
because  we  were  too  slow in learning to apply its fundamental
principles to the challenges of software_related inventions.

Thank you.

COMMISSIONER LEHMAN:  Thank you very much, Mr. Troesch. That  was
very helpful.

The next person on our list is Brett Glass, but  we're  not  sure
that  Brett  Gl....lis here. If you are, will you please stand up
and identify yourself and come forward?  If not, we will move  on
to  Robert  Sabath,  President of the World Intellectual Property
and Trade Forum.

Back to The San Jose Index

Forward to Robert Sabath