LEE HOLLAAR

UNIVERSITY OF UTAH

MR. HOLLAAR:  My name is Lee Hollaar, I'm a Professor of Computer
Science  at the University of Utah where I teach the Seniopftware
Development Laboratory and also teach computer intellectual  pro-
perty  law within the Department of Computer Science. I also con-
duct research into information retrieval systems.  I've been  in-
volved  with  computers  for almost three decades and received my
Ph.D. in Computer Science in 1975 from the  University  at  Illi-
nois.   I'm  also a registered patent agent working with the Salt
Lake law firm of Van Cott Bagley Cornwall and McCarthy  primarily
with  computer_related  inventions.  I hold one United States pa-
tent and I have another patent pending.  The views  I'll  be  ex-
pressing  are  my own and not necessarily those of the University
of Utah or any other organization.

First I'd like to thank the Commissioner's staff for  the  oppor-
tunity  to  testify regarding these important matters and to com-
mend them on holding these hearings. I'd also like  to  congratu-
late  the  office  for  making  their decision to accept comments
electronically and to make the comments and transcripts of  these
hearings  available on the Internet. today what I'd like to do is
cover a few points from my dual perspective as a computer  scien-
tist  and  also  a patent practitioner. I'll be following up this
testimony with written comments.

One of the things I would like to mention is  that  there  are  a
number of assumptions that seem to be taken for granted about the
differences of software and its patentability that may not  actu-
ally be true, and should be examined. One is that there's the as-
sumption that computer software is a fast_moving  technology  and
that  therefore  a  lesser  patent term __ I've heard three years
suggested __ may be appropriate.  Interestingly enough, about two
weeks  ago  Butler  Lamson  gave  an address at the University of
Utah, he's one of the inventors of the Alto Computer and a number
of  other innovations from Xerox Park, and he indicated that from
at least his point of view much of the  innovation  going  on  in
computers is not the result of computer software, but the comput-
er hardware  now  available  at  much  faster  speeds  with  more
memories,xbling  techniques  that  were known in the laboratories
many years ago to be possible now and  to  be  available  to  the
masses.   In  fact  he  made  the  statement  which  many  people
disagreed with, but it certainly caught people's attention,  that
from his point of view with the exception of spreadsheets there's
been no surprises in computer science since 1975.  It's interest-
ing  to  note  that anything which you filed an application on in
1975 and a patent would have issued, the patent would  have  been
expiring about this time.  It's also interesting to note that the
fundamental books on computer algorithms, written by Donald Knuth
came  out when I was a graduate student approximately two decades
ago.  Again, if everything in those books was patented,  the  pa-
tents would be expired by this time.

Much of computer science I see, my students and  so  forth,  con-
sists  of  reinventing  wheels. A large amount of that is because
people don't check prior art and a large amount of  that  is  be-
cause there's no good prior art collections t|eck, and I think __
and I will comment on this __ that this is one  of  the  problems
that has been caused by the two decades of the Patent Office hav-
ing at best ambivalent attitudes toward the patentability of com-
puter  software and not using the patent system to draw the trade
secrets and the other art into the printed publications  of  U.S.
patents.   I  was  involved  with developing a computer system in
1969 which is still running.  If I had patented  every  technique
in  that computer system, and it still represents the basis for a
state_of_the_art system, those patents would have expired five or
six years ago on it.

Comment has been made that patents can restrict  developers,  and
that's  certainly  the  case,  but that's true on every patent in
every area of art. It's not particularly true for  computer  sci-
ence  in  any respect.  The comments have been made about bad pa-
tents, patents which have prior art problems,  poor  examination,
again  these exist in every art unit.  Perhaps it may be worse in
the computer art arecause there was this period of time when  the
Patent  Office  was  denying  computer  patents and therefore not
building their prior art collection, not encouraging the  submis-
sion  of patent applications which would mature to patents and go
into the prior art collection on it.

In my recent practice I've found  the  inexperience  that  people
have  talked  about with the examiners to be diminishing and I've
found that the examiner's doing quite a good job given  the  con-
straints  of trying to examine very complex art in a wide variety
of areas. I think also that the patent office  has  received  bad
press due to both people who should know better and also the peo-
ple providing the information not knowing enough about the patent
system.  Too often comments are based on the title or at best the
abstract of the patent, and not the claims, which  indicate  what
the  true  invention  is.   Often  this  is  compounded  by press
releases from patentees trying to make their  patents  seem  more
important than it really may be.

There have be number of problems caused by the past PTO  position
on  patents.  One,  as I mentioned, is the poor prior art collec-
tion, which has been difficult, because in  many  cases  computer
inventions  aren't  as  self_revealing  as  mechanical invention.
When you sit down and sue  a  spreadsheet  or  another  computer,
you're  aware that you're using a spreadsheet, but you may not be
aware of the order that cells are being recalculated or the  par-
ticular  algorithm  that  is  being used for justifying the text.
The patent system would have eliminated this problem  by  forcing
inventors,  in  trade  for  the  patent monopoly, to reveal their
trade secrets, to reveal how this would work,  thereby  giving  a
prior  art  collection  for  people who want to develop follow_on
systems, and because of the distinctive claiming  of  patents  as
opposed to copyright, providing an indication of what an inventor
can do to avoid the patent and yet produce an improved product on
it.

Computer scientists suffer from a poor  tradition  of  publishing
their  algorithms,ecially  those involved in industry, and also in
looking at prior art. I often think  that  many  of  my  students
wouldn't know where the library is on campus if you ask them, and
when they start out on a project the idea of going to the library
is not the first thing that crosses their mind, and certainly the
idea of doing a patent search to see what's in the prior art  and
see how to accomplish things is far from their mind.

I think that  while  the  Software  Patent  Institute  should  be
congratulated  for  what  they're doing, their idea of collecting
prior art is doomed to failure because especially for old  inven-
tions, it is very hard to determine what is prior art or what are
the novel things. In the typeset system I wrote  many  years  ago
there  were probably hundreds of things which may be novel or may
at least be prior art for future inventions, and it would be hard
to enumerate them.

Other problems caused by the past PTO position  is  both  in  the
disclosure  and  in  particular  in  the claims' obtuse language,
unclclaims, as attempts to avoid a perceived  perception  of  the
PTO  position  on  Section 101, claiming things which are clearly
software programs as computer building blocks. I think of one pa-
tent issued to Thompson where not only is the source code listed,
but they tell you how to build it out of  computer  modules  from
Digital Equipment, out of flipflops for this searching technique.
No one believes that that's how they intended  to  implement  the
invention.   More importantly, the Section 101 babble often takes
the steam out of the examiners.  After  fighting  the  101  fight
there  may  be  very little fight left in them for the proper 102
and 103 questions.  I recently reviewed the file wrappers of  two
patents  for  a  client  who  wanted to know their scope, and was
shocked by how little prosecution history there was after the 101
arguments  had  been  resolved, even though one patent had a con-
tinuation application filed and had been pending for a number  of
years  on  it,  virtually no 102, 103 arguments after the initial
101.  think the patent system has produced distortions  in  copy-
right  laws,  courts have held it necessary to provide protection
beyond literal copying so that we have decisions like  Whelan  or
the  current  Lotus  against  Borland  decision, trying to extend
copyright perhaps too far from its intended  purpose,  and  as  I
said  we have the loss of past disclosures because the Patent Of-
fice wasn't accepting patent applications, which has caused  much
reinventing of past techniques.

The solutions?  I think one is to eliminate much of  the  current
101 confusion, and in fact go back to the basic principles of in-
ventorship, that patents are, if there is something useful  being
produced,  should probably meet the 101 test, and the real battle
of whether the patents should be issued should be  tried  on  102
and 103 issues.

The comments have  been  made  about  laying  open  files  during
prosecution,  and  I  strongly support that. I think it's the one
hope we have for getting the prior art.  Both laying it  open  at
some  fixed period of time sas eighteen months has been suggested
to eliminate the so_called submarine patents, and also laying  it
open  sometime after initial examination by the Patent Office.  I
would make sure that the file isn't laid open  before  the  first
office  action  such  that the applicant has a chance to withdraw
the application based on the position of the  Patent  Office  and
the prior art that's been found, and make an intelligent decision
of whether to keep it as trade secret.

I would recommend that there be a period after the office  deter-
mines that a patent is allowable that people can submit prior art
for consideration by the examiner. This could be  done  by  using
technology  to  widely  distribute the notice that this patent is
about to issue and a representative claim and has some method for
retrieving  of  the application, perhaps tied in with the project
ongoing  for  the  electronic  filing  of   applications.    This
shouldn't be an advisory procedure, but a way of having the prior
art brought to the attention of the examiner. I k in my  case  of
the old system that I have, I certainly could see if there was an
application and claims whether I had prior art on that and submit
that  to  the attention of the Office much easier than I could go
through a very complex system and identify every piece  of  prior
art such that it could be searchable by the Patent Office.

That concludes my remark, thank you.

COMMISSIONER LEHMAN:  Thank you very much,  Professor,  that  was
some  really  very  helpful testimony. Appreciate you coming over
here to California to talk with us.  Thank you.

Next I'd like to ask Dennis Fernandez. Is he  here?  For  Fenwick
and West?  Okay.

Welcome.

Back to Dennis Fernandez