W3C home > Mailing lists > Public > public-web-plugins@w3.org > October 2003

Court Ruling: Prior Art of the Viola Browser

From: <jim@idexer.com>
Date: Tue, 30 Sep 2003 22:29:40 -0700 (PDT)
Message-ID: <3741.>
To: <public-web-plugins@w3.org>

Here is the ruling on the Viola Browser in the MS Eolas case. Text is
converted from PDF format, so there can be some word spacing.

No. 99 C 626
Judge James B. Zagel


A Preface on the Law

Inequitable conduct is an applicant's breach of his duty to prosecute a
patent application with candor, good faith, and honesty. Li Second
Family Life L.P. v. Toshiba Corp., 231 F.3d 1373, 1378 (Fed. Cir. 2000).
All patent applicants owe this duty to the Patent and Trademark Office
("PTO"), and it exists throughout the entire prosecution of the patent.
See 37 C.F.R.  1.56(a); Semiconductor Energy Lab. v. Samsung Elec., 204
F.3d 1368, 1373 (Fed. Cir. 2000); Fox Indus., Inc. v. Structural
Preservation Systems, Inc., 922 F.2d 801, 803 (Fed. Cir. 1990).
Inequitable conduct "includes affirmative misrepresentations of a
material fact, failure to disclose material information, or submission
of false material information, coupled with an intent to deceive."
Baxter Int'l, Inc. v. McGaw, Inc., 149 F.3d 1321, 1327 (Fed. Cir. 1998).

With respect to inequitable conduct based on failure to disclose
information to the PTO, there is a three step analysis in which the
court must determine: (1) whether the withheld information or
misrepresentation meets a threshold level of materiality; (2) whether a
threshold level of intent has been shown by the evidence; and (3) if
these thresholds are satisfied, "whether the equities warrant the
conclusion that inequitable conduct occurred." Semiconductor Energy, 204
F.3d at 1373. Both materiality and intent must be found, and the more
severe one is found to be, the less severe the other has to be. See
Monon Corp. v. Stoughton Trailers, Inc., 239 F.3d 1253, 1261 (Fed. Cir.
2001); Critikon, Inc. v . Becton Dickinson Vascular Access, Inc., 120
F.3d 1253, 1257 (Fed. Cir. 1997).

Also, because a "mere showing that art or information having some degree
of materiality was not disclosed," is insufficient to establish
inequitable conduct, it is the intent to act inequitably that must be
shown. Thus, for a failure to disclose to constitute inequitable
conduct, there must be "clear and convincing proof of:

(1) prior art or information that is material;

(2) knowledge chargeable to the applicant of that prior art or
information and of its materiality; and

(3) failure of the applicant to disclose the art or information
resulting from an intent to mislead the PTO." FMC Corp. v. Manitowoc Co,
Inc., 835 F.2d 1411, 1415 (Fed. Cir. 1987).


I. The Inventors' Knowledge of Viola Michael Doyle knew in November of
1993 that something which was denominated as "Viola" existed. In two
separate accounts of what he knew, he first said that ORA said that
someone was trying to get something to work with program objects
embedded in documents and transported over the World Wide Web, but it
was not yet ready to show. A second acco unt was that he did not
remember ORA using the word "Viola" but did remember mentioning
something about a book browser or some kind of software they were
working on that related "to form submission on Web pages."

On May 20, 1994, Michael Doyle received an e-mail from David Raggett
which said: The EMBED tag was dropped after the WWW workshop in Boston,
late last July. It was felt by most brow ser writers that further study
was needed on how best to implement object level embedding in Web bro
wsers. This feature is still on most peoples' agenda though. You might
want to look at Viola which I seem to remember takes advantage of the tk
tool kit to provide a level of embedding. You can find a pointer to
Viola off the CERN WWW project page. Doyle stated he understood this to
mean that Raggett was referring to the "TK WWW" browser rath er than
Viola becau se the "TK WWW" was the only brow ser that used the TK to ol
kit. By May 20, 1994, another inventor, David Martin, had known of Viola
because he corresponded with Pei Wei, the developer of V iola, and was
discu ssing with Pei Wei a trade, to wit, an "updated Viola WWW
executable" in return for a guest compu ter account.

By August 31, 1994, Michael Doyle issued a press release to the VRML
e-mail list about what would become of the `906 patent. It said:
Researchers at the U. of California have created software for embedding
interactive program objects within hypermedia documents. Previously,
object linking and embedding (OLE) has been employed on single machines
or local area networks using MS Windows -TM-. T his UC software is the
first instance where program objects have been embedded in documents
over an open an d distributed hyperm edia environment such as the World
W ide Web on the Internet. A key feature of the inve stigation was the
em bedding of prog ram objects ove r the World Wide Web.

Pei Wei responded to the press release at 6:54 p.m. on August 31, 1994:
This is very interesting_ But, I don't think this is the first case of
program objec ts embedded in docs and transported over the WWW Viola
has had this capabilities for months and months now. Been meaning to
propose so mething for VRML ever since the Geneva W3 conf_ But anyway,
any body interested in learning m ore about how violaWW W does this
embedded objects thing can get a paper on it from:

Michael Doyle received Pei Wei's response on A ugust 31, 1994, he went
to the Internet address given by Pei Wei, found a paper describing the
Viola browser, downlo aded the paper o nto his computer, and read
it-the Viola Paper. At 9:06 p.m . on August 3 1, 1994, Do yle responded
to Wei by e-mail with this qu estion: "How many months and months? W e
demonstrated our technology in 1993." About two hours later, at 11:13
p.m., Doyle sent a further e-mail to the VRML distribution list,
commenting on Wei's pa per about Viola:

As Pei's pap er on Viola states, that pa ckage did not sup port what it
calls `embeddable program objects' until 1994. As our WWW server shows
(http: visembryo.ucsf.edu/), we demonstrated a fully functional volume
visualization application embedded within a WWW document in 1 993.
Furthermore, Viola merely implements an internal scripting language that
allows one to code "mini application" scripts that are transferred to
the local client, and then interpreted and run locally on the client
machine. As Pei correctly notes in his paper, this is similar to the use
of EMACS ' internal programming capabilities. What we have accomplished
is much different. Just as the Microsoft Windows OLE function allows any
OLE-compliant application to be embedded, in its native form, within,
for example, a MS Word for Windows document, we can embed ANY
interactive application IN ITS NATIVE FORM within a WWW document. At
11:16 p.m. on August 31, 1994, Wei responded to Doyle's "how many months
and months" query. Definitely by May 8, 1993 we had demonstrated that
plotting demo (the very one shown in the viola paper) to visitors fro m
a certain computer manufacturer_ This demo was memorable because someone
and I at ORA had lost sleep the night before the meeting, in order to
cook up that particular plotting dem o:) We had to show something cool.

That demo wasn't very hard to do because by that time the basic
capability was already in place for viola WWW to fetch viola objects
over HTTP (or whatever) and plug them into documents. Of course, our
wire-frame plotting demo isn't anywhere as com prehensive as you rs.
But, the point w as that there was a w ay to embed programmable &
interactive objects into HTML documents. After receiving this e-mail,
Doyle looked at Wei's Viola paper a second time to study the "wire-frame
plotting demo" (showing an airplane image) that Wei referred to in the
e-mail. Doyle then sent Wei a further e-mail at 11:36 p.m.:

Out of curiosity, did you publicly demonstrate this or publish any
results before 1994? I remember talking to people from ORA at the first
SIG-WEB meeting in November of 1993 and they said that no such features
were yet publicly demonstrable in Viola. I seem to remember that they
hinted at the time that someone was trying to get something to work, but
it wasn't ready to show yet.

Wei responded to Doyle's e-mail shortly after midnight on August 31,
1994, explaining that before 1994 he had "done lots of demos to many
parties in private," and had demonstrated the Viola browser "on the side
at InterOPS and on ce at SIG-WEB a few m onths ago. Nothing big." He
also told Doyle that by Novem ber 1993, "those embeddable objects
features w ere certainly in place" and "were being demonstrated to
visitors." Wei also responded to Doyle's 11:13 p.m. commentary
concerning the Viola browser, sending an e-mail to the VRML mailing list
at 8:19 a.m. the following morning, September 1, 1994. In response to
Doyle's claim that the Viola "package did not support what it calls
`embeddable program objects' until 1994," Wei maintained that "Viola's
model was *demonstrated* in 1993, *released* freely in 1994."

The `906 patent application was filed on October 17, 1994, some six
weeks after this round of e-mail exchanges. The `906 application
discussed the Mosaic brow ser in passing, and included an Internet
address (http://www.law.cornell.edu/) where information could be
obtained about the Cello browser. Neither browser enabled the use of
interactive program objects in Web pages. The `906 application made no
mention of Pei Wei or the Viola browser. A second round of e-mail
exchanges took place in 1995. On August 21, 1995, Eolas sent an e-mail
press release to the www-talk e-mail list, announcing that it had
"completed a licensing agreement with the University of California for
the exclusive rights to a pending patent covering the use of embedded
program objects, or `applets,' within World Wide Web documents." Doyle
said that one of the reasons he sent ou t this press release was tha t
he realized that if there was relevant prior art, "someone would likely
tell us about it."

Doyle understood that any material prior art that came to his attention
because of the issuance of the press release would have to be disclosed
to the Examiner. Wei responded to the press release with an August 21,
1995 e-mail stating: And for the record, I just want to point out that
the technology which enabled Web documents to contain fully-interactive
"inline" program ob jects was existing in ViolaWWW and was "released" to
the public, and in full source code form, even back in 1993_ A ctual
conceptualization and existence occurred before '93. Doyle rep lied
later that d ay: We've had this discussion before (last September, remem
ber?) You admitted then that you did NOT release or publish anything
like this before the Eolas demonstration. Wei replied:

Please carefully re-read my letter to you_ I said Viola was demonstrated
in smaller settings, but before your demo. The applets stuff was demo'ed
to whom ever wanted to see it and had visited our office at O'Reilly &
Associates (where I worked at the time). Wei then quoted his earlier
e-mail, where he told Doyle that "by May 8, 1993 we had demonstrated
that plotting demo (th e very one show n in the Viola Pap er) to
visitors from a cer tain computer manufacturer." He continued:

That date (May 93), at least, predates your demo if I'm not mistaken.
Then around August 93 it w as shown to a b unch of attendees at the
first W eb conference in Cambridge. So, it was shown , just not with
lots of publicity and noise. If you're talking about interactive apps
*specifically* on the web, ie applets in-lined into HTML docum ents
etc., and with bi-directional communications, then look at ViolaWWW as
it existed around late '92 early '93.

In 1998, while the `906 patent application was still pending in the
PTO, Doyle decided to do further research regarding the Viola browser.
In the course of his research, Doyle found a presentation describing the
Viola browser which Wei had delivered at Stanford University-the
Stanford Slides. The presentation described how "program objects can be
embedded into documents" and supplied two illustrations showing
interactive images (of a chessboard and the wire frame airplane)
displayed inside the browser window. Doyle testified that, in 1998, he
noticed the fact that the presentation showed the airplane image as
appearing inside the browser window.

Doyle created a file to hold all the information he found in 1998 about
the Viola browser, and he labeled his file "Viola stuff." The "Viola
Stuff" file included descriptions of two "beta" releases of the Viola br
owser, a versio n 3.0 release in Feb ruary 1994, and a version 3.1
release in March 1994. There w ere public announcements in both cases of
Internet addresses where "source and binary" code for the Viola browser
could be found. He also found extensive links for various purported
"demos" of the Viola brow ser's capabilities. Doyle never had possession
of the Viola code dated May 12, 1993 (earlier May code) or the Viola
code dated May 27, 1993 (later May Code) nor any executable binary code
asserted as prior art nor had he possession of any Viola source code or
executable binary code dated from either 1993 or 1994.

Neither the earlier nor the later May codes were capable of executing
the asserted prior art plotting demonstration in a system having at
least one client work station and one network server coupled to said
network environment, wherein said network environment is a distributed
hypermedia environment.

Doyle never had possession of any Viola reference, paper or other
publication of any sort dated in 1993 that described the functionality
of Viola nor was he given any details of any May 1993 demonstration of
Viola software, including: who was present, what was actually
demonstrated, the actual software, the functionality of the software
demonstrated or how that functionality was achieved.

Doyle was never given any written documentation or other corroborating
evidence of any May 7, 1993, demonstration of Viola software. The Viola
source code was hard to obtain. Wei did not want public access to his
1993 source code and he made it "nearly impossible" for the public to
get it.

<A Short Note on Why I Have Found These Facts>

Microsoft argue s that Doyle knew more than he w as saying at trial.
They challenge his statement that he thought Raggett was referring to
the TK toolkit. They reasoned that he had learned about Viola at the
SIG-WEB conference and thus he would have know n this was wrong. They
also contend that he could have and, inferentially, did get Viola source
code. What was proved convincingly (and is undenied) is that Doyle knew
of "Viola," but there is no proof that he knew how it purported to do
what it was said to do or that it worked at all. He was in the position
of one who knows the name of a computer file and is unable to access
more than a small, insignificant portion of its contents and then only
on intermittent occasions.

I do not believe it is likely, let alone convincing, that Doyle would
have, or should have, put together the fairly vague conversation's he
had at SIG-WEB with an e-mail from Raggett to arrive at a conclusion
that Viola was not the TK WWW browser-particularly since the meeting
occurred six months prior to receipt of the e-mail.

(Office actions rejected his application based on the U.S.C.
Mercury Project, Mosaic/Khoyi and Mosaic/Koppolu.)

To this I add the difficulty of getting the Viola source code. Microsoft
did not find it easy to track it down at a later time. I noted on the
record that there is Viola material today which I found on web history
sites, but it is difficult to know when it was posted there, whether it
works, and whether it is the 1993 source code. My observation was not
pursued by any party, so I disregard it. In any event, I believe Doyle
when he said he did not have the source code. Recognizing that this
might be the case, Microsoft argues that what Doyle did know about Viola
was enough to trigger his obligation to bring it to the attention of the
Examiner during the prosecution of the `906 patent. The argument is
addressed below.

<End of Note>

II. The Date of Invention

The `906 invention was reduced to practice no later than January 27,
1994. It was presented on that date at a conference "Medicine Meets
Virtual Reality II." The Materiality of What He Knew To the examiner,
Doyle consistently maintained that his invention was the first instance
where interactive ap plications were embedded in Web pages. It was
this assertion that was used to distinguish prior art cited by the

The key as he said was that the `906 patent made it. Were it to be
prior art, I would have to decide the question of whether SIG-WEB
preceded the other two dates that Doyle has offered-a reduction to
practice on November 13, 1993. "possible to embed fully interactive
external applications in Web pages thereby turning the browser into a
platform for the development of entirely new kinds of application."

III. The date of invention excludes the Viola Paper (dated August 16,
1994) and the Stanford Slides (dated September 20-21, 1994) from the
ambit of prior art. Also excluded are the portions of the Viola Stuff
dated after the date of invention. And all the Wei e-mails come after
January of 1994. The only item left in prior art is what Doyle was told
at SIG-WEB and this is clearly not prior art.2

The Viola Paper is not prior art in any event. The clearest indication
of Viola's capacity to deal with embedding objects is said to be in the
1994 Viola paper which also refers to the use of a tag. Interactivity is
not described in the paper, the embedding process is not described nor
is the embed text format. The use of type information by the browser to
identify and locate an executable application or how such an application
could be automatically invoked is also not revealed. The depicted
applications (a chess board and an airplane) are not shown in the
browser window. The likely reading of the Viola Paper is that Viola WWW
uses helper applications which were included in the Examiner's analysis.

The Stanford Slides come closer since the plane and chess board appear
to be in the browser window, although from the slides one cannot tell
whether the screen shots were moved into the depicted browser window
since the plane seems to be in the same Unix window as in the Viola
Paper. The chess board was not demonstrated as operational at Stanford.
The Viola Stuff includes references to a beta 3.0 and 3.1 release of
Viola WWW, but operational details are not set forth, and no source code
was found when Doyle looked for it. 3 I note in passing that even if it
had been done, it has not been shown that the Exam iner could have found
evidence of Viola's functionality that existed prior to the date of

There is no indication that he should have been able to find it. Other
references are to generic statements of how one might achieve some of
the goals of `906, but none of it has any meaningful detail. Wei's
e-mails are simply claims that he invented `906 first, but there is no
detail of what he did, how he did it, or to whom he demonstrated it. No
witness has offered an opinion that any of the items in question are
prior art.

<A Note on Why I Found These Facts>

Microsoft argues that if Doyle had disclosed what he knew about the
Viola browser, the Examiner might have considered that `906 was obvious.
He might have combined the interactive program object feature of the
Viola browser with the Mosiac browser. Doyle concedes, as he must, that
"the prior art taught that good ideas in one browser could be used in a
different browser." Microsoft makes something of this concession. I
leave aside that the word " idea" does not fit well in the discussions
of patentability. I assume Doyle meant "feature" or "functionality" in
one browser could be used in another browser. Yet what remains is that
Doyle would have disclosed nothing more than Wei's assertion that he
achieved something.

 An inventor need not tell the Examiner about every assertion that the
invention was made by another. Some inventors might have disclosed what
Wei had said and sent the Examiner off to hunt on his or her own for
source code or some other evidence of invention rather than a claim of
invention, but the law does not require this to be done.3 See Manitowoc,
835 F.2d at 1415 (stating that a "mere showing that art or information
having some degree of materiality was not disclosed" is insufficient to
establish inequitable conduct). 4 At trial I found there was
insufficient evidence to per mit the jury to consider in validity
defenses based upon Viola. I excluded evidence of the early May code for
procedural reasons.

I have considered both codes in this ruling only to the extent they bear
on the inventors' duty of candor before the PTO-a different issue
because the Exam iner is supposed to be told of things 12 <End of Note>

There was Viola source code.

At one time the operation of the earlier May code was shown to two Sun
Microsystem engineers. The source code itself was not shown to the
engineers, and that code was changed within a week or two into the later
May code. There was no formal confidentiality agreement with respect to
the demonstration, but circumstances demonstrate that it was private.
The audience was limited to two persons. The Viola project was, in the
eyes of its inventor, a work in progress. I infer that neither Wei nor
the engineers regarded the Viola project to be ready for or to have been
part of a public disclosure. The later May source code was given to Sun
engineers, but this too, Wei sought to keep from the public. Neither the
early nor late May codes were enabled for use in a distributed
hypermedia network environment.

The later May code did not perform a certain functionality of the `906
patent. The Sun engineers could not make it work. The source code did
not disclose to a person of ordinary skill in the art, in an obvious
way, how it would have to be modified to make the deficient
functionality (executing the plotting demonstration) work. Over time,
Wei made further changes to the code and tried to keep the public from
getting it.

Neither source code taught the `906 invention. Had either done so, Doyle
did not know it. which might affect the decision, not solely those which
are thought likely to do so. The procedural bar to introducing the early
May code which arose from rules applicable to adversary proceedings is
not applicable to the PTO process.

IV. The Inventor's Intent to Deceive Doyle was, from the beginning,
acutely aware of, and hopeful about, the commercial value of his
invention. He had, of course, a financial incentive to trim the truth in
this court and, as well, during the patent prosecution. Microsoft has
some degree of financial incentive to prevail in this court, but given
the evidence presented here, the financial loss Microsoft would bear
under a worst case scenario is large but easily bearable. The financial
incentive to Doyle is far greater than it is to Microsoft.

And Doyle knew throughout the prosecution process that it would be
Microsoft that, above all others, would be his adversary in litigation.
He sought to protect the value of the invention by committing Wei to an
invention date (and did a pretty good job of it). His testimony that
this reason for committing Wei to a date never crossed his mind is
untrue. So too is the message of his assertion that all he provided in
the prosecution process was "input" to patent counsel. If one views,
say, the order of MacArthur to his Pacific Theater commanders as
"input," then one could similarly say Doyle offered only "input." In
this sense, Doyle's statement is literally true, but the meaning he
intended to convey-that he was merely a resource to the patent
attorney-is not true. Doyle was truthful, however, when he said that he
did not believe that Viola showed an embedded program object running in
the browser window.

Doyle knew and believed only that Wei claimed to have invented the `906
invention, but nothing that Wei showed to or made available to Doyle
backed up that claim. Doyle was in the same position as an inventor who
receives a claim that someone else is the true inventor. The law does
not require that every such claim needs to be disclosed to the Examiner.
See Manitowoc, 835 F.2d at 1415 (stating that a finding of inequitable
conduct based on a failure to disclose cannot be established by "a mere
showing that art or information having some degree of materiality were
not disclosed") and Halliburton Co. v. Schlumberger Tech., 925 F.2d
1435, 1442 (Fed. Cir. 1991) ("materiality of an undisclosed reference
does not presume an intent to deceive.").

Wei was, to Doyle's eyes, a hostile rival who alleged he had done
something but avoided, if not evaded, any reasonable attempt to
corroborate his claims. Doyle could reasonably conclude, and did so
conclude, that the Examiner would face the same difficulties that Doyle
had in trying to corroborate Wei's assertion. Because no value would be
added to the Examiner's consideration of the patent, disclosure was not
mandated by law. This was in fact what Doyle believed. He had no intent
to deceive.

<A Note on Why I Found These Facts>

Ordinarily when a court credits the testimony of a witness, it simply
says that the witness was credible or, at times, offers a paean to the
candor and demeanor of the witness. I have done neither because I
thought Doyle was not truthful in every aspect of his testimony. In this
regard, I have mentioned the impression he sought to leave that he was
an almost passive consultant in the prosecution of the application and
his denial of patentability issues when he asked Pei Wei to give a date
of invention. I was unpersuaded by his account of an EOLAS commercial
endeavor which ended badly.

Testimony that falls short of the truth is not necessarily deliberate
falsehood. Doyle may have convinced himself of the truth of every thing
he said. And, even basically honest persons rewrite history on matters
they believe are not central to a dispute.

I reach the conclusion that he had no intent to deceive the PTO since he
wanted to act with candor before the PTO both because it was the law and
because it was too dangerous not to do so. From the beginning, he knew
that commercialization of his patent would require arduous effort
against very large enterprises. He could not have believed, for a
moment, that Wei's claim would fail to surface. Wei was not quiet about
what he had done. And Wei had the e-mails too. Doyle would have known
that he would, someday, have to face Wei. Had he intended to deceive, he
would, in my judgment, have destroyed nearly all of the Viola Stuff

In theory, he gains some advantage by conducting the dispute over Wei's
claims after the `906 patent issued. The presumption of validity works
against inequitable conduct and invalidity defenses and burdens of proof
are shifted to the defense. But the importance of this is not as great
as it appears. Before the Examiner, neither Wei nor Microsoft can
advocate the position taken by the inventor. At trial, those who put
forward Wei's claim can use subpoenas to get evidence to support it and
offer expert testimony to validate it. In the end, even the resources of
the defense could not make a passable case out of Wei's claim to have
invented the `906 patent.

So I do not believe that Doyle made a tactical decision to forego
disclosure because of the advantages of defending his invention after
patent issued. He made the decision because he believed that Wei was
full of hot air and he had nothing concrete to display to an Examiner.
So I credit his testimony that he had no intent to deceive the Examiner.
In this case, the Wei claim, as known to Doyle, was of limited
materiality at best, and he had no intent to deceive. <End of Note>


1. The Court has jurisdiction over the offered defense of inequitable

2. The defendant has failed to prove that the inventors or assignees of
the patent failed in their duty of candor in the patent prosecution
before the United States PTO.

ENTER: James B. Zagel United States District Judge DATE:
Received on Wednesday, 1 October 2003 01:29:44 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 18:09:06 UTC