|
!19980427 Rule 56 contempt: Action Technologies/Gail Hayes/Blakely Sokoloff
But first, a warning on telemarketing scams. A lot of people are getting calls asking if they want to switch their long distance provider for some better promised rate. If during the conversation you hear the phrase, "Can I put you on hold?", be very, very careful, and find out where the telemarketer is calling from. One firm, Home Owners Long Distance (H.O.L.D.) is preying on people's politeness and using a YES answer to the above question as sufficient permission to switch the victim to H.O.L.D.'s phone service.
====================
Last week, HR3723 was introduced as a bill to reduce PTO fees. Hard to be against this bill, but also hard to understand the logic of Congress reducing PTO fees while Congress is stealing PTO fees.
====================
If you want a good example of why the US software patenting system is mostly a joke, a recently issued software patent does so in many ways. Rule 56 violations, inept patent examination, token legal counsel, shoddy academic contribution, lack of real prior art searching, etc. Who am I talking about? Our good old friends at Action Technologies, who had a patent invalidated by the courts last year on trivial non-patent prior art grounds, and who obviously learned absolutely nothing from the experience. Just like the invalidated patent, this patent of Action's was examined by Gail O. Hayes and prosecuted by the Valley law firm of Blakely Sokoloff Taylor & Zafman.
THE PATENT
5,734,837
Method and apparatus for building business process applications in terms of its workflows
Assignee: Action Technologies (filed: Jan. 14, 1994)
Inventor: P. Flores, R. Flores, R. Medina-Mora, J. Vasquez, J. McAfee M. Kumar, M. Nunez, T. Winograd, H. Wong, R. Gift Abstract
The invention is a method and system which provides consultants, business process analysts, and application developers with a unified tool with which to conduct business process analysis, design, documentation and to generate business process definitions and workflow-enabled applications. The invention may be implemented using a software system which has two functional sets. One is a set of graphical tools that can be used by a developer or business analyst to map out business processes. The second is a set of tools that can be used to document and specify in detail the attributes of each workflow definition, including roles, cycle time, conditions, of satisfaction, cost and value, associated text, forms, application data as well as detail the attributes of links between workflows required to complete a business process map, and to generate a business process definition and a workflow-enabled application. In this manner, the invention provides the capability of performing application generation and generation of business process definitions in a definitions database. The invention also includes a workflow language scripting capability.
U.S. Patent References
3,736,409 5,182,705 5,216,592 5,233,513 5,249,300 5,301,320 Other References
Scherr, A.L.; "A New Approach to Business Processes"; IBM Systems Journal; vol. 32, No. 1; 1993.
RULE 56 VIOLATION
Even at the outset, this patent is suspicious for the paltry list of prior art cited. Five patents and one non-patent prior art item for a business CASE technology patent applied for in 1994? Get serious - nothing is cited from any IEEE or ACM software engineering journals and conferences, let alone lots of other journals and conferences. Examples, you ask? Examples you get, such as one very relevant piece of non-patent prior art that also is a nice blatant Rule 56 violation, since the author of this 1992 conference paper is one of the named inventors:
AUTHOR: Medina-Mora, R.
COMPNY: Action Technol. Inc., Alameda, CA, USA
TITLE: ActionWorkflow technology and applications for groupware SOURCE: Proceedings of GroupWare August 1992, 165-7
ABSTRC: The paper outlines Action Technologies' Action Workflow technology. It reviews the ideas behind the technology, complete with a review of its overall architecture and various components. The functionality that is added to groupware environments is presented; and the technology's integration with existing environments, including database and messaging environments, is discussed. The paper concludes with examples of business applications that have been successfully implemented.
with an accompanying, almost as badly blatant (the Action author wasn't a named inventor) Rule 56 violation from another conference:
AUTHOR: Dunham, R.
COMPNY: Action Technol. Inc., Alameda, CA, USA
TITLE: Business design technology: software development for customer satisfaction
SOURCE: Proc. of 24th Annual Hawaii Int. Conf. on System Sciences, 1991, vol. 3, 792-8
ABSTRC: Action Technologies has developed a new technology, business design technology (BDT), for developing applications that support businesses in managing and tracking the completion of business actions in order to produce customer
satisfaction. With this new technology, customer
satisfaction can become part of the design specification for mission-critical systems that produce revenues and manage costs for the company. Applications produced with this technology not only manage both the recurrent, structured actions and unstructured, improvisational actions in support of business objectives but also provide mechanisms for monitoring and managing the satisfaction of customers involved with the application. Moreover, BDT significantly reduces the time and cost required to develop, distribute, and modify BDT applications for business.
Gee, money to fly to a Hawaii conference to give a paper, but not enough money (like 10 cents) to do a search to find that paper, I guess.
No reference is made to the existence of these tools in the patent specification, which is surprising, given one of the statements in the specification that matches these abstracts: "The workflow application builder is the subject of the present application." How then, after having one patent invalidated by the courts on non-patent prior art grounds, could Action Technologies have the audacity to not submit in the above articles?
Maybe they guessed that there is a low probability that the PTO will do much checking and doesn't go after even the most blatant abuser of the rules. Of course "audacity" implies intention, while it may be that Action is just plain contemptous of the role of non-patent prior art, reflecting the general contempt for this integral part of the process.
INCOMPETENT EXAMINER
The examiner on this patent, much like the earlier Action patent that the court invalidated, plus a variety of other software patents that cite NO prior art of any sort (5,299,123 - 5,307,268 - 5,153,830 - etc.), was Gail O. Hayes. There isn't much of a risk in not sending in Rule 56 art when the examiner has a demonstrated inability to find non-patent prior art, like the above Rule 56 violations, or the following paper, found with a quick five minute INSPEC search for papers before 1993 containing "workflow" and "design" in the abstract:
AUTHOR: Hurwitz, J.S.
TITLE: AT&T's Rhapsody: targeting groupware and workflow automation SOURCE: Patricia Seybold's Office Computing Report, vol.13, no.6, p. 16-19, June 1990
ABSTRC: AT&T calls its strategy to become a major applications software supplier and systems integrator 'Business Orchestraction'. Rhapsody is the software implementation of this strategy. The author takes an in-depth look at the various facets of this strategy. Object orientation is at the core of AT&T's design. The advantage of this modular approach to design is that a company can more easily mix and match component parts without rewriting code. The company is picking up Hewlett-Packard's NewWave as the base user interface and its Object Management Facility (OMF) as a way to achieve object exchange over the network. The workflow automation software, a major part of Rhapsody, is provided by Workhorse Systems Limited, a small software company based in Dublin, Ireland. For imaging software, AT&T looks to Plexus's client/server-based image management system. AT&T uses its own mail API to tie these various components together. Rhapsody consists of many different components, but the heart of the product is workflow automation-also known as procedural automation or taskflow automation. In general, workflow automation is a series of consecutive-or simultaneous-tasks that follow the flow charge of a particular process or activity. Each step is automated, keeping track of the status of the process and plugging all the holes.
Recidivism applies to white collar crimes as well - a known inadequate search (the invalidated patent) tends to imply other inadequate searches by the same applicant. Admittedly a trite theorem for software patenting, since most applicants don't do much searching, but you think the PTO would have paid a little more attention to Action's patent applications after the court invalidation. Too busy issuing those First Office Actions, I guess.
ACADEMIC COMPETENCE
Much like the invalidated patent, one of the listed inventors for this new patent is Stanford Professor Terry Winograd. Prof. Winograd is the author of many journal articles, and a thesis advisor, so he is well versed and well familiar with bibliographics and their preparation. Further, Stanford University has one of the richest collections of non-patent prior art in the country. Yet despite all of this, these patents he is a party to cite little to no non-patent prior art. Maybe acceptable for a professor at Cal State Gilroy, but not Stanford. But again, why bother if the Patent Office isn't paying attention.
Inventors from large universities have much easier access to databases and libraries than most inventors. Thus it is harder for them to plead lack of access or undue burden to go searching, especially when they have written lots of articles and been thesis advisors. And given the gorgeous view from the patio of Stanford's Math/CoSci library, you're a fool not to want to spend a few minutes there anyway.
LEGAL COUNSEL
Having had their client's earlier patent invalidated on trivial non-patent prior art grounds, you have to wonder what kind of counsel Blakely Sokoloff gave to Action on this patent. It couldn't have been something along the lines of "Look guys - you lost one patent for not searching much of anything - how about putting a little more effort in this time around." To allow their client to make the same mistake repeatedly is poor counsel.
Some other recent Blakely client's patents just as bad: Sun Microsystems, 5,734,831, filed in 1996 titled "System for configuring and remotely administering a Unix computer over a network", citing only one patent, and NO non-patent prior art (including nothing from USENIX, which is pretty much an impossibility); and Ricoh, 5,710,816, filed in 1995 titled "Method and apparatus for ensuring receipt of voicemail messages", citing only three patents and NO non-patent prior art.
I am sorry, but even to some minuscule, microscopic level, I believe that patent lawyers should be champions of the patent system. Firms should whack their clients over the head at least once during the patent prosecution stage to get them to do some searching. Sometimes the client is wrong, and has to hear it.
PATENT OFFICE
Since Compton's software patent stunt in 1994 (which is still being argued in the courts), about 30,000 software patents have issued, and frankly most aren't too much more valid than the Action patents. The PTO has made ZERO progress in these last three years, despite the stories they spin in patent bar circles. The amount of non-patent prior art being cited is grossly inadequate, all the more sad in light of the PTO's continued support of the fraudulent Software Patent Institute.
What is all the more upsetting about these non-patent prior problems is the PTO's inability or apathy to even doing the unambiguous search of the inventor's names in the online databases. The PTO should have found the reference to Medina-Mora's paper. There can be nothing more simpler than this type of check. And yet I repeatedly see patents issuing where one of the authors had non-patent prior art in his name germane to the patent, yet not submitted to the PTO. Clear Rule 56 violations going undetected and unsanctioned.
So anything the PTO says about progress it is making should not be taken too seriously - they do not understand the problem. Which is all so ludicrous, since part of the solution is accessing the contents of large proprietary databases such as INSPEC, the type of copyrighted databases Commissioner Lehman wastes too much time on trying to develop suet generis forms of protection for. Bruce, why bother new protections for databases if you ain't going to use them?
A COMPLETE JOKE
With the PTO and patent bar washing their hands of the problem of software patenting, and applicants looking for a way to save a few bucks, no wonder software patenting is so dysfunctional. Since Compton's, since the 1980's, since the 1960's, everyone has known that examiners do not have enough time, money and resources to do a decent patent examination for software, and that requirements like Rule 56 meaningless. And yet nothing changes. Thus the complete joke, with issued US software patents should being presumed to be invalid.
Which is a real damn shame for the small software companies who increasingly will be relying on their software patents for protection in markets where they can't compete with marketing dollars. The patents they are being issued are no more useful than trying to duel with wet spaghetti. Not very American if you ask me.
Greg Aharonian
Internet Patent News Service
====================
Here is my story on the court invalidation of an earlier Action patent, showing the trivial amount of non-patent prior art used to invalidate the patent.
DATE[1997072
TITLE[District Court invalidates Action's groupware software patent TEXT[
One part of the lecture I give on software patents reviews patents that that cite no prior art of any sort on the front page - no patents, no journal articles, no conference proceedings, no mass media. I have long argued that such patents are of questionable validity and reflect bad faith on the part of the patent applicants.
One such patent in my talk is patent 5,208,748, an early groupware patent assigned to Action Technologies, one of a few groupware patents that Action has been awarded, titled "Method and apparatus for structuring and managing human communications by explicitly defining the types of communications permitted between participants", and has a sibling, patent 5,216,603, with the same title. Both patents have the same inventors and mid-1980s filing dates, but different claims.
Both patents have many characteristics that portend invalidity:
1) No patent or non-patent prior art was cited
No software idea is an island - there is always prior art.
In this case, ACM journals/conferences on office automation,
OA work, Xerox PARC, studies from the Human Computer Interface communities, etc., all are relevant to this patent.
2) The examiner was less than thorough
The patent's examiner, Gail Hayes, has been the examiner
for a number of software patents that cite no prior art of any sort (5,299,123 - 5,307,268 - 5,153,830 - etc.).
3) A lack of diligence is evident
One of the inventors, Carlos Flores, in 1982 submitted
a thesis to U.Cal Berkeley titled "Management and
communication in the office of the future". While only
a precursor to the techniques in the patents, it should
have been cited to help establish an obviousness basis.
Another inventor, Terry Winograd, is a professor at
Stanford (which has nice libraries), an author of many
journal articles, and a thesis advisor. In particular,
SU's libraries have both sources of the prior art items
the court invalidated the patents with.
The invalidated patent, 5,216,603, was a 1990 continuation
of a 1988 patent, which was a continuation of a 1985 patent.
Yet in 1986, the two main inventors, Carlos Flores and
Terry Winograd co-authored a book, "Understanding
computers and cognition", chapter 11 of which was titled
"Management and conversation", and which listed 162 items in the bibliography. In the 1990 application that became
the patent, neither the book nor any of the bibliographic
items were submitted as prior art.
Despite these questionable circumstances involving the prosecution of the patents, in November 1995, Action sued Novell for infringing the 5,216,603 patent. Not surprisingly, (given that Action didn't do a search before filing the lawsuit), last month the United States District Court - Northern District of California, invalidated the patent (5 June 1997, case number C-95-4093-VRW). What follows are excerpts from the court's ruling.
I find the ruling interesting because there are thousands of software patents of worse quality than the '603 patent (like the 3D pie chart patent) as measured by the amount of non-cited prior art divided by the broadness of the exemplary claim. In this case, it only took a few non-patent prior art items to invalidate the patent. Good detective work by Novell's lawyers (the ruling I have doesn't mention the law firms involved), and reasoning by the courts with regards to the commonality of computer applications.
====================
United States District Court
Northern District of California
June 5, 1997
This patent infringement action was filed by plaintiff Action Technologies, Inc ("Action") in November 1995. On December 3, 1996, the court issued an order construing the claim at issue and granting Action summary judgment of infringement. Action has now filed a motion for a preliminary injunction and defendant Novell, Inc ("Novell") has filed a motion for summary judgment of invalidity. Because Novell is entitled to summary judgment that the patent is invalid due to anticipation by the prior art, the court does not reach Action's motion.
[....................]
{Note: an explanation of summary judgments appears here.
[....................]
The dispute in this case concerns claim 1 of patent number 5,216,603 ("the '603 patent"), issued on June 1, 1993. The claim states that it encompasses:
1. A computer system which implements a method for assisting users of the computer system to manage at least one of their business, social and personal communications, wherein said computer system includes for each of said users a corresponding input and output interface, processor, memory, and storage device, said method comprising the steps of:
a) creating at least one conversation type record structure for at least one conversation, said at least one conversation having a plurality of states, said conversation type record structure for defining a type of conversation, an identification of each of said users who are to participate in said conversation, a role to be assigned to each of said participating users, the state of the conversation, and a specification of conversational moves permitted for said role at each state of the conversation;
b) creating at least one conversation instance record for a first one of said participating users and for at least one other of said participating users based upon data entered by the first user using said corresponding user input and output interface and the at least one conversation type record structure, wherein each of said at least one other participating users is assigned a predetermined one of said roles defined in said at least one conversation type record structure, storing said entered data in said corresponding conversation instance record for said first user and making said corresponding conversation instance record available to each of said at least one other participating users;
c) updating the corresponding conversation instance record for said participating users based upon data entered by each of said participating users using said corresponding input and output interface, wherein the data which is permitted to be entered by each of said participating users is determined by the user's assigned role and the state of the conversation when said data is being entered.
The court's order of December 3, 1996, construes various terms in this claim. The construction provided by that order governs the court's validity analysis. {Intervet America v. Kee-Vet Laboratories, 887 F2d 1050, 1053 (Fed Cir 1989).}
[....................]
{NOTE: A discussion of 35 USC 102(b) and 35 USC 282 appears here.}
[....................]
Novell claims that claim 1 of the '603 patent was anticipated both by description in several printed publications and by public use in this country more than a year prior to the date of the patent application. The court need not address the argument regarding publications because the evidence is clear and convincing that a program called MONSTR anticipated claim 1 and was in use, indeed on public display, more than a year before the application for the '603 was submitted.
The MONSTR program was developed by two men, Anatol Holt and Paul Cashman, as part of their work on a project developing an operating system called the National Software Works ("NSW") for the Arpanet, the precursor to the Internet. Novell has submitted two publications in which Holt and Cashman describe the MONSTR system: a paper entitled "A communication oriented approach to structuring the software maintenance environment" ("Communication Oriented"), published in the January 1980 edition of Software Engineering Notes, a newsletter produced by the Association for Computing Machinery; and a paper entitled "Designing systems to support cooperative activity: an example from software maintenance management" ("Designing Systems"), published in the proceedings of the 1981 Computer Software and Applications Conference held by the Institute of Electrical and Electronics Engineering, a professional organization. In addition to these publications, Novell has submitted a declaration from Paul Cashman ("Cashman Decl") describing aspects of the system and its development.
{Greg Aha note: Both publications are in the Stanford libraries and the 1981 CompSAC paper is in the Berkeley Engineering library, before the patent was submitted, and when the lawsuit was initiated.}
Holt and Cashman created MONSTR to assist in the software maintenance aspect of developing the NSW operating system for the Arpanet. Designing Systems ("DS") at 184. Essentially, the participants in the development of the NSW found their existing communications system inadequate and Holt and Cashman sought to create a more efficient way to handle communications concerning software maintenance. Id.
The program first went into use in July 1980 and all its major features were in place by the end of 1981. Cashman Decl at pp 8-9. The two papers cited above explained the nature and basic methods of the program to the general public in 1980 and 1981. Cashman states in his declaration that he demonstrated MONSTR to at least two groups of visitors at his company's offices. Cashman Decl at pp 23. In addition, the NSW itself was a reasonably public use of the system, as the project involved a variety of public sector and private sector entities. Cashman Decl at pp 4. These uses are more than sufficient to find that MONSTR was in public use more than a year prior to the filing date of November 15, 1985. Compare Beachcombers v Wildewood Creative Products, 31 F3d 1154, 1159-60 (Fed Cir 1994) (demonstration of invention to 20-30 guests at party is public use).
There is no question that MONSTR is "[a] computer system which implements a method for assisting users of the computer system to manage at least one of their business communications", as required by claim 1. Cashman states that the system was developed "to coordinate and discipline communication between end users, project managers and developers." Cashman Decl at pp 6. In the Communication-Oriented paper, the authors describe MONSTR as "a medium for controlled communication among project members who are carrying out their role-defined responsibilities." Communication-Oriented ("CO") at 13. The manner in which MONSTR functions, explained more fully below, buttresses the conclusion that the system is "for the purpose of" managing communications regarding software problems.
In order to anticipate, the system must implement a method which consists of a number of steps. First, it must create a "conversation type record structure". As noted in the court's December 3, 1996, order, this "record structure" is part of the computer program. Order at 9. A system which includes a "conversation type record structure" must define:
(1) a type of conversation;
(2) an identification of each of the users who are to participate in said conversation;
(3) a role to be assigned to each of said participating users; (4) the state of the conversation; and
(5) a specification of conversational moves permitted for said role at each state of the conversation.
The structure of MONSTR encompasses each of these elements.
[....................]
{NOTE: the ruling goes on to explain in detail how MONSTR does encompass items 1 to 5 in the above list, noting one objection from Action's expert:
Plaintiff's expert, Michael Malcom, asserts in his declaration that a user's organizational affiliation is not the equivalent of a conversational "role" as used in the '603 patent. In some systems, organizational affiliations might be entirely separate from conversational role, but not in MONSTR. The different organizations in the NSW serve differing roles vis a vis software problems. End-users might, for instance, find and report bugs, but would not be in a position to generate solutions........... Due to these distinctive roles, the organizational affiliation supplies a conversationally significant "role" or the user by identifying him as a generator of new STRs, a solution finder or a solution approver. }
[....................]
Having determined that MONSTR incorporates a "conversation type record structure", it is clear from the foregoing that the system creates this structure for "at least one conversation", and that "said conversation" has a plurality of states" as required by Claim 1. The next issue is whether MONSTR creates
at least one conversation instance record for a
first one of said participating users and for at least one other of said participating users based upon data entered by the first user * * * * and the at least one conversation type record structure, wherein each of said at least one other participating users is assigned a predetermined one of said roles defined in said at least one conversation type record structure, storing said entered data in said corresponding conversation instance record for said first user and making said corresponding conversation instance record available to each of said at least one other participating users;
The MONSTR system creates a conversation instance record called an STR. This use of the term STR applies the second of the two definitions outlined above at 8. The STR includes identifying information regarding the problem, called classifiers, and a history of the course of the communications, called action segments. DS at 187. "Each segment names an actions such as CREATE, DISPATCH or ACCEPT; the role player who performed the action (person's name and organization); the time of the action; a pointer to any text that may have been created, and information about the STR's state * * * * after the action occurred. Id. The various actions noted within the segment, as well as any text referenced by a pointer in the segment, constitute data entered by the first user and the other participating users. The entries in the segments are "based upon", meaning a product of, the data provided by the various users and the "conversation type record structure" which dictates the "moves" available to each user. Finally, the STR is available, at least in part, to all users through the STATUS and HISTORY commands. Cashman Decl at pp 17.
The last requirement of the claim is that the "conversation instance record" is updated based on data entered by the various users, where the data permitted to be entered depends on the role of the user and state of the conversation. As noted above, at least some of the data entered by the participants in an STR conversation is limited by the role of the user and the state of the conversation. Because there is really no distinction between the STR conversation and the conversation instance record, the record is updated with each piece of data added by a participant. The MONSTR system, therefore, satisfies this aspect of the claim as well.
In evaluating the evidence of anticipation offered by Novell, the court notes that the existence of MONSTR and the publications describing it were not brought to the attention of the patent examiner. See Exhibit C to Declaration of Laurie S. Hane. The patent examiner did not, therefore, have the opportunity to determine whether the work of Holt and Cashman anticipated the '603 patent.
In addition, the court notes that the distinctions between MONSTR and the elements of claim 1 suggested by the Action's expert were singularly unpersuasive. The bulk of his commentary was directed to the notion that MONSTR was developed using a different theoretical paradigm than was used in the '603 patent. The theoretical undergirding of the patented method are not, however, part of the claim. The court's concern is whether the program created by the designers of MONSTR followed the method set out in claim 1.
{Greg Aha note: additionally, some of the theoretical undergirding of the '603 patent was disclosed in Flores' 1982 thesis, reducing the theory's patentable aspects.}
The court has been presented with undisputed evidence regarding the existence, nature and public use of the MONSTR program. A reasonable jury could not conclude anything other than that Novell has proven by clear and convincing evidence that MONSTR utilized the method stated in claim 1 of the '603 patent. For that reason, Novell's motion for summary judgment of invalidty is GRANTED and Action's motion for a preliminary injunction is DENIED as moot.
IT IS SO ORDERED - Vaughn R. Walker - US District Judge |
