Saturday, February 14, 2009

This IBM patent clogs sewers:

One rumored candidate to be the next Director of the Patent and Trademark Office is David Kappos, one of the head patent lawyers at IBM. Given IBM's many abuses of the patent system and patent policy over the past few decades, I think it is inappropriate, nay, wrong, for anyone from IBM to be head of the PTO. Might as well as make Bernie Madoff head of the SEC as part of his upcoming jail-time work-release program. IBM patent lawyers for too long have abused the patent system. Case in point. Last week IBM was issued its usual batch of patents, many of which are crap - crappy patents whose sole value is to clog the PTO's patent examination pipelines to the detriment of everyone else. A patent application policy actively embraced by David Kappos. If I was an IBM investor, I would applaud David for doing his best to help IBM. In fact, sometimes I recommend people to buy IBM stock because the company will do anything to maintain its market value. Kudos to David for his efforts in this regard. But the reward for abusing the patent system for the benefit of IBM should be a gold watch at retirement - and should not be the reward of being appointed head of the PTO. One of the many crappy patents issued last week is the one below, an expert system at a server that analyzes incoming messages (such as news), checks lists to see which subscribers want to be alerted to such messages, and sends the message to such subscribers. That is, methods Marimba used (and patented) to push software in 1996, IBM decides six years later to patent as methods to push other stuff. As usual, the patent cites an inadequate amount of non-patent prior art, based on IBM's inadequate IDS and the examiner's lack of experience in searching the non-patent prior art. Further, it looks like this patent, crappy as it is, was a First Office Action issuance - usually a good indicator of crap. After the application was submitted, an amendment was submitted, maybe with a bit of talking with the examiner, and then the patent was allowed to issue. Don't you all wish you had such clout with the PTO to get such crap issued? Part of the problem is that the claim language is IBM's usual excessively wordy, baffle-them-with-bull's shirt, confusing verbiage - again, another IBM patent policy actively supported by Kappos. Let's look at the abstract, which is almost easy to understand, and then claim 1, which gives me a headache to read:
United States Patent 7,487,550 Methods, apparatus and computer programs for processing alerts and auditing in a publish/subscribe system Abstract A message broker receives a published message from a publisher program. Responsive to identification of one or more subscriber programs subscribing to messages of the type of the received message, the broker forwards the received message to the one or more subscriber programs. Matcher components compares the received message with stored subscriptions to identify subscriber programs, generates an alert when an alert condition is satisfied, and compares the generated alert with stored subscriptions to identify subscriber programs subscribing to the alert. The alert is then forwarded to the subscriber program subscribing to the alert.
The patent only cites five or six prior patents, inadequate, grossly inadequate, especially in light of a Sun Microsystems patent whose title kills IBM's patent outright:
United States Patent 5,761,662 Dasan, June 2, 1998 Personalized information retrieval using user-defined profile
Now, for those of you who know anything about expert system database alert systems (i.e., a database with alert triggers, a decades old field), look at the crappy non-patent prior art considered:
Other References "Design of a General Clinical Notification System Based on the Publish-Subscribe Paradigm", A conference of the American Medical Informatics Association. By, A. Geissbuhler, M.D., W. W. Stead, M.D., Oct. 25, 1997, pp. 126-130, XP002179981. cited by other . "Exploiting an Event-Based Infrastructure to Develop Complex Distributed Systems", Proceedings of the 1998 International Conference in Kyoto, Japan Apr. 19-25, 1998, Los Alamitos, CA, USA, IEEE Comput.Soc, US, Apr. 19, 1998, pp. 261-270. cited by other . Icc.net Internet Commerce Corporation on website ICC.net/Services/Infosafe and ICC.net/Services/Infosafe/
Technology, 2001. cited by other . Arnold et al, "Discourse with Disposable Computers: How and Why You Will Talk to Your Tomatoes", USENIX Proceedings of the Embedded Systems Workshop, Mar. 29-31, 1999. cited by other.
Not one article from the any ACM publications and conferences on database systems, alert systems, message analysis systems, push, or expert systems, and only one article from the IEEE. Grossly inadequate, and IBM knows it. To cite nothing from SIGMOD or DEXA makes this patent crap. So maybe IBM should spend less time doing searches against other companies as part of its scam public patent review project, and more time doing searches of its own crap. Excrement, as in the language of claim 1 (the only claim as well, probably too long to not be workaroundable and thus unenforceable):
The invention claimed is: 1. A data processing apparatus for providing a publish/subscribe message dissemination service on behalf of publisher and subscriber programs comprising: means for receiving a published message from a publisher program; means, responsive to identification of one or more message subscriber programs subscribing to messages of the type of the received message, for forwarding the received message to the one or more message subscriber programs; and one or more matcher components for: comparing the received message with stored message subscriptions to identify the one or more message subscriber programs; generating an alert when an alert condition is satisfied; and comparing the generated alert with stored alert subscriptions to identify one or more subscriber programs subscribing to the alert; and means for forwarding the alert to the one or more subscriber programs subscribing to the alert, wherein the message and alert subscriptions are stored in data storage in association with message topic information, the one or more matcher components including means for retrieving stored subscription information by reference to message topic information of a received message, rules procedures for generating and determining required dissemination of alerts are stored in association with the message topic information,
NOTE: so far, this claim is nothing more than the many push systems popularized in the Internet era - no innovation here. And as a good example to deceive the Patent Office, the word "push" doesn't appear in the patent, depriving the examiner the opportunity to think about PUSH and search for Marimba's patents.
wherein the one or more matcher components are adapted to identify a relevant rules procedure by reference to the message topic information and to forward to the identified rules procedure: a message subscription list; a list of authorized recipients; and an identification of one or more subscribers for alerts; thereby to enable generation and determination of required dissemination of an alert; the one or more matcher components includes: means for performing an authorization check to identify a subset of the identified one or more message subscriber programs which subset of programs is authorized to receive the message; and means for generating an alert when the authorization check identifies an unauthorized message subscriber.
NOTE: actually, even this far, this claim is nothing more than the many push systems popularized in the Internet era - no innovation here. This patent is nothing more than an IBM ploy to clog the patent system with a patent application embodying little to no innovation - a tactic IBM has used for thousands and thousands of patent applications. A tactic warmly embraced by David Kappos. Who should not be appointed next Director of the PTO.