Take the 2-minute tour ×
Ask Patents is a question and answer site for people interested in improving and participating in the patent system. It's 100% free, no registration required.

US patent #6098065 is the first in the series of patents used in Rockstar's legal action against Google and Android device manufacturers. It is for an "Associative search engine" and contains just one claim:

A method of searching for desired information within a data network, comprising the steps of:

  1. receiving, from a user, a search request including a search argument corresponding to the desired information;

  2. searching, based upon the received search argument and user profile data, a database of information to generate a search result;

  3. and providing the search results to the user

wherein searching the database includes

  • correlating, as a function of a fuzzy logic algorithm, the received search argument and user profile data to particular information in the database, and

  • providing the particular information as the search results.


Off the cuff, the key here would appear to be the "fuzzy logic algorithm", as the rest of the claim certainly was used prior to February 13, 1997 (including various bits of prior art mentioned in the abstract, like Yahoo).

I am interested in collecting relevant definitions of "fuzzy logic algorithm" and any prior art for the same. It is unclear to me if stemming or synonym matching would qualify as "fuzzy logic algorithms" or not, and if they do, whether there is documented evidence of such techniques being used in database searching prior to February 13, 1997. I would assume that simple wildcards (e.g., SQL LIKE operator) would probably not qualify.

share|improve this question
    
So every time I use Sed,AWK, and/or grep, I am infringing on this patient? –  hdrins Oct 10 at 14:20

2 Answers 2

Fuzzy logic would involve assigning a probability of a hit that was between 0 and 1. Just using stemming or synonyms to decide if something hit or not wouldn't be fuzzy. Giving a record extra "points" for having multiple synonyms hit or for degree of proximity of terms or having terms in a title vs in a footnote would be fuzzy, in my opinion.

While it makes sense to look for a "point of novelty" in trying to understand a claim, they are taken as a whole. That is, there may be no single new element in a claim but it can still be novel and non-obvious as a whole. Is a wind surfer like a sailboat but flatter or like a surfboard with a sail added?

Besides the fuzzy search it may have been novel at the time of filing to take user profile information into account in interpreting a search string.

share|improve this answer
    
Good points -- thanks! –  CommonsWare Nov 1 '13 at 17:08
    
I don't think I concur with your analysis. First fuzzy logic is not about probability of "hit", it's about degree of "hit". e.g. "This water is 95% potable", if we were talking probability, then there would be a 1/20 chance that the water is cyanide. If we're talking fuzzy logic, we're saying that it is (with 100% certainty) something like beer, something that is mostly potable, but has some things in it that aren't exactly good for human consumption. –  McKay Nov 4 '13 at 13:57
    
"Giving a record extra points for having multiple synonyms..." wouldn't necessarily count as fuzzy really. The whole point of searches, especially on large databases is not about what the hits are, but how good the hits are. e.g. no one cares if they are indexed on google, they want to know if they're on the top page (or 2?) of google, everything after that mostly gets lost. Standard search engine techniques like stemming, and number of occurrences don't use fuzzy logic, at least any more than standard search techniques available since the 60s. –  McKay Nov 4 '13 at 13:58

I first thought of 'spamassassin', which is a tool to do fuzzy email filtering, but is newer than '97. However, the program is based on Jason Rennie's technology published in 1996:

The first known mail-filtering program to use a naive Bayes classifier was Jason Rennie's ifile program, released in 1996. The program was used to sort mail into folders.[1]

http://people.csail.mit.edu/jrennie/ifile/old/README-0.1A

I would suggest this was close enough: the claims basically say: a search engine that does fuzzy matching, and 'ifile' is just that.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.