Google Patents Caching MLK Day Search Results 113
theodp writes "Google remembers Dr. Martin Luther King, Jr. not only with its Doodles, but also with its patents. 'Right around the Martin Luther King Holiday,' explained Google in its application for a recently-granted patent on Discovery of Short-Term and Emerging Trends in Computer Network Traffic, 'there may be many searches about "Martin Luther King"...Thus, it would be useful to have better methods of detecting short term trends for the purposes of caching search results to making them more readily available to users.' You may call the invention of detecting and caching 'MLK Day Sale' search results patently obvious, but the USPTO calls it U.S. Patent No. 8,082,342. Hey, at least it's arguably better than the patents issued to Microsoft and Google for avoiding walking or driving down Martin Luther King Boulevard!"
Re:U.S. needs to get rid of software patents (Score:2, Informative)
Companies like Google and Apple are collectively abusing the system and patenting every single thing they can think of, most of which are outright obvious.
Google isn't abusing the system. They're trying to protect themselves from people who are abusing it, like Apple and Microsoft. The patents that they have filed and acquired are solely for defensive purposes. Google has never used a software patent offensively, and is very outspoken about the need for patent reform: https://www.google.com/search?sourceid=chrome&ie=UTF-8&q=google+patent+reform [google.com]
As usual, not a vague patent. (Score:5, Informative)
The patent as written claims the use of a particular formula (Which I read as "N'=N/Df^((T'-T)/Ti)") to predict trends.
It is not a patent on prediction in general, or prediction with computers, or caching, or caching based on predictions. It is a very specific design of a non-obvious system, applied in a specific way. Implement the system differently, and you're not violating the patent. The MLK mention is an example, which in no way affects the actual claims. In fact, it's a trivial example, as well. Here's an excerpt containing all references to MLK from the patent itself:
Short-term trends are, however, important to consider, as they are often the result of external activities dominating the time of day and date, as well as current events. For example, during the days preceding and following a space shuttle launch there may be many searches relating to "space shuttles," "NASA," "space," and similar terms. Right around the Martin Luther King Holiday, there may be many searches about "Martin Luther King." If a celebrity was just arrested for drunk driving and assaulting a police officer, it is reasonable to expect a significant increase in queries involving the name of that celebrity. Thus, it would be useful to have better methods of detecting short term trends for the purposes of caching search results to making them more readily available to users.
Sorry, but the typical Slashdot patent hate is yet again unjustified, and the reference to MLK is likely unintentional, as the patent was filed in December of 2006 and granted in December of 2011. It looks like the submitter just search for "Martin Luther King patents" and wrote a story around the results. Well done, sensationalist headline!
Re:Way to be racist slashdot... (Score:5, Informative)
Neither. Given the widespread white racism at the time of King's death, it was politically impossible to rename any street held dear by the white majority to honor the man. Pretty much without exception, MLK Blvd runs through a black section of town. Economic realities being what they are in the U.S., that section is frequently run-down.
Sucks, but that's how it is in the places I've been.
Re:hmm, patent-commenting system (Score:0, Informative)
Sans the "patent it yourself" part. I recently started the process of publishing my own thesis; one thing the university actively encouraged me to consider was delaying the release date so as not to infer with pursuing a patent on some of the material.
I told them where to stuff it - I don't think mathematical truths, no matter how clever I might think they are - should be patentable. The point remains: once the thesis is published, that counts as prior art against any patent I might pursue. I imagine there's some short window where I could still go after it - but I don't think 1997 would be included in that window.
Still - requesting re-exam and invalidation based on previous work would be a noble pursuit for citizen-technologists.