Ok, I made that last part up. But it’s only .2 degrees more ludicrous than what IBM was actually granted a patent for has applied for a patent on.
Ian Foster points to a smelly piece of patent action by IBM, and says really all that needs to be said:
We all know that the US patent system has its weaknesses: that every now and then, people get granted patents for things that are either well known or obvious.
…IBM has finally gotten around to applying for a patent on grid computing. For example:
a computer-implemented method of providing access to grid computing resources available to a plurality of users comprises receiving, from a requesting entity, a request to use a specific grid computing resource to perform a defined function; and routing the request to the specific grid computing resource in accordance with the request.
It’s not April, but I still read the patent grant application. Ian’s not fooling us.
The base-10 bit was closer than you thing… There were concerns about IBM’s DPD encoding in the IEEE-754r work. But people at IBM (e.g. Mike Cowlishaw) went far out of their way to produce a guarantee from IBM that DPD would be available to anyone with no royalties. Sounded like that required a lot of legal footwork and convincing inside IBM. This then forced Intel’s later BID contribution to have the same terms.
It’s an application by IBM to the USPTO. It’s not a granted patent. The two things are significantly different.
David: you are absolutely correct. I blocked the word “application” when I read the document. I’ve updated the post.
I haven’t waded through the whole thing, but the claims look to be more along the lines of scheduling into a grid computing system than grid computing itself. Either way, there’s a lot of published prior art out there going back quite a few years that many might say indicates the technology in the application is well known.
HPCer: that’s the same general conclusion that Ian drew in his post as well. I’m going to be interested to see whether this is a successful application. Not having been in a company that does as much patent work as IBM, I don’t know what percentage of their applications turn into granted patents, but on the other hand they have a LOT of patents so one suspects that they wouldn’t apply for something they didn’t feel like they’d get.
Just guessing, though.