IBM, USPTO Trying To Improve Software Patents?
from the maybe,-maybe dept
Recognizing some of the criticism facing the patent system, it looks like the US's largest patenter for thirteen years running, IBM, and the USPTO are trying to work together on a system that improves software patents, without having to reform the entire patent system through laws. The NY Times has more details -- but it's tough to tell from these two articles how helpful this will really be. There are three parts to the plan, and the first two are nice to have, but merely deal with some of the symptoms, not the root causes of bad software patents. The first is that people can sign up for email or RSS alerts on certain types of patents, so they can find out about them quickly. The second is an "open source as prior art" database, that will contain plenty of open source software that could be useful for patent examiners in determining if there's prior art -- though, not necessarily if an idea is "non-obvious" to those skilled in the art. The final piece of the puzzle is the one with the fewest details. It's the Patent Quality Index, which claims to be a quick and dirty algorithmic method of giving your patent application a quality score. The idea is that patent applicants can run their patent through this system before they submit it for real, and can be quickly told that their patent is lame. Of course, that all depends on how the PQI works -- and so far, that's still a big secret. It's also not clear if it's just patent applicants who will use this system, or if patent examiners will also use the PQI to "score" patent applications. Obviously, that might be tempting since patent examiners don't scale, but it probably puts way too much faith in a algorithm. While it's nice (really!) to see at least some attempt at more innovative ways to deal with the patent issue (other than "throw more money at the patent office"), it seems unlikely that this is really going to help all that much. Update: In the comments, someone suggests that the PQI isn't algorithmic, but rather a group of volunteers who rate patent quality. Would be nice if there was something clear to back this up.