Patent examiners are not the enemy, and someone needs to do the work regardless. I feel bad for them myself, because their tooling is apparently terrible. Makes me wonder if I should take a look into the space and see if I can come up with some ideas for a halfway decent set of tools.
I've seen a lot of patents whose prior art could be found by spending a bit of quality time with a thesaurus, or a halfway decent index.
> I feel bad for them myself, because their tooling is apparently terrible. Makes me wonder if I should take a look into the space and see if I can come up with some ideas for a halfway decent set of tools.
The search tools are acceptable, but could be improved a lot. To get ideas for improvements, you should talk to actual examiners and try searching for patents yourself. Unfortunately, too frequently people who don't know much about how patent searching actually works propose "improvements" that aren't actually improvements.
While well-intentioned, this site is doomed to mostly be unused by examiners. The classification search doesn't work. Classification search is a critical feature for patent examination. And the classifications are produced by machine learning, which usually produces poor quality classifications (despite loud pronouncements about how great machine learning is for this task).
The internal search tools work mostly by keyboard, and this search site works mostly by mouse. Mouse is much slower in my experience, and this matters a lot for time-constrained people like patent examiners. The main advantage of the internal USPTO search tools is speed, not anything fancy: https://news.ycombinator.com/item?id=30978043
The documents on the "Prior Art Archive" have tons of broken images, too.
> I've seen a lot of patents whose prior art could be found by spending a bit of quality time with a thesaurus, or a halfway decent index.
Most of the time this is not the case. If it's something that simple, it would be easily rejected. Keep in mind that when the media says a patent covers X, it probably doesn't actually cover X. It probably covers something far more specific that isn't a problem for anyone.
(Again, like my other comments here, this is just my opinion, not that of the USPTO or US government.)
