In the field of IR I was considering US8214361 B1: if I can resume the claims in two words the invention is a search engine interface that not only provides a set of results but also a set of categories in which results are grouped, and it allows the user to customize the categories so that the next time she would search she will be provided the customized set of categories.
My questions are:
how this can be considered not obvious? It's a simple combination of features well known to everybody. Isn't it like to say that I want to patent a car painted with yellow and purple stripes: just because nobody did before is it an invention?
in the case somebody would like to implement such a system the point actually would be the algorithm that makes possible to categorize search engine results in a meaningful way. In this invention instead this algorithm is completely abstracted. Isn't it like to say that I want to patent time machine. A system and method for travel in time, not saying anything about how this would be possible?
if such kinds of patents are granted, given that somebody estimates that every day 110 software patents are granted by USPTO, then how to be sure that every kind of web interaction somebody would want to implement in her web application isn't yet patented? Patent search isn't always reliable and it can also be that patent application hasn't been yet published. I mean: if somebody puts money and efforts to develop a web product there's the risk that somebody else has filed the same idea less than 18 months before. This risk can be acceptable if limited to real invention (how the time machine work) but instead becomes too high if extended to such broad concepts (time machine itself).