Prior art patents offer a wealth of information that goes untapped for virtually all inventors. There is a very small minority of inventors that actually search other patents looking for ideas.
One reason for such a minority is that searching prior art patents is a very difficult task, and one that should be much easier than it is.
One problem with searching is that the classification system is perpetually behind, especially in rapidly evolving areas. While it is understandable that new patents and publications are being added to evolving areas very quickly, there is a limit to how large some classifications should be allowed to grow. When a subclass has 1000 or 1500 patents in it, there is almost no likelihood of being able to read, understand, and digest that amount of information within a reasonable time.
Another fundamental problem is that the classification system may not be able to fully capture the content of a published application or patent. One application may touch on 10 or 15 different classes and subclasses, but may only be classified in two or three.
These problems will never be solved, mostly because the classification system’s evolution will naturally lag the inventions and no classification system is ever perfect. There will always be holes in it.
In many classes and subclasses, such as simple mechanical devices, the classification system is very highly developed and very effective. For new fields such as telecommunications and software related technologies, the classes are unmanageable and make searching a hit or miss proposition.
If I were king for a day, I would ask for much more detailed subclasses and begin to whittle away at these giant subclasses with 1000’s of patents in them.