I had not read that. Thank you. Those guidelines, if followed, will address some issues with software patents, but not the problems I mentioned. This is understandable, because the problems I mentioned are intractable.
The key problem you mention, of not having enough specialist knowledge, applies to /all/ patents. Actually, like every other field, 99% of engineers will understand 99% of patents. The specialist areas are the small minority. I'm sure I can find 100 UI or web patents for every sound or graphics one... for instance. So I agree this is an intractable problem, but I disagree that its a serious one...
It's not a problem of understanding the patents. The problem is determining novelty and non-obviousness. There are so many open-source jQuery plugins coming out all the time that a web/UI patent's novelty is going to be very hard to determine. Similarly, this ongoing flood of progress will help give new ideas to ordinary skilled programmers, rendering lots of potential "inventions" obvious.
I still contend that this is a serious problem in a legal system that includes presumption of validity.