Let's review
Let's review the idea behind patents -> that they are necessary to promote R&D in areas where the results may be copied for little to nothing; they are meant to protect the inventor, and secure their future by way of completing their designs,and manufacturing them (whatever that entails).
So, are software patents succeeding here? Anyone?
We know that software mutates a lot faster than hardware: the rate of change is every 1.5 years, or somewhere thereabout; we know that the USPTO can take 3-5 years to send out a FOA, and possibly several more years to grant the patent; as such, unlike in many other industries, by the time a patent is ready to be used, the software industry may have already run the gambit, and implemented the various routines contained therein, banking that the patent application would fail. This contributes to the perception that these patent lawsuits are insane, as an entire industry can spring up before they are told "this thing is patented."
Where does this leave us? Well, the question of whether software patents should exist is up in the air; they are meant to promote the arts and sciences, and seem to be getting in the way of them; however, the question then is, is this because software people are so good at reverse engineering their competitor's code (when they see something they like), or is it because the various algorithms / code / designs are so trivial? The software industry does have some very intelligent people, but then, it also has tools which can decompile a program into code in a matter of minutes, and give us access to various hidden or obfuscated methods; I should know, I've used several of them before, and it's almost trivial to figure out what the decompiled code does, even allowing for the lack of symbols, etc.
As for the USPTO, they have some serious funding problems, which is kind of sad since they tend to bring in a lot more revenue through voluntary means than what they spend. The USPTO, IMHO, needs a lot more funding, and a lot more examiners; the waiting period for a software patent, if we are to keep them, should be no more than a year at most, from receipt to publish, if they are to be of any use to anyone. There are an awful lot of startups out there who would find their second or third rounds of funding easier to secure if the products they are developing were not so...well, 'left to the wind' as it were; investors might be more forthcoming if they knew that a startup actually had some assets to back up their bold venture. As it is, I imagine a number of startups file for software patents, and go bankrupt, for lack of being able to protect their new market from better funded also-rans; by the time the patent is published, that's all that's left of the company, which the added insult that it may be sold off to patent trolls to offset losses.
Whatever the case, this is probably not how software patents should be used, within the various contexts of the meanings of their existences.