If you cannot understand how people could possible think that software patents are acceptable, I recommend reading this article. [1] Not because it will convince you that software patents are good — it won't. But it will help you understand the thought process of lawyers who do think that there should be some software patents. For example, Goetz writes that he supports "patents where the preferred implementation of a…
Personally, I support software patents. I'm just against idea patents, as opposed to invention patents. Screen rotation is no invention. It's an idea; once you have it, it's trivial to implement. That, IMO, does not afford patent protection. However, certain things can, and should be patents. RSA, for example. It's a non-trivial procedure that was non-trivial to invent and there are not very many alternatives.
For instance, determining the optimum layout of a mail application for a four inch screen probably takes a lot of tweaking and user testing. But there's no way that you can call it an invention, and your competitors are free to copy the layout, so long as they change the style enough that customers can clearly distinguish the products.
On the other hand, using a magnet to hold a power connector to a laptop seems like a really obvious idea once you've seen it. You don't have to study detailed documents to see how you could make it. But calling it an 'invention' seems reasonable - and it is patented. And it's perhaps the only thing that attracts me about Apple laptops, but that's not important right now.