Indeed, but why should the algorithm be protectable? What if the great mathematicisions of our history had "patented" their mathematical concepts, their proofs, etc. Would we be where we are with mathematical and scientific knowledge? I doubt it. I just don't see the value in extending from "protection of the unique expression of an idea" to "protection of the idea itself". The unique expression is the true creation of man and should be protected as "property" so long as we have a concept of such a thing, IMO. But the idea is no-ones property and cannot be proved to be such. It is a much more arbitrary way of dealing with ideas IMO.
I don't quite agree with your wording ("the unique expression is the true creation of man"), because that suggests that ideas are somehow less creative, and I don't think that's true. However, when it comes to protection, I do agree that protecting ideas is much more problematic than protecting unique expressions of ideas. The main problem is that a single idea could be used in a number of unique expressions, and so protecting an idea can hamper a lot of further work, which can be undesirable.
And I guess, when it comes down to fundamentals, I'm generally against protecting ideas. The way I see it, protecting an idea gives you two things: 1) The inventor of the idea gets a fair shot at exploiting the idea, and 2) Extensions of the idea and unique expressions of the idea tend to be restricted too much. If the benefits of (1) were more significant than the issues of (2), then I might be for protecting ideas, but I'm not convinved that the benefits of (1) are that great in reality.
I generally believe that the inventor of a new idea will tend to have a natural advantage in terms of exploiting it anyway, since he will have a head start and was smart enough to come up with the idea in the first place. OK, so there will be cases where a large company will see someone's great new idea, steal it, and overtake the inventor, but inventors can combat this to some extent by not releasing details too soon, and history has also taught us that radical new ideas are rarely stolen, mainly because people don't believe in new ideas until the evidence is staring them in the face.
The problem with this viewpoint, though, is that some ideas are so fundamental that they have lots of applications besides the one the inventor was considering at the time of invention, and the inventor's natural advantages won't normally stretch to all applications, so something like patents are the only way to give them a chance to exploit these other applications first as well. However, you could argue that this gives the inventor more than a "fair shot" at exploitation, but I guess different people have different definitions of what a "fair shot" actually means. I do think the current patent systems go far beyond any reasonable definition of fair shot, though.
That said, I'll happily admit that my opinion of how things should work is pretty idealistic, as is much of the anti-patent talk I've witnessed. There's nothing wrong with idealistic opinions, of course, but I think we'll have to accept that the reality of the situation is very complex, and that any change will be very slow and will probably only ever go so far.