[ejames]

In theory, a patent is supposed to be specific enough that the patent itself accurately describes the object, to the extent that a reasonably knowledgeable person could actually manufacture the object using only the patent as a blueprint. In other words, you would not be able to write the patent application unless you had actually invented the object in question, because otherwise you wouldn't know how to describe the theoretical object well enough for a person to manufacture it. Therefore you cannot patent imaginary future inventions.

In practice, this rule is effective only to the extent that the Patent Office enforces it.

[suprasanna]

As you point out, the issue is that the UPO doesn't enforce that as rigidly as it should (or maybe, is reasonable to expect). As tjr pointed out in a separate thread, 'This American Life' did a great piece on software patents specifically, definitely worth a look if some of you are interested in learning more. Certainly gave me better perspective. http://www.fsf.org/blogs/community/tal-when-patents-attack

[lmarinho]

Building on that, I pose a question. Has anyone here ever used a software patent as a reference to help solve a problem?

[glimcat]

Software, no. Hardware, fairly often.