Having worked in a research lab before (HP not MS), I really feel their pain with this one. Typically product divisions are completely consumed with making incremental improvements to existing products. Even if they do recognise the work you've done as strategic to their business they are often unwilling/unable to do anything about it.

I think the best option for a research lab is to have a more formal marketplace for ideas. Product divisions can bid for an idea, but also ideas can be spun out into new buisiness areas, or even related startup firms. Olivetti research in Cambridge (UK) did a good job with this, they spawned quite a few innovative and successful firms. Researchers who really believe in an idea are often keen to move to a spinout to make it a success.

As a previous poster pointed out, anti-patents already exist in the form of "technical disclosures".

A few years back when I was working in a research lab, even if we weren't going to go the whole hog and patent something, then we would make sure that the idea was published as a technical disclosure. Once this was done then basically the idea was in the public domain and was not patentable.

-- Harry