> You don't understand how patent trolling works, do you? The point is to make it so expensive and painful for the innovator to continue without giving the patent troll a big slice of the pie that the innovator capitulates and pays off the troll.
It is impossible to do anything today without infringing on somebody's bullshit patent. However, one can identify patents that can be demonstrated to be bullshit for a reasonable cost.
There is also an element of poker. If someone has a bullshit patent, but they have successfully used it, or think they can, then they risk losing that revenue if it is struck down as invalid. They do have something to lose, in addition to legal fees.
Trolls live on fear. Don't be afraid.
So back to these particular patents: They are bullshit. They are obviously bullshit. I would represent myself if hit with them. I think Google might be able to find a few people better than me to prove it.
I suggest that these differences are what makes them different as applies to patents. Patents are usually either so broad as to have prior art, or so narrow as to not apply to even "similar" implementations.
6,192,476: Too broad. Applies to any OS since the 70's. 6,125,447: as above. 5,966,702: Too specific (.NET doesnt need to optimize data it creates properly in the first place) and too broad at the same time: it simply describes data normalization. 7,426,720: Unix forks. Either way, .NET does not do this. RE38,104: Not applicable at all. .NET is always fully JITed. 6,910,205: See above. 6,061,520: Specific to JVM implementation detail.