Thursday, April 20, 2006

Definition of "patent troll"...

In the wake of yet another "patent troll" case, Microsoft, Autodesk lose patent case, I decided to take up Matthew Buchanan's challenge to define a patent troll. (I've submitted the following as a comment as well).

I'll take a stab at a definition of a "patent troll":

A company, individual or other entity whose current sole business objective is the exploitation of a patent or patent portfolio acquired after (but patented before) the adoption of a technology to extract royalties or settlement agreements at values calculated to be near the cost of patent litigation.
Under this definition, companies can move from "legitimate" inventive businesses into the realm of "trolls" and vice-versa. I think that I've incorporated several important distinctions:
  1. Simply seeking out licensors and/or infringers cannot be enough to be a troll;
  2. Simply seeking reasonable royalties cannot be enough;
  3. Simply defending one's actual business by using a patent or patent portfolio cannot be enough (purpose of; and
  4. Being a invention farm cannot be enough.

I think that this captures companies that are essentially patent holding companies, with no real business objective except the acquisition of patents (usually not of their own creation), to enforce against wealthier, productive entities.

The Slashdot crowd may not like this def'n since they use the term to include lots of entities--or even one commenter using the term against universities.

Frankly, I'm not sure that there is such a thing as a patent troll, or even if they're necessarily a bad thing. I think largely the idea of a patent troll is really about perception. The small patent owner is always a troll when you're the infringer.

Additional material: I might also add to the definition that the patent or patent portfolio often contain weak or seemingly invalid patents. By quickly glancing at the software patent at issue in the Microsoft/Autodesk case, I'd have to say that Microsoft/Autodesk are correct that these patents, 6,044,471 (filed 6/4/1998) and 6,785,825 (filed 5/31/2002), seem invalid. For one thing, the first patent seems to cover methods of "product activation" long used by shareware software.