I couldn't pass up a discussion of the failures of the American patent system -- certainly not at the Red Hat Summit, where questions of IP law, licensing, and copyright are filling the air. What I got was not a fiery invective against the USPTO or a fulminating war cry against patents in general, but a much more nuanced and well-argued case for selective patent reform than I've heard in a long time.

The speaker, Michael J. Meurer, Boston University School of Law, is co-author of a book titled Patent Failure; his presentation was a condensed version of many of the assertions argued in that volume. He didn't need to persuade us that patent lawsuits filed in district federal courts are exploding, but the reasons this is happening had little or nothing to do with piracy or a growing contempt for IP rights.

The problem is actually far more complex, which is part of why it has thus far not been an easy fix. The patent system, Dr. Meurer asserted, does not work in its current incarnation because it fails to act like a property system. Worse, most of the people caught in its gears aren't deliberate offenders -- they're companies that don't do patent investigations because of the cost, time, and effort involved. (And even if their detective work turns up anything, it often isn't of much use.)

The core of the problem, according to Dr. Meurer, is that while patent law superficially resembles land law or property law, it breaks down in the specifics. The boundaries of debated patent filings -- what's covered, what's not -- are often poorly described, both in the original filing and in subsequent interpretations by the judiciary. Examples cited included a patent originally filed for a system to create custom mix tapes in record stores, which passed into another set of hands and was broadly abused by a patent troll outfit.

What to do, then? Spot reforms -- greater clarity of property rights, requiring more precise filings -- will likely achieve more real progress than junking the system wholesale, Dr. Meurer added. The system already has provisions for penalizing those who try to game it by filing overly broad patents, but those provisions are toothless and underused. Put some bite into them, whether at the judicial or legislative level, and a good many of these problems might be alleviated.

Maybe now that patent litigation -- especially involving software -- is at an all-time high and is creating real headaches for companies, individuals, and judges alike, it's high time for solutions that don't involve bulldozing patents as we know them.