Patents have become a minefield that inhibit software innovation, a fact recently highlighted for me in a conversation with Rob Tiller, Red Hat's vice president of Intellectual Property and assistant general counsel.
Many software patents are of poor quality and are difficult to interpret, Tiller explains, made worse by the fact that patent boundaries are often vague: a patent can even cover an invention the patent holder never conceived.
Compounding this morass, it's difficult to impossible to know if code you have written could be covered by a patent, as search methods are unreliable (and damages are structured such that it may not even be advisable to search because you hit treble damages for willful infringement, as one Microsoft employee has counseled in the past ("Ignorance is bliss and strongly recommended").
Even worse, unlike copyright law, which is more suitable to software, patent law isn't concerned with knowledge of infringement: you can create a completely original invention with zero knowledge of someone else's invention and still violate his patent.
What to do?
One option is Microsoft's: engage the industry in cross-licensing agreements. These may simply be nothing more than safety blankets, as there's really no way of knowing the quality of Microsoft's patent portfolio in advance of being sued and having a court decide.
Another is to engage in disarmament, which is unlikely but probably the right thing to do. The software industry doesn't need patents. Patents are actually counterproductive to innovation.
In fact, as Tiller explains, "Most of the important enterprise software in wide use today rests on ideas that predate the explosion of software patents in the 1990s." Lotus 1-2-3? Released in 1982. Microsoft Word (a facsimile of WordPerfect)? 1983. Oracle version 3? 1983. Windows 3.0? 1990.
Several of the world's largest software franchises, in other words, were built without the aid (or inhibition) of patents. Microsoft, for all its recent bluster about patents, is probably the world's biggest beneficiary of a patent-free development environment.
Of course, post-Bilski, we may be entering a period of court-ordered disarmament, which would be fantastic. The Bilski decision puts software patents on the defensive, and it hopefully will help to clear the minefield that currently helps only incumbents--and arguably hurts even them more than it helps them.
In the meantime, it would be fantastic to have more community-reviewed patent applications, since the U.S. Patent & Trademark Office seems incapable (due to both expertise and workload) of adequately reviewing the flood of patent applications. Something like the Peer to Patent project, which ex-Red Hat general counsel Mark Webbink runs at New York Law School, perhaps? Or how about Post-Issue Peer To Patent and Linux Defenders?
We have a serious patent problem on our hands. It's time to do something about it.
Follow me on Twitter @mjasay.