Update – 2024-05-09:

In the 14 years since I posted this, I’ve changed my mind. Without patent protection against low-effort knock-offs, it can be really difficult for small firms to get VC funding. At the very least, I overstated the case.

Somewhere else I’ve said “Everything is more complicated than it seems.”. This is one of those.

…indeed the best thing about patents is that they eventually expire. (And I say this as an inventor with multiple issued patents.)

See, for instance, http://news.cnet.com/8301-10784_3-57374941-87/litigation-lunacy-silicon-valleys-lost-its-collective-mind/?part=rss&subj=news&tag=2547-1_3-0-20.

Or http://en.wikipedia.org/wiki/Nest_Labs#Litigation.

And in general: http://www.techdirt.com/.

Having said that, I guess I have to also say that I’m not against patents in principle, only in practice.

I usually try to stay out of politics on this blog, but this affects us nerds in particular.

If we could somehow have a working patent system that limited patents to truly original and (above all,) non-obvious inventions, ideally ones that involved genuine investment (instead of off-the-cuff ideas), then I’d be in favor of that. But the current system is supposed to do that already, and fails miserably.

The result is worse than no patent system at all. Ironically, the current patent system, which is supposed to encourage innovation, instead stifles it – the risk of company-killing lawsuits over genuinely independent inventions (and therefore, in my book, obvious ones) far outweighs any encouragement.

In my view, to qualify as non-obvious, an applicant should be required to show that her invention solves a long-standing (not recent) problem which other people have had ample opportunity to solve, but have been unable to. Too many modern patents are obvious solutions to new problems which either never existed before (because a new technology raises new problems) or which only recently became solvable because of new technology. For example, sending voice over the Internet is obvious once you have an Internet. Nobody should get a patent on that just because they were the first, as this is obvious. Heatsinking a LED for domestic lighting is obvious – no patent should issue simply on the basis that nobody did it before, as that is because LED use for domestic lighting is a new application, and therefore the problem never came up before.

Apparently somebody was issued a patent on using a rotating mirror to scan a laser in a polymer 3D printer – because it’s a “new application” of the invention of the rotating mirror. There’s at least one guy who therefore is using a rotating prism instead, to work around the patent. In his own words “even though this would have been obvious for any bachelor physics student looking into this topic”. Using a rotating mirror (or prism) this way is not in any sense “invention” – it’s workaday engineering.

Solutions should be considered obvious if they appear very quickly after appearance of the problem, or if multiple independent “inventors” come up with the same solution over a short period of time.

While I’d prefer real reform of the patent system along these lines – which would reject 98%+ of currently issued patents (including most of mine) – political reality seems to make that unlikely. Given the choice between the current system and no patent system at all, I’d choose none.