The Economist has two recent articles on the US patent system and its
problems. One, titled Patent
wars, describes the increasing business importance of having a
patent portfolio and the rapidly growing rate of patent applications and
grants. The other, titled Patent
nonsense, talks about a few of the more well-known bad software
patents, and ends on a note of warning suggesting that the system should
be reformed. Both pieces ran in the Apr 8-14 issue of The Economist.
Another curious fact exposed by the articles is that Amazon has a
relatively new patent on
referral click-throughs, issued Feb 22 of this year.
My guess is that, by the time the dust settles, nearly everything
interesting to do on the Internet will be patented by someone or
another. You won't be able to swing a cat without running into
infringing a patent.
It's not at all clear to Advogato what's the best thing to do
about the patent mess. In the meantime, I think, staying informed and
tracking the important patents seems like a good thing. And you can't
beat David Turner's approach - simply create something new
that gets around existing
patents. That's also a large part of the motivation behind the xiph.org projects.
(Hi lkcl!)
There is nothing like encountering a software patent in one's own project
to give one a visceral disgust
for the whole system. I would explain the problem domain to people,
and have them say ``ah, that's easy, you just need to do X''. But sadly
I'd already thought of that, and it was outlawed because there was an
incredibly broad patent already covering the idea.
My personal favourite ridiculous patent is this,
by Unisys.
We are heading down the Road to
Tycho.
Seriously, I think the issue here is largely one of good patents vs. bad
patents. I personally think the idea of patents is a pretty good one,
and, if implemented reasonably well, would be a nice tool to compensate
inventors for their work.
Of course, as we all know, the current situation at the US Patent and Trademark Office is
royally screwed up. Thousands upon thousands of overly broad, trivial,
or already-invented patents issue every year. The cost of playing the
patent game puts it well past the means of many small inventors, while
at the same time patents are becoming so important for business that
it's almost a breach of fiduciary responsibility for big companies not
to participate.
What, in my opinion, differentiates a good patent from bad?
There is work involved
So many of the bad patents are for just the idea of doing
something. I think a good patent should reflect actual research and
implementation work. The result of sitting around bullshitting new ideas
should not be patentable.
There is cleverness involved
Even if the implementation takes work, a straightforward implementation
shouldn't be patentable.
It is not overly broad
Some patents are issued with relatively narrow claims that protect the
specific invention at hand. Others have broad claims that effectively
prevent anyone else implementing the same idea in a different way. Most
of the responsibility for the overly broad patents lies with the patent
office - if you're going to go to the trouble and expense to get a
patent, naturally you're going to want the broadest claims you can get.
Note that these criteria aren't all that different than the stated USPTO
guidelines, which call for new patents to be novel, unobvious (to "one
skilled in the art"), and useful. To these criteria, I'd add another
that has less to do with the patent itself than the context it's used
in:
Patented technology doesn't belong in standards
Any new standard that requires patented technology to implement should
be subjected to intense scrutiny. A patent-free alternative, if
available, is almost always preferable. Much of the trouble that patents
have caused for free software have to do with their inclusion in
standards. Just think GIF, MP3, RSA, etc.
Lastly, I'm in the process of granting rights to use all my patents in
GPL software. It's time-consuming and costly to do something like this
right, so it's not going as fast as I'd like. But they will get there.
[ Home | Articles | Account | People | Projects | FAQ ] 
del.icio.us
Digg
Google bookmark
reddit
Simpy
StumbleUpon
Furl
Newsvine
Technorati
Tailrank