Brad B never clearly stated what distinguished software patents. (The USPTO hasn't figured it out either.) He said software patents didn't cost millions of dollars to develop nor was the technology subject to approvals. But those statements are incorrect. Additionally many non-software patents didn't cost millions of dollars or were subject to approvals. He seems to accept other categories, but he doesn't define what he doesn't like.

Maybe there are more junk software patents than other patents. A question of degree, not kind. Big difference. If so, instead of wiping out the whole deal, why not work instead to find ways to blow up bad patents (software or others)? And improve ways of determining reasonable royalties.

Can a good "software" patent exist? If so, why would you want it not to?

Some examples to contemplate: (1) the RSA patents, which enabled Internet commerce among other things; (2) various GPS processing/monitoring (not hardware) patents, which help keep planes from crashing (accidentally); (3) early image compression patents, which enabled digital video. These patents were developed by staffs of many Ph.D. researchers who were dedicated for years to their research. Isn't that stuff okay?

To amplify – software patents don't just say "here is an algorithm for processing GPS data." Rather, they say, "the very idea of processing GPS data are belong to us."

Nope. Patents and copyrights do different things, and pretty much none of the technology I listed can be protected sufficiently by copyrights. Did RSA, Garmin, Allied Signal, Honeywell, Bendix, and Compression Labs just not understand copyright?

For more venture capitalists against software patents, see this (archived) PR that used to be hosted by MySQL (before they were acquired by Sun): http://web.archive.org/web/20070804233425/http://…

One of the signatories of the PR, Benchmark Capital, was the main investor in, a.o., eBay.

They don't patent because they "understand" anything. They patent because they can.

I guess the USPTO thought that technology was patentable.

SHOULD it be patentable? Doesn't sound like it.

Plenty of bad patents are out there — obviously. That's why I like working on ways to blow up bad patents instead of throwing out the idea of patents altogether. Some new case law makes it much easier to blow up bad patents.

Patents are not the problem. BAD patents and meritless litigation are.

Our society through legislation gives certain monopolies to developers of intellectual property. To me, the only valid reason for doing so is a cost-benefit analysis. If we grant developer A a legal monopoly (patent, trademark, servicemark, copyright) for X years concerning idea B, then overall we will benefit more than if we do not.

Patents for inventing drugs are to me a good idea. Without patents, a lot less drugs would be invented and drugs are hugely expensive to develop (in part due to the FDA approval process). By allowing patents, a lot more drugs are invented and the benefit they provide vastly outweighs the extra prices we all pay because the drug manufacturer has a monopoly for X years.

When Brad says software patents are a bad idea, I think what he is really saying is, "On a cost-benefit basis, we would be better off if there was never a legal monopoly in the form of a patent for software." Although a few more good ideas might be generated, the benefit of those few ideas is much less than the harm caused by software patents. That is a position I agree with.

The problem is compounded by the fact that there is absolutely no doubt that the patent agents who work in the software department of the patent office are total absolute idiots, literally some of the dumbest people ever born. Is there a software patent they will not grant, no matter how blantantly obvious?

In addition to the stupidity of the patent office when it comes to granting patents, software ideas are in so many cases cumulative — standing on the shoulders of giants.

It makes no sense to try to fix the system, it is not fixable. It's better to simply abolish software patents (and business process patents), retroactively to the first one granted. Very little if any harm will occur and vast benefits will accrue.

As for the RSA and GPS algorithms, I am certain if RSA did not write these algorithms, someone else would. RSA can still make lots of money from its work without patent protection — almost any company would rather deal with the inventor directly, even if they have to pay a fee.

James Mitchell
http://www.jmitchell.me

Follow-up discussion at http://www.feld.com/wp/archives/2010/02/sawyer-we… .

Brad,

Another great post on the problems of software patents. What's missing is a concrete plan for getting things changed – at its heart it's a *political* problem.

It's possible to get some fixes from the Supreme Court (the KSR decision was a step in the right direction), but ultimately any change has to be passed by Congress or rules changes need to be implemented by the PTO.

What would really help move the debate forward is not just suggestions on how to "fix" the "problem", but a realistic set of steps to pull together the political support to improve the situation: who the major players are, how their support can be obtained, and what compromise will build a coalition that is likely to gain majority support in Congress.