Software Patents

Tom Lee has some more background on software patents, for those of you who seem to think these are a good idea. Remember that, yes, stronger IP protections create larger financial incentives to innovate. But they also make it more expensive to innovate and ultimately reduce the level of competition. You can already copyright your software so people can't just go and copy what you've done.

Adding a layer of patent protection on top of that makes life too difficult for programmers (who don't know what they're allowed to do and may not be able to afford to do it) and too comfortable for incumbents.

Comments

You can already copyright your software so people can't just go and copy what you've done.

I don't think the enforcement mechanisms are really robust and easy to enforce. From what I understand, the code itself is seldom released for commercial software. It's possible to borrow a big chunk of someone else's software and bury it in yours, without anyone knowing about it (unless maybe they sue you).

That said, it sounds like software and tech patents are really getting a bit out of hand.

Posted by ben | May 16, 2008 4:57 PM

Matt, I am a computer scientist and have earned my living by programming and worked in the IP business. You are of course totally right that patents do not encourage innovation in programming--innovation in trying to get bog-standard techniques past patent examiners perhaps. They bog programmers down in the bureaucracy of processing patent applications and trying to work out what patent applications have been filed by other programmers. They stifle innovation in the art and have been intensely disliked by programmers who generally think of them as a fraud. (At least that is how it used to be; some younger ones have been bought off with patent bonuses).

Other than that they are mainly a mechanism for corporations to attack individuals and other corporations.

Posted by Chris Dornan | May 16, 2008 5:03 PM

A couple of questions:

1) If software can be copyrighted, then why should we be concerned that it can be patented too? Patents are for a fairly limited period of time (used to be 17 years; I'm too lazy to see if that's changed); copyrights are for essentially forever - life plus 70 years, or some similarly monstrous length of time.

I agree that being able to do either one is stupid: with the life cycle of software being as short as it is, even 17 years is close enough to forever so as to make no difference.

2) When you copyright or patent something, don't you have to show the world what it is that you're copyrighting or patenting? Given what ben said about the weakness of enforcement mechanisms and the ability to bury code, it seems to me that it would be easy enough to steal copyrighted or patented code to use in an application, and just not copyright or patent your own code.

Posted by low-tech cyclist | May 16, 2008 5:14 PM

I'm not entirely sold on software patents being a good or bad thing... Here are a few scenarios:

Small company innovates in some way... big company copies the innovation and pushes the small company out of the market.

Small company innovates. Big company separately innovates. Small company sues big company and basically makes innovation challenging.

Unfortunately, I'm not entirely sure what the best mechanism is.

Posted by Joe | May 16, 2008 5:16 PM

low-tech cyclist - Yes, a 17 year patent is an outrageously long period of time in the software world for ideas to be off limits. As someone who writes code for a living, I can tell you there is a constant, if subtle, fear that I will implement some great new idea which I thought was original, only to find that someone else came up with it independently and has patented it. As a little guy, I'm not so much worried about being sued as I am being denied the ability to use my innovation.

But this is only a problem with patents, and not copyrights. If somebody else copyrights some code, thats fine, I can't use their code without a license. But I can write my own version - possibly a better version. But with patents, the entire line of innovation becomes either off limits, or very expensive.

Posted by Spike | May 16, 2008 5:34 PM

Copyright is kind of a dumb way to protect software. So too is the patent.

There is nothing (legally speaking) to prevent the creation of a sui generis form of protection geared to the way software innovation really works. Congress should hold some hearings to educate itself on how such a regime could be constructed, and should make it so.

Posted by dj moonbat | May 16, 2008 5:38 PM

I've been a code monkey and systems analyst for 20 years. I second Spike's comments.

I'll reiterate what I said yesterday: the examiners at the USPTO are overwhelmed. So, even if we grant that s/w patents are a good tool for advancing the art and science of software, at the moment they're granting patents on a raft of vague, broad, obvious, and/or widely used algorithms and techniques.

Posted by cmholm | May 16, 2008 5:58 PM

As for the need for software patents: I'll argue that most of the real innovation in software these days is released under the General Public License (GPL). This commie/libertarian concept allows you to use GPL-licensed software in just about any way you want.

The caveat is that if you make changes to GPL'ed s/w and want to sell, distribute, or embed it in a widget that you're going to sell a million units of, you need to make the software's source code available to others to copy... just like you did when you originally copied it.

The GPL recognizes that everyone stands upon the shoulders of those that came before them, and looks to multiple the shoulders.

Posted by cmholm | May 16, 2008 6:13 PM

The problem with software patents is that software models the way the real world operates. So there is no real invention. This is analogous to translating a program from one language to another (or a book). The original is what is/was unique.

But there is an even more dishonest set of software patents, which abuse the reason for patents: to improve upon the state of the art.

On example is the Amazon 'one-click' patent. The fact that Amazon reduced the number of steps to place an order was considered worthy of patent protection. But this is the whole point of programming in anything but machine language. In fact the point of programming is to simplify something, so that isn't a good reason to provide patent protection.

Posted by tomj | May 16, 2008 7:31 PM

Spike - that makes sense. Thanks for reducing my ignorance!

Posted by low-tech cyclist | May 17, 2008 7:22 AM

I agree with Spike that the idea of software patents is flawed, but add that the implementation is worse still. Patents are granted for things that are obvious and are re-invented independently over and over, or worse still for techniques that are already common practice. The result is that any sizable program contains patent violations. The only way to avoid that would be to have a second staff the size of the programming staff constantly doing patent searches.

My favorite example of a software patent: Picture an old-fashion black-and-white monitor with a cursor that's a rectangular block. If you make the cursor black, you can't see it against a white background. If you make it white, you can't see it against a white background. How to solve this intractable problem? After about 30 seconds reflection, I think we'd all come up with making it black when on a white background and vice versa. Unfortunately, IBM has that idea patented.

Posted by Mike | May 17, 2008 11:35 AM

The biggest problem with software patents are the patent trolls. Among companies with actual products, patents are a bit of unnecessary friction, but don't have any lethal side effects. Each company wastes some time and money building up a portfolio of mostly trivial patents. If company A takes its bunch of patents and sues company B, B then countersues A with its bunch of patents. The lawyers from the two companies confer and settle out of court, with a manageable sum of money changing hands. There's waste, but neither company's core business is effected. But the patent troll is a parasite genetically engineered to live inside the software patent system. They have no business and no intent of ever starting one. Their only business is suing people based on a bunch of (bogus)patents that they bought. Thus they can't be countersued and a comapany's patents don't protect it against them. When Research in Motion settled with the troll that had sued it, it paid out more than it had earned in its entire corporate history. So despite the fact that RIM had done the hard work of building the business, the profits went to the troll - and all because of a software patent that had subsequently been declared invalid. As an economy, we'd be better off with no intellectual property at all - in the current system, filing a patent is more lucrative than building a business.

Posted by Jim | May 18, 2008 1:24 PM