Patent Wars

And so it begins.

Slashdot posted an article today about some patent claims against Open Source developers. They linked to an article by Bruce Perens, a well known OSS advocate, detailing some of the issues surrounding 2 particular patent cases currently pending. The first case is a recent case against RedHat regarding their Hibernate software package. Firestar Software is claiming that they hold a patent on what they call Object Relational Mapping. If I understand correctly, this is a programming technique used to hide the implementation details of a database behind an object. In other words, it’s basically encapsulating the database within an object.

Umm.. yeah. Duh. Ok, so let me get this straight. If I create an object in a programming language that can be used within the program to prevent having to write direct SQL calls, then that falls under this patent? Well, I guess I’ll have someone banging on my door pretty soon. phpTodo uses this same technique! Isn’t this an obvious extension of the object-orientation paradigm in most modern programming languages? It’s the next logical step from creating procedures or functions to accomplish the same thing!

According to the article by Bruce, there is plenty of prior art that covers this. And rightly so! The problem here seems to be the US patent system as a whole. Patents on their own seem, at least to me, to be something useful. At least, useful to a degree. I don’t hold any patents so, if anything, I’m biased against the system. But I do see some worth in it. I can see the need to defend a new, unique idea, at least for a time. However, it seems that patents are being granted on the most ridiculous things! For instance, check out patent number 6,368,227. WHAT? Are you kidding me? A patent for swinging on a swing? Sure, it’s side to side instead of the traditional forward and backward swing, but give me a break. I did this when I was a kid, and probably in the same manner.

Check out this excerpt from the patent itself :

“It should be noted that because pulling alternately on one chain and then the other resembles in some measure the movements one would use to swing from vines in a dense jungle forest, the swinging method of the present invention may be referred to by the present inventor and his sister as ‘Tarzan’ swinging. The user may even choose to produce a Tarzan-type yell while swinging in the manner described, which more accurately replicates swinging on vines in a dense jungle forest. Actual jungle forestry is not required.”

It seems to me that the patent system needs a major overhaul. I swear I’m not trying to jump on the bandwagon here, but when larger companies start leveraging these ridiculous patents, I get a bit scared. I’m just as open to getting sued as RedHat is. I think most of the uproar over the Firestar patent has to do with them suing an Open Source company, but the same remains true for any other company. For instance, the patent dispute against RIM. My main issue with that case isn’t so much the content of the patents, but rather the company that held the patents. NTP is a holding company. The entire reason NTP exists is as an entity that owns patents and collects fees based on usage of those patents. From my point of view, this is extortion. Basically, these companies hold the patents and require the user of the patent to pay fees for continued use. But they never use the patent themselves! In fact, given the task, I doubt any patent holding company could ever hope to implement any of the patents they hold.

But even patents that are blatantly obvious and are easily overturned are still extremely harmful. The second case that Bruce mentions is against a small open-source developer, Bob Jacobsen, who makes no money from his creation, JMRI. KAM, the company that filed the claim, holds the rights to patent 6,530,329 which outlines a method for sending commands from a computer to a model train.

This *sounds* like a patentable idea to me, but, upon further inspection, they haven’t really invented anything. First, they seem to be using pre-existing hardware and merely writing software to control it. Second, it’s basically a queueing system. Essentially, the patent outlines how a queue works. User 1 sends a command and the digital controller sends an acknowledgement; A second user sends a command and the same process occurs; And so and so forth. The interesting part here is that the patent language makes it a point to explain that these acknowledgements are intended to inform the user that the action requested has taken place, when, in fact, it it merely queued. I can think of some other ways to do this, but the idea generally works. So where’s the new invention? It sounds to me like they took a pre-existing system and added a queue. That’s patentable?

So, because they have this patent, they have decided to sue Mr. Jacobsen. They are asking for $19 per user of JMRI. I’m not entirely sure how they determined how many users JMRI has, but my guess is that they merely looked at the number of downloads the software has received. It looks like version 1.4 received about 11000 downloads which is about right for the $200,000 they’re apparently asking for. However, it appears that there may be plenty of prior art to fght this claim, so what’s the big deal? The problem here is that Mr. Jacobsen probably doesn’t have a few thousand dollars lying around that he can use to defend himself. Depending on how the lawsuit proceeds, it can possibly take several months or years to either overturn the patent, or lose the case. Either way, it would cost Mr. Jacobsen a lot of money he likey doesn’t have.

This type of patent abuse only serves to hurt everyone in the long run. Some developers may stop developing, or at least stop releasing their code out of fear. If small developers can be sued like this, even for patents that were so obviously granted without proper review, then they run the risk of losing more than just the right to develop a product. OSS developers are usually independent and don’t have the luxury of a corporate umbrella to protect them. They run the risk of losing everything they own. Something needs to be done about this system.

Here are some of my ideas for patent reform. They are listed in no specific order :

• Existing patents should be re-examined for validity.
• Any patent over a certain age should be considered public.
• Any patents held by companies that are not implementing them should be given two choices. Either start working on an implementation of the patent, or sell the patent to a company that will implement it. Either way, a deadline should be set to prevent the company from sitting on the patent. If they exceed the deadline, the patent should be placed into the public domain.
• All new patents should be scrutinized for validity beyond the current methods. If insufficient expertise is available at the patent office, then an expert in that area should be consulted.
• All new patents should be open to public review. (I believe this is already the case, but I may be mistaken)
• All granted patents should have a shelf-life. This shelf-life should be the same across all patents regardless of what the patent is on.
• Patents on software should either not exist at all, or should be very critically and very carefully reviewed before being granted. There are too many ways patents like this can be exploited.

I’m sure there is a lot more that should be covered, but this, at least, is a start. This would put everyone on a level playing field and help prevent the stifling of innovation. Let’s get real here. If patents such as the Object Relational Mapping patent are allowed to survive and are enforceable, then innocent developers such as myself and others are in danger. I have no prior knowledge of the existance of that patent, and I never would have bothered to check. This, to me, seems to be a common sense bit of programming!

Hopefully we’ll see a larger movement to reform the current patent system, or to do away with it entirely. While there is worth in the system as it is today, I think it has much more potential to do harm.