Impossible software

If you produce software – any kind of software – the issue of patents has probably come up. The fact is that it is impossible to produce a piece of software that actually does anything practical without infringing on at least one patent – and for complex systems, probably hundreds of them.

Let’s say you wanted to write a quick program to index music collections. Congratulations, you’ve just run afoul of… umm, let’s see… I count at least four patents for that, at a glance. What about indexing images? Well, depending on how you do it there are at least 30 patents that apply to that.

Think of something obvious you can do with software. There’s a patent for that. Probably a lot of them. You probably owe someone money.

Worse, if you stopped to check what patents existed that might cover what you’re doing, you owe three times as much as you would have if you didn’t, because then it becomes wilful infringement. Ever wonder why most software developers have a rule against checking the patent registries? That’s why.

Technically, the rules are that you can’t patent without demonstrating a physical invention; that you cannot patent something obvious; that you cannot patent a process.

The rules, then, don’t work, because the majority of patents filed at (eg) the USPTO are obvious processes without a physical invention. There are loopholes. Huge ones. In some cases, patents can be completely rewritten to cover something else, without changing their original date.

Worse, there’s a tangled web involved. You might have patented something – which gives you the right to prevent anyone else from producing it for the term of the patent… but that doesn’t mean that you can produce it yourself. Your patent might be partially covered by someone else’s patent, and they (in turn) have the right to exclude you.

Say, just as an example out of the air, your patent covers a relational database to index and check Web URLs to issue a warning to a user before they wind up at a malware site. You can prevent anyone else producing such a system without giving you satisfaction first. However, using a relational database to check Web URLs at all might be someone else’s patent (there’s dozens of those actually), and the database itself might be covered under yet another. So, while you could prevent someone else from making the thing you patented, you can’t make it either, until you’ve satisfied other patent holders.

How does this work in practice? Well, in practice (as you have probably gathered by now) pretty much all software infringes on patents, usually in the hope that nobody will notice. If software makers followed the law as the law exists, there would be no software makers.

Meanwhile, some businesses exist solely for the purpose of coming up with random ideas, filing patents on them and seeing who ends up falling foul of them later on, then suing them for money. That’s called “growing your IP portfolio” or “patent trolling” – take your pick. The business doesn’t intend to produce anything, other than throwing more or less obvious ideas together and suing people who end up having the same more or less obvious ideas.

I firmly believe that a patents system is a good idea, and that such a system serves a good and necessary purpose. The system that we have though? That isn’t it. The current system is a convoluted cesspit that stifles innovation, and paralyses industries who aren’t willing to break it. That’s just not right.

Tags: , , , ,

Categories: Law, Opinion.



Got a news tip or a press-release? Send it to [email protected].
Read previous post:
Close