EU Ministers Went Off-Brief In Patent Vote

MartinB writes "Several EU ministers reportedly went against the wishes of their nations in voting for the proposed EU Software Patent legislation in May. Among those misleading the council of ministers were representatives from Holland, Poland and Germany. The Dutch parliament is going as far as asking to change its vote, which was originally in favour of making software patentable."

Comments

Some questions ...

[s20451]

I'm not trolling. But I think the software patent issue is more complicated than you think.

Let me start by saying that I think submarine patents are a gross abuse of the system and should be abolished. And I agree that, in general, intellectual property law needs to be reformed. Both these ideas are beyond the scope of this post.

Now, you ask the average open source advocate what s/he thinks about software patents, and s/he will be opposed to them, on the grounds that they stifle innovation. I can buy that to some extent. However, if you ask the same advocate why s/he wants a particular patent invalidated, it's usually to copy a patented algorithm and incorporate it into an open source product. That doesn't seem like innovation to me. It's true that open source would let others learn about the algorithm and improve on it, but there's nothing preventing you from studying a patent -- in fact that's the whole point of the patent process. If you're keen enough, you can take the ideas in the patent and implement a free work-alike (like png, gzip, or the free equivalent to rsa), and innovate away.

Looking deeper, I don't see that it's consistent to be in favor of patents but opposed to software patents. This is because software blurs the line between a device and a description. For example, consider an integrated circuit. This is clearly a device, and hence patentable, right? But it can be described using a language, such as VHDL. In fact the VHDL can be used with a programmable chip to instantly implement a work-alike to the device. Hence, if the chip is patentable, the VHDL should be patentable too.

It's as though you had a description of a tool (a drill, say), which could be instantly implemented on some universal machine. The description is only trivially different from the tool. (This may sound ridiculous, but with 3-d printers and related technology, the day may not be far off when we see such a thing.)

To take a Touretzkian view, this means that either all patents are valid (including software), or none are ... though I would prefer a middle view that recognizes software as a distinct hybrid of an "idea" and a "device". Such a legal status would also resolve the question of "is code speech?" with the answer "yes and no". (Can you tell I'm Canadian?).