Story: Software's nuclear winter is at hand

• 
• 

> Software patents are not automatically evil.
> It makes little sense to say that if you build
> something ingenious, non-obvious yet
> useful from hardware then you can patent
> it, while exactly the same functionality in
> software cannot be protected.

< SNIP >

> The trouble comes from a lack of expertise
> and foresight in the way software patents
> are issued. More than 20,000 new
> software patents are passed every year in
> the US: this is either an orgy of fecund
> invention or a cynical push for software
> development's nuclear winter. The evidence
> to decide which is at hand. Take a look at
> the software you use every day. Any
> evidence of a tsunami of fresh technology?
> No. Was the software industry pre-patents
> a sluggish, unprofitable, ineffective mess?
> Again, no.
>
> Software patents are failing because they
> are being misapplied, and they are being
> misapplied because the US patent office is
> unable to detect true ingenuity. Instead, it is
> left to the courts -- and by extension, those
> best equipped to pay for them -- to apply
> the basic checks of novelty, utility and
> invention.

This "how do you distinguish hardware and software" line is an analysis that indirectly emphasizes "protection" rather than the fact that patenting abstract instructions given to a generic logic device is an absolute impediment to innovation. This is not to speak of questions regarding freedom and a certain kind of understanding which people have been encouraged to forget of late.

The reason why there's so much talk about software patents right now is because in Europe they are not legal and there is presently very clueful opposition being kicked up to efforts to make them legal.

Europe has a much stronger tradition of recognizing that abstractions are not patentable.

In Europe right now, the EU Commission has been trying to tell the public that they are not trying to allow patents on software "as such," while at the same time trying to pass a Directive making "computer-implemented inventions" patentable.

Incidentally, I'll tell you the reason: it's just because their key performance indicator for "innovation" under their "Lisbon Strategy"
for the European Union, includes a simple count of "patents!"

Note that right now, in Europe, individual computer users may hold that the presumption of the law is expressly on their side, that they may program their computers without regard for the ridiculous prospect that someone may claim a patent on the logical algorithms that they thereby express.

A patent on software in a device is a patent everywhere. This is because software is abstract, and that's what happens when you patent abstraction. Making software in a device patentable, is not in any way different from making software patentable.

It's not just because that's what a patent is, and patents are an unvarnished good. An abstraction is a general rule by its very nature, freely available for anyone's use. A patent is a rule under the law.

It's not that you can patent hardware, but you can't patent software; it's that you can't patent abstraction. The nature of the medium in which you express your ideas makes no difference. A specific kind of hardware, the
generic logic device, has brought about the existence of software, the ability to code such a machine in those terms. The generic logic device provides abstract logical functions that may be directed by sequences of instructions which are themselves inherently abstract. The reason not to patent software is because it's abstract -- it's logic! Whatever level language you use and whatever medium you express it in, it still translates into logic. Code is certainly embodied in all sorts of physical forms, but it is a representation of abstract ideas, just as surely as written mathematical calculations are.

It does not even matter if the abstraction is a major, breakthrough discovery of a nonobvious solut

• 
• 

• 
• 
• 

Full Talkback thread