May 19, 2004
Software Patents

Oh, shit: EU votes through software patent changes. This is bad.

Update 21st May: The Computer Weekly coverage is good: European Council approves 'controversial' software patent draft directive.

Via perfect.

Posted to Software development by Simon Brunning at May 19, 2004 04:02 PM


Posted by: Andy Todd on May 19, 2004 05:58 PM


Posted by: Simon Willison on May 19, 2004 06:59 PM


Posted by: Alan Green on May 20, 2004 12:47 AM


Posted by: Paul on May 20, 2004 09:05 AM

Eloquently put, guys. I forwarded this to El Presidente, who didn't see why it might be a problem. This is how I see it.

Currently, software is copyrighted. If someone else writes software, I'm only in trouble if I *copy* it without their permission. I can write software which does the same thing, either deliberately or inadvertently, and I'm fine.

Now, imagine software becomes patentable. Someone writes some software and patents it and the techniques they have used. If I later write a program which does the same things or uses the same techniques, then I'm liable, even if I've never heard of the original program, and invented everything totally independently. Ignorance is no defence. No one will be able to write software without the risk of being sued. This could kill off small software companies as well as open source. Only the big players would be in the game.

What's more, companies could (and will) patent techniques that are already in use. Think data cacheing, pooling, that sort of thing. If you were to infringe *this* patent, you *would* have a defence - "prior art". But you'd have to go to court and *prove* it; which only big companies can afford to do on a regular basis.

If this passes, we are all fucked.

Posted by: Simon Brunning on May 20, 2004 10:03 AM

The conversation between El Reesidente, Mark 'the burger flipper' Matthews and I continued. Whis is it, I was asked, that I feel that patents are appropriate in the case, for instance, of driugs, but not in the case of software? OK, I'll give it a go.

Drugs, as patented, consist of a single active component. These components are very expensive to develop, and even more expensive to test. Many drugs fail during testing, so the cost of developing a working, patentable drug is very high. In the absence of patents, no one would develop them in the first place.

A drug company is very unlikely to violate a patent by accident - in fact, it doesn't happen. It's easy to see if someone else has already patented something, and even big drug companies are only working a few new drugs at a time.

Imagine: drug company Foo invents and tests drug A. It costs them a fortune to do this, but it sells well. Without patents, drug company Bar can copy it, virtually for nothing, and sell it dirt cheap. Foo makes no money, and never researches another drug. This is bad for *everyone*.

A piece of software consists of a great many components. There is no realistic way to check if a new piece of software might violate a patent - not only are their a great many components, but each of them is a complex thing, and it's difficult to see how to check each of them.

For instance, when we wrote Evo, I'd not heard of MVC. (Paul, this a pattern for separating business logic from presentation logic.) Nevertheless, the prompts are written according to that pattern - I came up with the design myself. If someone held a patent on MVC, they could sue, and I can't see any way of checking for this. After all, we couldn't have just checked for an MVC patent - I didn't know it was *called* that!

Patenting of higher level concepts is more of a grey area - but we already have too much monopoly in software. I can't see that we need any more. Besides, copyright law already prevents direct copying: Software company Foo writes a "word frobinator". It costs them a fortune to do this, but it sells well. Software company Bar can't just copy it - that's copyright violation. They *can* write their own word frobinator, but that'll cost them money (and take them time) too. So, Foo gets its reward for coming up with the idea of the word frobinator - it's first to market. Bar *can* compete, but only if it has written a *better* word frobinator. This is fair enough.

Phew, better get on with some work now. ;-)

Posted by: Simon Brunning on May 20, 2004 11:55 AM

The difference is, as I see it, the patenting of actual things versus algorithms. It is difficult to accidentally copy a production technique (or an active ingredient in a chemical) but you can re-create an algorithm completely independently.

As usual the big fans of this are the large commercial and non-commercial organisations who can afford the extensive lawyers fees. Bastards.

Posted by: Andy Todd on May 20, 2004 07:52 PM

> As usual the big fans of this are the large
> commercial and non-commercial organisations who
> can afford the extensive lawyers fees. Bastards.

Well, Microsoft was one of the sponsors of the council meeting, so maybe the result shouldn't have been such a surprise...

Posted by: JanC on May 21, 2004 02:21 AM

Surely the old drug company anecdotes are so much hot air, anyway - I imagine that a lot of research gets done in universities with public funding (there was an interesting article in Wired with an example of this [1]), and that the real costs are in the trials and approval processes. Whilst I accept that someone has to put down the cash to "ready" the drugs for the market, and if a private entity does so then they should be reasonably compensated, the granting of patents which could then be applied worldwide is obscene. If GlaxoWhoever puts up the readies to test some drug or other, why not issue them with a temporary national sales monopoly rather than broad patent rights?

Indeed, what the patent fans either conveniently ignore (either through ignorance or by being distracted by the sensation of having wads of cash stuffed into their pockets by lobbyists) is that patents are convenient anti-competitive weapons, available for misuse by the most cynical of companies.


Posted by: Paul Boddie on May 21, 2004 11:41 AM

I would just like to raise the point that you are not patenting the code or the algorithm but the innovation behind the code- the idea of it.

This is a great tool for small software companies that have an idea to ensure that their investment in a project will come to fruition- to pay the employees, etc. Also, patents are even more valuable to small companies than large companies, because what other protections does a small company have? Patents level the playing field.

Posted by: H on January 27, 2006 08:38 PM
Post a comment

Email Address:



Remember info?