Linus leads charge against software patents

[arthur]

signal processing can be written down as mathematical formulas.

 

I think the "gray area" you mention is that computer code sometimes works more "mathematically" than "mechanically" which is where the current patent system is based on. I mean, the design for a DVD system (high-precision stepper motor, laser operating at around 700nm, etc) works well with patents. However, computer code may go like..."generate 2 prime numbers m and n, then raise the data value by the power n..."

 

But the patents you've mentioned you said yourself that there are hundreds of ways to circumvent. Then how could a software patent be in any way protective? Although I agree, trying to "get around" patents will jumpstart innovation in the FLOSS community. Although I think it already has a lot of that.

 

And it may be just the mathematical property of programming that's driving the programmers on the anti- side absolutely nuts...

[aRTee]

If you can implement similar functionality, like ogg vorbis is similar to mp3, you still are not using the IP of the developers of mp3.

 

However, you can still not make any mp3 capable player without violating the patents.

 

You can, however, create a competing standard. Which means you have to do the hard work of figuring out lots of things, instead of piggy backing by using the results of the existing work.

 

Ogg vorbis has many advantages over mp3, but it took a very long time to create a decoder that works in integer mode only (necessary for portable devices - think ARM7 devices), and ogg vorbis en/decoding takes more calculation power.

Because of this it's behind in the race with mp3. So technologically, mp3 has some big advantages.

 

With copyright, you can't protect mp3 technology - one would only have to write their own code and it's done. So copyright doesn't suffice.

Here patents play a big role - no one can leapfrog / avoid development cost due to patents. But you can invest time/money/effort yourself and create a competing system (ogg vorbis), then try to get that used.

 

The basic fear is this: that companies manage to protect an idea that only needs the time to develop the code -- there I don't see why copyright doesn't suffice. The 'copying' party will have to write their own code, so there is no need for extra protection - there is no competitive disadvantage for the original developers if others are allowed to incorporate the same functionality (written by themselves) into their software.

 

If all you need to do from the general idea to the implementation is to write the code, I'd say this is trivial (obvious) and so it violates the rule of nonobviousness.

So from the general idea of having to click only once to add something to your shopping list to the implementation can be seen as trivial.

 

On the other hand, from the general idea of saying: we're going to use the human ear model, do subband coding and compress audio/music by a factor of ten, and still have the result sound acceptable, to having an implementation of this is imho _not_ trivial and needs lots more work to be done than just coding...

 

Hence there is a true investment that needs to be protected, else the competition can just pluck the fruits of the labour of the developing party and copy the whole thing.