Slashdot Mirror
EU Board Votes To Allow Software Patents
scamp was one of the folks who wrote from Europe with the news that
an administrative board for the European Patents Office has voted 10-9 to allow patents for software in Europe. There's still a final conference to be held in Novemeber to ratify the decision - so there's still time to sign the petition against it. The conference in November should be close - the multinationals, US and Japan are applying heavy pressure. BTW, if you can't read German, use the fish.
I can't believe that so many governments could be bought out by corporations. Like it or not, computer code is a written work, not a device. The only "device" involved in the execution of software is the computer itself, which reads the instructions embodied by computer software and takes actions accordingly.
:= 1;
/* MaryLamb.c */
Written works are certainly protectable by copyright, and this is a Good Thing. A person's work should be protected. But copyrights, patents, and trademarks are meant to be mutually exclusive (you can't patent something that you could instead copyright, nor copyright something that should instead be trademarked, etc), and for a good reason. To claim patents on written works is to claim ownership of thought itself (not a specific thought, mind you, but ideas themselves); this can never be allowed.
How to prove software is a written work? How does this strike you:
"Mary had a little lamb."
This is obviously a written work. Well out of copyright in this specific instance, but that's irrelevant; all we're looking for is written work (and preferably a short one, as we're going to be doing a lot with it). Were I to use the whole rhyme, and were it an original work, it could be copyrighted.
María tenía un cordero pequeño.
The same sentence, translated into Spanish (my Spanish is, sadly, somewhat rusty, so forgive me if I got something wrong). Still obviously a written work; it's well established that a translation of a written work is still a written work.
program MaryLamb(output);
type
lamb: integer;
var
Mary: array[1..10] of lamb;
begin
Mary[1]
end.
The same thing, translated into Pascal (as with Spanish, my Pascal is very rusty, so I'd appreciate any corrections). 1 is certainly a "little" value relative to the set or integers, or "lambs" as I'm calling them here, so we still communicate that Mary had a little lamb. It's been established that source code is a written work. Translation, therefore, still has not violated that, and source code as itself should, as a written, copyrightable work, not be patentable. But let's take this a little further...
#define LAMB int
int main() {
LAMB *Mary;
Mary = new LAMB;
*Mary = 1;
return 0;
}
A very rough translation of the MaryLamb program to C (and once again, my C is rusty). Translation from language to language doesn't make this a non-written work, does it?
"Zmel umq m yvggyr ymzo."
The English sentence again, but now it's ROT13'd. A simple one-to-one mapping of characters onto other characters; in other words, a cipher. However, it has been established that putting a written work into a cipher like this still does not affect its status as a written work.
"Blue green red yellow orange."
Same sentence, but I've mapped whole words now instead of characters. I've encoded it. But again, I haven't affected its status as a written work. I could get really crazy and map it to, say, Japanese katakana, but I'm not good with katakana and I would be willing to bet that the browsers of most Slashdotters wouldn't be able to display it anyway. But this is another type of encoding, where I've mapped parts of words (the syllables) rather than the whole words. And I still haven't affected the status.
Now, let's go back to the C-language version. Suppose I were to translate that into assembly (substitute your favorite architecture's ASM here; I'm not about to start flamewars by picking one in particular). Still human-readable, still the same message. And still not in dispute over whether or not it's a written work.
But now, let's run that through an assembler. What does this do?
In simple terms, it maps the individual instructions in ASM to their machine-language equivalents. Nothing but a re-encoding (I'll deal with linkers momentarily). This is still human-readable, though relatively few people except chip designers ever take the required study to do it (reading machine code is hard though by no means impossible). In other words, simple translation and encoding. We still have a written work, unless you're going to apply a double-standard.
Now, a program is more than the object code alone; other things are added to it to put it into an executable format (this is why linkers are necessary). But this is all still human-readable, if you take the time to learn it. In other words, by the definitions set earlier in this post, a program, in source or object form, is still provably a written work, and written works have been legally defined as unpatentable.
There are other ways to prove that software cannot be legally patented, such as proving the fact that all software is, at heart, a mathematical equation (these are already defined as unpatentable). I've just taken a more unusual route, and one which negates the argument that programmers' works need to be protected by still allowing for that protection, but only for the actual work (by means of copyright).
Comments? Corrections for any of the translations above?
----------
Herbie J.
It is, indeed, fundamentally illogical to suggest that one should be able to patent hardware but not software that does the same thing.
The problem in the US is not so much software patents *per se*, but how badly the US Patent Office has gone about issuing and regulating them.
I have no problem with people being able to patent something truly innovative, whether the product of lots of hard work or a sudden flash of inspiration.
One should not be able just add '... on the Internet' to existing ideas and patent them, for example.
One thing I will take issue with: your bringing up the millions of dollars companies spend. An idea is not and should not be patentable simply because you've spent millions on it. In fact, patents are MORE defensible for the little guy, the backyard inventor, the small startup with a good idea.
If you spent a million dollars and came up with a poor, lame-ass idea that's not all that original, you don't deserve a patent for it.
The philosopher Robert Nozick made a fascinating argument on patents in his book "Anarchy, State and Utopia"
The gist is that a patent is perfectly just as long as the idea is sufficiently non-obvious that nobody else would think of it for the duration of the patent.
1 click shopping wouldn't take 20 years to come up with, but vulcanized rubber is another matter entirely.
I think that 5 year patents for software are perfectly reasonable, it's the 20 year deal that makes it absurd. The industry just moves too fast.
What do you guys think?
--Shoeboy
The board of directors of the European Patent Office voted to allow the unrestricted patenting of software. The authority recommends the cancellation of the current clause of the European patent convention which states that computer programs "as such" are not patentable.
In doing so, the EPA placed itself in the same boat as the proponents of software patents -- mostly international corporations that want a change in the European legal situation towards one more similar to the United States and Japan, where where software is practically patentable without restriction. The German delegation in the 19-member EPA committee voted against the raid in 10-to-9 decision.
Programmers of free software as well as smaller software houses stand against software patents, since the potential danger of patent suits from larger corporations makes their work practically impossible. Already today large American companies protect themselves with numerous patents on minor developments - a move which also allows them to resist by counter suits if another enterprise files a patent infringement suit against them. However, free programmers and small companies do not have the financial means in order to be able to exact such a strategy and therefore formed an alliance against software patents.
The endorsement of software patents by the EPA modifies nothing in the existing legal situation, but serves as a preliminary decision for a conference in November, where the nations that have taken part in the EPA want to finally rule on the batter. At the beginning of of July, leaks from the European Union indicated that the organization did not want to permit American-style software patents in Europe. However, organizations, like the EuroLinux alliance and the Linux federation LIVE, pointed out that the commission had, at the time, already ruled itself as being basically unopposed to software patents.
A couple years ago, Bruce Perens penned Preparing for the Intellectual-property Offensive for LinuxWorld. It's an interesting perspective on the potential for the subversion of the patent system by unscrupulous (is there any other kind?) proprietary software vendors.
Some noteworthy ideas, including that of "open patent" development, which keeps resurfacing whenever patents are discussed, but doesn't really seem to have taken hold yet.
--
Violence is necessary, it is as American as cherry pie.
H. Rap Brown