http://www.osnews.com/story/7442/Dutch_Parliament_Considers_Revoking_Support_for_Patent_Directive Exploring the Future of Computing en-us Copyright 2001-2009, David Adams adam+nospam@osnews.com Mon, 30 Nov 2009 21:12:49 GMT http://www.osnews.com/images/osnews.gif http://www.osnews.com http://osnews.com/thread? http://osnews.com/thread? Is the first petition I sign,and I see some results! Mon, 21 Jun 2004 21:47:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? If you can win the fight over there, then maybe we can win it in the US too. Keep hope alive! Mon, 21 Jun 2004 22:57:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Excellent work! Is there any protest planned for the date of the next debate? Mon, 21 Jun 2004 23:02:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Sooo... Lets say that over-broad software patents take hold, what is to stop the EFF or OSS folks from patenting everything under the sun and forcing businesses to prove how bad the patents are? Or just opening them up to everyone? Mon, 21 Jun 2004 23:42:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The problem is that the big companies like Microsoft already have the papers filed for consideration. They are only awaiting the "GO" signal to patent everything they can get their hands on. It's too late for a move like you suggest. Mon, 21 Jun 2004 23:52:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Glad someone has their euro-boots on. The patient (EU) may not emerge from the anaesthetic without a good diplomatic kicking. Tue, 22 Jun 2004 00:02:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I guess cost is the main reason OSS folks don't patent their ideas.

If you really, really know your stuff, the patent filing cost alone is around $350. Chances are that it will be rejected for some detail.

If you don't have years of experience filing patents, a patent agent's time will be around $3000-$20000. Tue, 22 Jun 2004 00:42:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I really hope the democratic powers of the EU can win over these greedy bastards, they make a disgrace of the human race Tue, 22 Jun 2004 00:49:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? As with everything to do with software patents, the issues are all in the details. To many people, it's just fuzzy abacadabra (for me too mostly, and clearly for some of our politicians).
The author does a good job at dissecting what's what, especially if you read the linked documents.
I think opening up patents for software as such is generally a bad idea, and even more for European business.

@DigitalStatic: the problem with that is in the MONEY: a) it takes lots of money to obtain patents, and b) it could take way more to enforce those, once you got them.
Big businesses have that money, but small/medium businesses, Free/Open Source people, or the hobby programmer, don't. And that creates an unfair advantage for big companies. Then it's not the best products, but the biggest fish in the pond that win. Tue, 22 Jun 2004 01:09:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I don't agree that, if a company spends years and money trying to find a solution to a software problem, they shouldn't be granted a patent for exclusive use of that problem for a set period of time -- just as a pharmaceutical company has exclusive use of its chemical formula for a set period of time, or a mechanical company has exclusive use of its mechanical design for a set period of time. This seems only fair and just; otherwise innovation would slow to a crawl. A decent system of software patents would have the same beneficial effects that the patent system has generally had since its inception. Software strikes me as inherently mechanical: it's not merely an idea; it's the implementation of an idea, and innovative implementations deserve a measure of protection. In this respect, I think the author's argument comparing software with the 15th century seizure of "commons" is flawed.

I do agree that the current system of software patents as implemented in the US, and the system that seems to be pending in the European Union, is a legalistic mirage that will have the opposite effect, giving "a letter of the law" that will completely evade its spirit, much as current copyright law seems to be on the verge of doing. In this sense, such patent law would certainly stifle innovation.

I just hope we won't throw the baby out with the bath water. Tue, 22 Jun 2004 02:40:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? No, software patents do not only cover implementations. Your specific implementation is always protected by copyright law, whether there are software patents or not. We really don't need software patents for that. Tue, 22 Jun 2004 05:20:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Very nice article. And interesting information about Microsoft. Now it seems proven that they don't play fair at all. I think we all should boycott M$ products, also in our companies. Why use Word when you can use OpenOffice or AbiWord or LaTex! Dont buy an XBox or buy it and dont buy any games, just run Linux on it... Man, this all makes me angry. Tue, 22 Jun 2004 06:09:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? So Microsoft spend many years to invent the click and double click on PDA?

You can't take a pharmaceutical example and apply it over informatic, that doesn't work, they are different Tue, 22 Jun 2004 06:11:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Protests don't cut it as it is...

Software Patents is an area that appeals to very little people and thrus a protest will never get much adderance (even if it gets 100% of the people involved there).

In the end it will always be viewed as a "failure" by the press because of so little people there.

That is bad, because problems like this shouldn't be viewed by the "adderance" of the masses, one people is one too many! Tue, 22 Jun 2004 06:19:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "So Microsoft spend many years to invent the click and double click on PDA?"

Well, that's not entirely true. They are talking about a device with "limited resources".

So, unless your pc has unlimited resources, it would also apply to normal pc's. Tue, 22 Jun 2004 07:35:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "I don't agree that, if a company spends years and money trying to find a solution to a software problem, they shouldn't be granted a patent for exclusive use of that problem for a set period of time"

When was the last time you met a company that spent years and money trying to find a solution to a software problem?

In my experience, that is just theoretical. So I'd like to see a real example for a change proving me wrong. And even if you manage to find one, it's just one exeption on millions of programmers that do depend on re-using existing ideas to earn their dayly bread.

The point is that yes, writing software costs money, but the money goes into debugging and fine tuning. 99.99% of software is just a re-use of existing ideas. You couldn't do anything if you couldn't re-use great ideas of people that did realise they were also standing on the shoulders of others.

I like to compare it to architecture (my father in law is an architect). They do nothing but re-use existing ideas to design buildings. Often, beautiful designs, but no inventions.

It's because they use existing ideas in different combinations that they make completely different products.

So, their money and work is in the artistic part, not the incentive part. To me, this also goes for software.

Besides, software patents are not allowed by international treaties. The TRIPs treaty forbids patentability of software. On the website of the WTO, article 10.0 of TRIPs says:

http://www.wto.org/english/tratop_e/trips_e/intel2_e.htm :

'The obligation to protect computer programs as literary works means e.g. that only those limitations that are applicable to literary works may be applied to computer programs.'


There is no patentability for literary works, so there shouldn't be any for computer programs either.


On top of that, the WIPO copyright treaty says in article
10:

http://www.wipo.int/documents/en/diplconf/distrib/94dc.htm :

'(1) Contracting Parties may, in their national legislation, provide for limitations of or exceptions to the rights granted to authors of literary and artistic works under this Treaty in certain special cases that do not conflict with a normal exploitation of the work and do not unreasonably prejudice the legitimate of the author.'

The normal exploitation of a computer program means using and selling it. Both can be banned by software patents if a third pary owns a software patent a program breaches.

Therefore, software patents are in conflict with both the TRIPS and the WIPO copyright treaties.

So, why have these treaties if you don't live to them? Tue, 22 Jun 2004 08:12:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Except that the sentance > is proven to be wrong.
If you just don't trust the fact that computerscience made a hell lot of innovation before those patents came, the FFII has an impressive list of studies (inclusive many officials one) to assert this fact :

Since Fritz Machlups report to the US congress of 1958, a considerable number of studies about the economic effects of the patent system has accumulated. Some studies deal with certain types of innovation (sequential, complex systems) or with special areas such as semiconductors, genetics or computing rules (algorithms, mathematics). None seems to claim that the patent system has a positive effect on innovation in these fields. Most find strong indications for negative effects. Some governmental studies (e.g. by intellectual property institutes and the like) combine such negative findings with a recommendation to legalise software patents.

followed by an impressive list.
http://swpat.ffii.org/archive/mirror/impact/index.en.html Tue, 22 Jun 2004 08:28:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I'm finally getting to know what's really happening with software patents in Europe.

I hope we can do something like this in Spain too. Tue, 22 Jun 2004 08:45:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? als ik tijd heb kom ik donderdagavond nog even kijken naar het debat. even zien: een microsoft t-shirt met een linux button erop? Tue, 22 Jun 2004 08:56:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "I hope we can do something like this in Spain too."

Actually, Spanish professors have played an important role.
Make sure to take a look at the Fajardo Lopez report:

http://www.vrijschrift.nl/Members/arend/Fajardo-Lopez-Law-Office_Pr...

This is why the Secretary of State had to admit the parliment *could* revoke their vote!

Kuddos to our Spanish friends! Tue, 22 Jun 2004 09:03:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The costs of patents are another point. In Europe patents are very expensive, because they have to be translated in all official languages spoken in the EU.

That's one of the reasons patents costs tenths of thousands of Euros here. And that's also why it is rubbish that they come up with the argument that it is so good for the little companies. They can't afford it.

Actually, 75% of Europe's controversial software patents are in the hands of American and Japanese multinationals. And still, they keep up that this is going to be good for the European industry. Tue, 22 Jun 2004 09:13:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Well, I'm going to wear something special, since we are planning a not-so-common protest on the square before the Parliament, thursday at 19:00.

I'll be whistling (is this the correct word) an Oxen Horn with the idea of representing our forefathers who also saw their common possessions being appropriated:

http://www.vrijschrift.nl/Members/arend/pics/ossenhoornblazers

We normally do this around Christmas accoding to the tradition and we always get mentioned in the press, but the Oxen Horn was also traditionally used by the farmer leaders to call upon the community to work on the commons.

Hopefully this will also get us some press this time.

But anyway, otherwise I would be wearing a suite. After all, we are important guys for the software industry! Tue, 22 Jun 2004 09:29:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I am today a proud SouthAfrican dutch descendant speaking Afrikaans-actually went to amsterdam a couple of weeks ago and could speak my native toungue there and understand what they said-*MUCH* better than here in BERLIN!!! and the folks are as friendla as back home!!! just a few days more and I will be home, yipeeeeee!!! Tue, 22 Jun 2004 09:51:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? So Microsoft spend many years to invent the click and double click on PDA?

Actually, this is precisely the sort of thing I don't like. Patentability would be more like -- a company that devises a genuinely new thing, not a reapplication of old art on a new device. I mean that the invention of new art should be protected; I'm talking about that as the implementation, not about the code itself (which as someone pointed out is adequately covered by copyright law).

Some people have said, well, precisely when has a company devised a genuinely new thing? That's a fair argument; how about the RSA algorith, Be's filesystem, or Apple's Exposé? As I understand it, all of those were innovative (I don't buy the Exposé = window tiling argument at all, having seen both). I don't think that other companies should have been able simply to re-implement RSA's algorithm after their effort; or Be's filesystem on their OS; RSA and Be did a lot of work and ought to have some exclusive, temporary rights. Likewise, if a company devised a new, fantastic [search | sort | garbage collection | resource allocation | etc] algorithm, I think they ought to have some rights to that. I don't accept the notion that software is only an idea, and therefore like journal articles in the sciences and mathematics, ought to be privy to the "commons" approach (which by the way is not a 100% accurate description of the process, as the RSA patent shows, among others).

Someone else said that a lot more innovation was done in the early days than in the last few years. Hmm, perhaps. But there are lots of ways to measure innovation, and I'm sure someone can make the argument that there was more innovation in photo camera technology in its earlier days than there was in the 20th century, until the invetion of digital camera technology. That doesn't mean that patents are evil; it may simply mean that, after the initial flurry of work in any new field, it's much harder to invent something new. That has nothing to do with patents, but with the human mind.

It ought to be clear that I'm not entirely certain what the theory is behind patents -- I may be arguing things based on incorrect understanding. But, I am convinced that a sweeping generalization like "Software Patents are inherently and morally evil" needs more discussion. Just because the US Patent Office shows no interest in doing it right, and the EU Commission doesn't seem to be doing much better, doesn't mean that they ought to be discarded willy-nilly. Tue, 22 Jun 2004 09:55:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Keep up the good work! For once, I'm proud to be Dutch. :-) Tue, 22 Jun 2004 10:02:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? to all involved. This is real politics, for a worth while cause.

Questions do need to be asked about the financial influence of Microsoft in Europe. Tue, 22 Jun 2004 10:14:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I understand what you mean. The problem is that it is extremeny difficult to define what's new and what should deserve protection.
The daily practice is that the EPO, which is supposed to be an honest judge, gets payed for every patent it allows. So, they are not saying "is this really new?". They are saying "can we get away with allowing this patent?". If yes, $$$.

But even if the EPO would be able to judge adequately, I am not in favor of software patents.

This is the principal discussion about whose interest comes first:

a) the person that happens to have a new idea first and should therefore deserve a monopoly on that idea.

b) the interest society has in maximising innovation, because that spurs economic activity to the benifit of all.

Since I think the interest of society should come first, I think it is important the patent system does a good job in *maximising innovation*.

Now it seems logical that innovation is stimulated by patents, but that is not the case. There are several studies that show that in fact research en development gets lower because companies invest in patents instead of R&D.

See Kahin http://www.si.umich.edu/~kahin/mip.html:

"A recent study by James Bessen and Robert Hunt shows that patents actually substitute for R&D in the software."

So, software patents have a *negative* effect on innovation and are therefore bad for society. Tue, 22 Jun 2004 10:14:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? This protest you are talking about, will it be this thursday? If so, I'll be there! Tue, 22 Jun 2004 10:23:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Yes. The debate will be at 19:45, so we're having a little action at 19:00.
It's not a big demonstration or something, just a little playful action that will hopefully attrack some press.

Should you want to help, consider joining the nl-parl mailinglist at: http://lists.ffii.org/mailman/listinfo/nl-parl

There are also lists for other countries. They are mentioned on:
http://swpat.ffii.org/group/todo/index.en.html Tue, 22 Jun 2004 10:33:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "This is the principal discussion about whose interest comes first:
a) the person that happens to have a new idea first and should therefore deserve a monopoly on that idea.
b) the interest society has in maximising innovation, because that spurs economic activity to the benifit of all.
Since I think the interest of society should come first..."

No! No, No and No! I think that is completely wrong, at least according to my values. Individuals rights come first. The opposite is the basis of communism (i.e. State or society before individuals). Society isn't a bee hive, people don't have to contribute for the "common good" if they don't want to. Everyone has the right to be greedy.
It would be in society interest to forbid carmen electra from wearing clothes, that doesn't mean they (society) have the right to do so
Also, you cannot apply that reasoning to software pattents IMHO because it's someones individual rights vs. another persons individual rights. Tue, 22 Jun 2004 11:09:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I think you just named a few good examples where a patent would be applicable.

The examples you named are bound to a specific implementation. As far as I know 'normal' patents are a pretty detailed and technical description of the innovation and someone should be able, when the patent expires, to reconstruct your 'thing' by reading the patent.
A patent should also be something that gives you the feeling "I would've never thought of", or even "Why the heck didn't I think of that first" and it should be bound to a specific implentation of that idea.

A patent is give the inventor the time to recoup costs and earn a few bucks and then the idea is in the public domain.
So in the world of rapidly evolving software 7 years or so is much too long. If the patent was really innovative and quite specific I think that it still should be patented for about 2 or 3 years maximum.

If you look at mechanical engineering then there are ways to work around an invention. Your product might not be exactly the same, but you could create a new design to get the same function.
So a patent should also leave space for a competing and different implementation which results in similar functionality.

The thing that bothers me most is the way the European Commission ignored the European Parliament and the impact of software patents on Open Source software. I think patents should *improve* innovation instead of a means to stiffle it.
I will try to be at the Tweede Kamer this Thursday (in suit and tie :-) ). Where can I find more info? Tue, 22 Jun 2004 11:37:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "Individuals rights come first."

Hmmm. Eh, you are absolutely right about that. Let me try to put it another way.

Historically, patents are ment to stimulate innovation by rewarding people that invest in R&D with a temporary monopoly on their invention.
This seems to originate from the thought that society would benifit from extra innovation.

Now, they are talking about Intellectual Property, as if someone should have the right to own an idea.

Therefore, the fundamental question is: Should someone be able to own an idea, just because he was first to think of it?

If you can't answer that question with "Yes", the only reason left that could justify software patents is that society as a whole would benifit from the extra innovation, which is actually not the case (see previous reaction).

Since my answer is "No, I don't think anyone should have a fundamental right to own an idea" and society also has no benifit, I am against software patents.

If your answer is "Yes", then that's fine. That would mean you would not fundamentally reject software patents, which is just another opinion.

However, you would probably want a reasonable system to judge patents and for a price anyone could afford. With the current directive, this is not the case. That is why there is such a broad resistance against the current directive. Tue, 22 Jun 2004 11:37:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Thanks for bringing this into the public with such a well-written article!

Problem with the EU is this only helps the overal outcome slightly because every nation has a vote while the influence of the Netherlands is _incredible_ small. With this power centralized, we need to organize more over the borders as well and i think the Internet is a perfect platform for this. It happened already as well.

The Netherlands loses power in such way that even Dutch politicians like members of VVD feel uncomfortable with it they'd rather have only the economic benefits, but as we all understand that doesn't work very well if things that relies on are missing.

Any news on the new countries who joined the EU? What's their point of view? Tue, 22 Jun 2004 11:40:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Dieter promised to keep http://softwarepatenten.be/ up to date. Tue, 22 Jun 2004 11:43:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Good points! A few I didn't think of.

On the one hand I don't think someone should be able to own an idea.

OOTH if you have a great idea for a new rear axle or something, how would you prevent other people from copying you axle. They wouldn't have the same investment in research that you would have, so you'll be the one to end up broke. That also stiffles innovation.

I think the biggest danger with software patents is that they will be overbroad, like they already are in the US.
Another problem is that big companies, with more resources, will be able to patent more, so not a lot will be left for small companies. Not a good thing.

Again another problem is that a lot of software patents are already owned by non-European companies. As Europeans we would be stupid to allow those companies to build up a monopoly and kill off our own economy. Tue, 22 Jun 2004 11:45:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Commissievergaderingen:
Maandag, dinsdag, woensdag, donderdag
U kunt u melden bij de balie aan de ingang Plein 2 Tue, 22 Jun 2004 11:49:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? One of the problems I see into people reacting to this is lack of knowledge of both its technical and its political aspects. For instance, first question: how the heck does one know what his or her country voted? Tue, 22 Jun 2004 12:01:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? If you have an idea about a mechanical innovation, then you patent the PROJECT of a very physical thing.

Your patent will be a classical industrial patent, and of course it is already perfectly legal in the EU.

Let's say you invent an innovative steering system for cars that solves the problem of understeering and patent it.

If your competitor then devises a DIFFERENT steering mechanism that solves the same problem as yours, then his project is prefectly legal, and does not infringes your patent.

Software Patents are different: it is more like patenting the IDEA of removing understeering.

Your competitors won't be able to fix the problem with ANY different steering system project, because you own the very concept of removing understeering.

Sound silly?
That's exactly my point.

IT companies have always failed to show examples of GOOD and USEFUL software/algorithm patents.

BTW, in the USA you could easily patent the idea of the turing machine, and by that basically own all the computing industry. Sure, there will be prior art and you will eventually lose, but how many millions of EUR will go down the drain (and the lawyers pockets) before that? Is this promoting innovation in any way?


(sorry if some parts maybe make little sense, but english is not my native language)

Bye, Renato Tue, 22 Jun 2004 12:10:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Reading the article I spotted the following google ad at the top of the screen:
---
Have a new product Idea?
Protect your great new idea, Don't let someone else patent your idea.
---
How weird is that !

Anyway, will more people present at this debate coming thursday make a difference ? In that case I'm thinking of going as well. Tue, 22 Jun 2004 12:16:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The point is, the vote actually was only just met, because it needed a more then 50% majority. I don't know the exact details about how big the majority needed to be, but it comes down to this:

The vote of Poland seems to have been counted wrong. It seems to be that there is a rule that if you don't say anything, it is assumed you are in favor. Apparantly, Poland didn't know this and has been counted "in favor" while this is not what they wanted.

The Netherlands has also voted in favor, because the Parliament was misinformed.

Now, it is *not done* to revoke you statement once a "political agreement" has been made. But, it is legally possible. And of course, it is very well possible that due to the maternity leave of the Secretary of State some things went wrong which caused the confusion.

So, it is a realistic possibility (I'd say 1 to 3) that the Netherlands will indeed revoke their support in the formal meeting that will take place in September. This would be a very strong political signal, since this has never happened before.

*If* the decsison to do so is taken coming Thursday, it might very well be that Poland will also join the Netherlands and that will be enough. It might also be that for example the opposition in Germany will take their chances and make trouble in the Bundestag. Who knows?

Another possibility might be that The Netherlands ask the Presidency for a new voting procedure. If they do that, the Council will have to vote about wether or not to allow a new voting procedure. Since a respected member state apparantly thinks there is enough reason to open a new voting procedure, they might just be able to get a majority for a new voting procedure with the accompanying new negotiation round.

If you're really interested in how everybody voted, just look up the transcriptions of the actual meeting, mentioned in the article. There's even an audio recording. Tue, 22 Jun 2004 12:24:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Thanks for the explanation, Renato.

I think you're right and we agree. It's what I call 'overbroad patents'.
However, I don't agree with your statement "IT companies have always failed to show examples of GOOD and USEFUL software/algorithm patents"

Yes in the case of overbroad patents, but no in the case of very specific algorithms as Lempel-Ziff (sp?) encoding (zip) and the RSA algorithm. I don't think they're algorithms that anyone could think of, are obvious or are too general.
So in my opinion these algorithms are patentable. They are precise, specific, not-obvious and represent an algorithm more than an idea.

Patenting One-Click paying is obviously bizarre, so shouldn't be possible here.

Also I think it should be obligatory for standards bodies to enforce that proposed standards are not encumbered by patents. How can something be a common standard if it forces everyone to pay a single party.
So either you patent something and you won't be able to propose it as a standard, or you make it a standard, but lose the patent.
A lot of lawsuits in the US are because of standards that have patents attached. Tue, 22 Jun 2004 13:59:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The algorithms you talk about are basically mathematics. They use prime-numbers in specific ways. The ideas are in the mathematics, not in the implementation in software.

Since mathematics are not considered as a field of technology, these algorithms should not be patentable. Or at least that is what is generally accepted.

Because if they were, then you would also need to consider allowing patents on all mathematics. Why only allow patents on mathematics that happens to be useful in software?

The point is that with the example you show, you would allow a patent on the _use_ of an idea that actually lies in a non-patentable field: mathematics.

That is one of the dangers of sofware patents: they are a backdoor to allow patents on ideas that should not be patentable because they are not technical.

The same reasoning goes for business methods. These should not be patentable, too. But, if you're a smart patent lawyer, you can wrap it into a piece of software so to speak and get a patent anyway. Tue, 22 Jun 2004 14:21:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The problem is that it is extremeny difficult to define what's new and what should deserve protection.

WIth this, I agree :-) But to my mind, that only means we need to work hard to determine what it should be. Let me state: I think your effort is a great thing, because our governments are conceding too much too quickly to the software companies. Someone said software patents should only last 2 or 3 years; that alone is a good start. I simply don't agree with the notion that software patents shouldn't exist at all, and I think it's a fundamental disagreement over the notion of software. You have said several times that software is an idea, and no one can own an idea, so patents are philosophically wrong. I agree that no one can own a general idea, such as "encrypting messages" -- but the RSA is not a general idea; it's not something someone was just "lucky" enough to "think of before anyone else". There was a lot of hard work that went into it, and that hard work ought to be rewarded, not simply placed immediately into the public domain so any hack programmer can write their own code to implement the program.

The mathematical ideas that underlie the algorithm are not patented (for all sorts of reasons, not the least of which being that exponentionation of integers goes back centuries). Nor did anything stop someone from implementing a similar algorithm for encryption of messages, using exactly the same mathematical ideas -- they simply couldn't implement the RSA, without obtaining a license.

This is vastly different IMHO from (say) Amazon's attempt to patent "one-click shopping". No real effort went into that, and it was quite possibly prior art (I don't know the details).

Now it seems logical that innovation is stimulated by patents, but that is not the case. There are several studies that show that in fact research en development gets lower because companies invest in patents instead of R&D.

I looked up the web page, but got a 404 error. Can you tell me whether the study is of general R&D or of software R&D? The difference is crucial to the argument I think, because I believe firmly that, in general R&D, patents have been beneficial, whereas I agree that with software R&D patents have been harmful. But, to repeat myself, I think that's because software patents are being implemented irresponsibly by the US Patent Office, not because software patents are inherently flawed. Tue, 22 Jun 2004 14:28:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Originally you weren't allowed to patent ideas. I'm sorry, but software should not be patentable. Its written, not invented. There fore it should be covered by copyright law not patents. Second, patents should have to be filed BEFORE wide spread use happens. The people who owned LZ compression knew it was being used for years, or at least should have. Permitting it to happen is the same as giving consent as far as I am concerned. Once you allow something to start being used by the public, through either neglect or good will, it should become public domain. But thats just how I see it.

TG Tue, 22 Jun 2004 14:28:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The examples you named are bound to a specific implementation. As far as I know 'normal' patents are a pretty detailed and technical description of the innovation and someone should be able, when the patent expires, to reconstruct your 'thing' by reading the patent.
A patent should also be something that gives you the feeling "I would've never thought of", or even "Why the heck didn't I think of that first" and it should be bound to a specific implentation of that idea. ...So in the world of rapidly evolving software 7 years or so is much too long. If the patent was really innovative and quite specific I think that it still should be patented for about 2 or 3 years maximum.

I agree with you mostly, except that I think patents ought be on something that give you the feeling: "Wow. Now, that is brilliant." So for example I think Xerox had every right to patent the GUI -- and, as you say, such a patent should have expired quickly. I don't think (based on what little I know of the case) that the patent on XORing bits to move a mouse should have been granted. Tue, 22 Jun 2004 14:31:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Therefore, the fundamental question is: Should someone be able to own an idea, just because he was first to think of it?

I'm going to repeat this so many times that everyone's going to hate me: this is not the fundamental question. Software is not merely "an idea", and a programmer is not simply the "first to think of it". There is a mechanical design process involved.

I am a mathematician. We write proofs for a living; we "create knowledge" (in the words of one of my advisors). Theorems are ideas. Likewise, the ideas discovered by science (strings, probability electron clouds, the Big Bang) are pure knowledge. To patent these ideas is not merely "unethical" (to use the words of some) it is quite frankly impossible. Someone can use the theorems I create, without understanding diddly-squat about the proofs that underlie it -- and so long as we mathematicians and scientists get credit for our knowledge, we aren't terribly offended, because for us the reward is more in the knowledge itself, and in the credit for the discovery. I suppose one could argue that the proof is mechanical design -- but only a fool would copy a proof, because any proof proves only one specific theorem, and there's no sense in trying to reprove a theorem with the same proof.

This is not at all the same as algorithms, that could possibly be patented in computer science. Algorithms require design and implementation; to re-implement them requires copying the mechanism itself (without necessarily copying the code -- I can copy the algorithm using a clean code implementation), and it is therefore easy to profit off another's labor. This is much closer to mechanical engineering or chemical engineering, than to mathematics proper, and that is inherently unfair to the software engineers themselves.

In sum: you can't patent the idea of encryption of email -- but you can patent a particular encryption method. So far I haven't seen any argument that really refutes this. Tue, 22 Jun 2004 14:44:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? http://www.si.umich.edu/~kahin/mip.html

The study is about software R&D. See http://swpat.ffii.org/papers/bessenhunt03/index.en.html :

"Bessen & Hunt show that strategic, anticompetitive and defensive use of patents tends to concentrate in software patents, because they are easier to obtain (they don't require experimentation or prototyping, not even writing a program). They are also broader, because software is not subject to physical constraints and can therefore be composed into more complex systems, potentially infringing on hundreds of patents per program. This causes a patent buildup simlar to a cold war arms race that discourages innovation and competition, and instead of bringing new products to consumers, reduces their choice and their access to infomation society, resulting in significant costs and less productivity for businesses."


"There was a lot of hard work that went into it, and that hard work ought to be rewarded, not simply placed immediately into the public domain so any hack programmer can write their own code to implement the program."

Not having a patent is not the same as relasing into the public domain. Copyright allows you to keep the source code secret and only release binaries, which are very difficult to work with. It is a lot of work to "decompile" a binary as our Secretary of State likes to call it and paints it as if hackers are doing that as an afternoon exercise.

Actually patents *should* allow to to write your own code to implement the program. That's what it's all about: because if you get a patent, you *publish* your idea, so that others can *build* further on it. That's how innovation is supposed to be futhrered and it is assumed that if you can make a derived product, you can also afford to license the original idea.

With sotware patents, the practice is that there is no source code or otherwise a description of the invention that any sane programmer can read. It is ususal just lawyer rubbish that noone but a lawyer can understand. That undermines the whole principle.

It is not for nothing that Dutch pro-patent organisations wanted a reference source code implementation to be included with a software patent application. By the way, a wish not-granted. Tue, 22 Jun 2004 14:48:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Jack Perry wrote; " ...how about the RSA algorith..."

Thank you for providing a wonderful example of why software patents are NOT a good idea.

Back in the day there were a few simple rules regarding patentability that worked well for us in the U.S. of A.

Had to be a physical object.
Had to be non-obvious to someone in the field.
Had to be truly inventive.
Could not be a "discovery" (i.e. God had invented it first, that ruled out people, animals, and plants).

Math, science, genetics, lifeforms, software could not ever be patented under these rules, nor could "business methods".

The RSA algorithm isn't an invention, it's a discovery. It's a previously unknown fact of math. If you believe that algorithm should be patentable, why not the Pythagorean Theorem, or a method of solving quadratic equations? The fact that an invention, the programmable computer, can execute an idea doesn't make it any more patentable than doing that same thing in your head.

Just look to the U.S. of A. to see just how ridiculous the world becomes when you allow "anything" to be patented. Plants are patented in the U.S. Gene sequences are patentable. You can't keep seeds from one year to the next, as farmers have been doing since the beginning of farming, if you happen to be growing genetically modified soybeans. If you do, you are infringing on a "gene" patent. If your field gets contaminated with genetically modified corn, pay up or burn your crops to the ground. You are infringing on someones patent. Some day the gene for blue eyes will be patentable and everyone with blue eyes will owe some multinational money for "infringing" on their patent by being alive.

Business methods patents, like the "one-click" patent, or the "to-do list" patent. How about the patent for downloading software? Yep. While we are at it why don't we allow a patent on the exchange of goods or services for money. Too obvious you say, how about if we add "... with a computer, or through a computer network." That should definitely make it patentable.

"Delivering goods to customers" - not patentable

"Delivering goods to customers with a computer, or through a computer network." - we have a winner.

"Linking a customers billing and shipping records to a customer number to make new orders by the same customer more efficient" - not patentable.

"Linking a customers billing and shipping records to a customer number to make new orders by the same customer more efficient with a computer, or through a computer network." - not only is it patentable, but it's the basis of Amazon's "one-click" patent.

Therein lies the current EU problem. A mathematical equation isn't patentable in the EU, but if you add the magical, ".. with a computer" like magic you can patent any equation you want. The people in favor of software patentability say you can't patent software, just software that runs on a computer.

Hmm, isn't that a little like saying you can't patent breathing, only breathing done living entities. The difference will provide little solace as you hold your breath, or infringe on their patent.

I hope the Dutch, the Polish, and the rest of the EU wake up before it's too late.

Just my $0.02 (Canadian, before taxes)
someone247356 Tue, 22 Jun 2004 14:49:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Software Patents are different: it is more like patenting the IDEA of removing understeering.

Your competitors won't be able to fix the problem with ANY different steering system project, because you own the very concept of removing understeering.

That's simply wrong. No one was prevented from implementing a new algorithm to encrypt and decrypt messages, because of the RSA patent. In fact, several different methods exist, and some have been used that operate on "similar" mathematical principles: Diffie-Hellman, ElGamal, Elliptic Curve. Moreover, the strengths & weaknesses of RSA have been regularly discussed and analysed in the mathematical community. The idea was still free; the mechanism was not.

BTW, in the USA you could easily patent the idea of the turing machine...[i]

That's almost certainly wrong. For all the defects of the US patent office, they're not [i]that dumb. Tue, 22 Jun 2004 14:55:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "you can't patent the idea of encryption of email -- but you can patent a particular encryption method. So far I haven't seen any argument that really refutes this."

The point is that if you want to make this patentable, you should make the underlying algorithm (the idea :p) patentable, not the specific implementation in software. Tue, 22 Jun 2004 15:11:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The RSA algorithm isn't an invention, it's a discovery.

How is an invention different from a discovery? It's a subset: an invention is in fact a kind of discovery. So your argument is flawed from the beginning, but also in the middle.

[RSA] is a previously unknown fact of math.

No; it's not a fact; it's a method, based on some very well-known mathematical principles, as well as on the suspicion that factoring sufficiently-large integers is impractical. It may in fact be the case that factoring sufficiently-large integers can be done quickly, using a method we don't yet know -- and believe you me, when that is discovered, RSA will collapse like a house of cards in the wind. The genius of RSA was to assemble the known facts into a mechanism that solved a pressing problem.

If you believe that algorithm should be patentable, why not the Pythagorean Theorem, or a method of solving quadratic equations?

A theorem is not a kind of algorithm, nor vice-versa. In mathematical journals, algorithms are published alongside theorems that state, and prove, the algorithm's correctness.

As to solving polynomial equations, such methods were kept secret in Renaissance Italy, because Italian mathematicians derived prestige from being able to solve polynomial equations that other mathematicians could not. Today on the other hand, the prestige of a salaried researcher comes in accumulating published solutions to purely mathematical problems, because ever since Fermat's last theorem (and other gaffes like that by mathematicians) no one believes you if they can't check & verify the proof. -- And, as I argued before, only a fool would copy the proof of a theorem, once it was published, because it only proves that theorem, and there's no profit in it. With algorithms, however, there's a completely different context: there's no inherent reward to developing an algorithm, and you can demosntrate the algorithm's effectiveness quite effectively, without other people's being able to copy & redistribute it. Tue, 22 Jun 2004 15:13:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The point is that if you want to make this patentable, you should make the underlying algorithm (the idea :p) patentable, not the specific implementation in software.

Hunh? I thought you considered copyright (the specific implementation in software) sufficient, rendering patents (the underlying algorithm) superfluous. Tue, 22 Jun 2004 15:16:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "it's a method, based on some very well-known mathematical principles"

Ok, then let me refraise my previous comment:

The point is that if you want to make this patentable, you should make the underlying algorithm (the *method*) patentable, not the fact that it happens to be implemented in software.

If the method is technical, it should be patentable.
If it is not, it should not be patentable.

Exactly what the European Parliament said. And hopefully what we will get.

Oh, and "technical" basically means that in some way forces of nature have to be involved. Tue, 22 Jun 2004 15:19:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Thank you for the extracts. As far as I can tell, there's nothing you've quoted that contradicts my fundamental position (ie that it's the implementation of software patents that is wrong, not the idea itself) but mebbe I'm wrong :-)

Actually patents *should* allow to to write your own code to implement the program. That's what it's all about: because if you get a patent, you *publish* your idea, so that others can *build* further on it. That's how innovation is supposed to be futhrered and it is assumed that if you can make a derived product, you can also afford to license the original idea.

I still don't agree. To "build further on" the algorithm, is not at all the same as to distribute your own version of the code to the public. It means what it says: to build further on it, by modifying it, or implementing differently the ideas underlying it.

I don't think that a patent on a machine prevents a researcher from building his own model of the machine, instead of buying it from the company -- the patent simply means that he can't distribute his own production models. Patents give inventors the exclusive right to market their invention, and to profit from it. Am I wrong on this? Tue, 22 Jun 2004 15:24:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? All I'm saying is that what is patentable, the method, the idea or whatever you want to call it, should be patentable on it's own, regardless of wether or not it is implemented in software.

What happens now is that methods that are not patentable by themselves, suddenly become patentable just because they are executed by a computer. Tue, 22 Jun 2004 15:25:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The point is that if you want to make this patentable, you should make the underlying algorithm (the *method*) patentable, not the fact that it happens to be implemented in software.

I was going to say, Bingo! we agree -- as I said, I have deep reservations with the looseness of the US Patent Office -- but then you added your opinion that "technical" means that forces of nature must be involved, and I find myself disagreeing anew... After all, one could (if I understand you correctly) manufacture and patent an "RSA machine", just as companies manufacture and patent all sorts of other machines; well, the software that implements RSA, is, IMHO, just that: a machine.

So actually I'm perceiving weakness both in your argument and in mine, but I can't quite put my finger on it. Tue, 22 Jun 2004 15:29:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? You wrote, "No one was prevented from implementing a new algorithm to encrypt and decrypt messages, because of the RSA patent."

Let me see if I am understanding you correctly. You claim it is O.K. to patent an algorithm as long as you can get the same answer some other way.

"In fact, several different methods exist, and some have been used that operate on "similar" mathematical principles: Diffie-Hellman, ElGamal, Elliptic Curve."

So you would have no problem if I patented adding two numbers to arrive at a sum. Because as you have written several different methods exist and some have been used that operate on 'similar' mathematical principals."

2 + 2 = 4 Patented
1 + 1 + 2 = 4 Ok.
8 / 2 = 4 Ok.
2 * 2 = 4 Ok.
(2 * 4) / 2 = 4 Ok

8 - 4 = 4 Might be ok, could be argued that it's really 8 + -4 = 4 infringes on my patent.

Yep, you can still get to the number 4, I guess you'll be supporting my patent application on the addition of two number to arrive at a sum.

Thanks for your support.

Just my $0.02 (Canadian, before taxes)
someone247356 Tue, 22 Jun 2004 15:30:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The patent issue is bad enough, but worse is the democratic process. This process is something that people on the street do have a chance of understanding - if only more people over here would take notice of whats happening in the EU!

Could this be presented to the public as a case of the European Parliament doing something thats good for us, then getting overruled by a quango somewhere that is acting in foriegn interests? Tue, 22 Jun 2004 15:31:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? What happens now is that methods that are not patentable by themselves, suddenly become patentable just because they are executed by a computer.

With this I understand you, as well as your point (quite salient) that the method ought to patentable in its own right. That gives me pause: is that at all the theory behind how patent law works? If patent law works differently -- since if it works differently, there's a terrible weakness in my understanding & thus in my argument. What I wouldn't give to have a competent patent lawyer to talk to... Tue, 22 Jun 2004 15:32:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? So you would have no problem if I patented adding two numbers to arrive at a sum.

No, for several reasons. To begin with, there's prior art. Secondly, you haven't patented a method; you've patented an idea. If you were to invent (say) the method of adding tally marks, then I'd say that could be patented -- or if you were, in the 19th century, to invent a mechanical cash register to add numbers quickly for stores. In fact, such patents were granted then.

http://inventors.about.com/library/inventors/blcash_register.htm Tue, 22 Jun 2004 15:37:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? I wrote:

No, for several reasons.

I mean, "No, I would have a problem, for several reasons." That would be obvious from what followed, I hope. Tue, 22 Jun 2004 15:38:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Prepare to claim your free PC with every software title you buy...! Tue, 22 Jun 2004 16:05:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? You wrote;
"A theorem is not a kind of algorithm, nor vice-versa. In mathematical journals, algorithms are published alongside theorems that state, and prove, the algorithm's correctness. "

So the theorem stated as "The square of the hypotenuse of a right triangle is equal to the sum of the square of the other two sides." isn't patentable, but an algorithm for determining the length of a hypotenuse of a right triangle should be?

y = SQRT( x^2 + z^2)

where y = length of right triangle
and x = one side of a right triangle, not the hypotenuse
and z = another side of a right triangle, not x or y

Perhaps I am a little slow this morning, but practically speaking how's one different from the other?

You wrote, "As to solving polynomial equations, such methods were kept secret in Renaissance Italy..."

Great, I'm all for trade secret protection for algorithms. You want to keep it to yourself, go for it. There's no impediment to me "discovering" it independently and using it. If it's patented, then even if I come across it while scratching lines in the sand on the deserted beach where I've spent my entire life, I still can't legally use it. Of course, the rest of the planet will pass you by as they share, test improve on the body of mathematical knowledge that's in the public domain. In fact I believe that's just the situation that the NSA is in. When someone there discovers a wonderful new algorithm to encrypt a message, or to break an otherwise encrypted message, they keep it to themselves. I remember reading something about a "secret-service" that had discovered the idea of public key encryption, and some of the algorithms that were later discovered in the civilian sector and patented (uggg..) they had "prior art" but couldn't reveal that they already knew about it.

You will notice that I've used the term discovery as opposed to invention when talking about algorithms. You asked, "How is an invention different from a discovery?" You then proceed to answer yourself ("It's a subset: an invention is in fact a kind of discovery.") and based on your answer declare my argument flawed.

I will try to illucidate. A discovery is the process of gaining knowledge that exists independently of the discoverer. An invention is using a discovery, usually several, to create a device that didn't exist before, and wouldn't have been obvious to someone working in the field before you invented it. This is the standard definition of the difference between discoveries and inventions. Notice the emphasis on "device", I'll come back to it in a minute.

A mechanical calculator is a device to automate the computations of mathematical operations. You can patent a particular calculator design, you can't (or shouldn't) be able to patent the mathematical algorithms themselves.

You can patent a design for a steering wheel, you shouldn't be able to patent the idea of using a round device to steer a vehicle.

You can patent a particular physical computer, you shouldn't be able to patent the idea of a computer.

All the "facts" of math, and yes, I am using the term "fact", exist whether or not you are aware of them, as do all the laws of science. 2 + 2 equaled 4, plants used light to make sugar, and the earth continued to trace an elliptical orbit around the sun long before we understood, before we "discovered" the laws and theorems that explained it, long before we derived the algorithms to predict them.

Back in the day, when patents were more sensible, people understood this. Software, like math, or science was never patentable.

This Perl code (by Adam Back aba@dcs.exeter.ac.uk)
"print pack"C*",split/D+/,`echo"16iII*oU@{$/=$z;[(pop,pop,unpack"H*",
)]}EsMsKsN0[lN*1lK[d2%Sa2/d0 Tue, 22 Jun 2004 17:09:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? So the theorem stated as "The square of the hypotenuse of a right triangle is equal to the sum of the square of the other two sides." isn't patentable, but an algorithm for determining the length of a hypotenuse of a right triangle should be?

The algorithm you propose does not require any "new art" apart from the theorem. If you can't see that, go look at what the RSA algorithm actually does, and compare that to the theorems that support it. There is a much greater difference than there is between the Pythagorean theorem, and using algebra to solve the problem. If you can't understand that, you really should stop arguing with me.

On the other hand, if I create a device that solves for a hypotenuse quickly and conveniently, in a meaningful way, and uses some innovative art that is not obvious from the theorem, then yes, I'd say it should be patentable, just like the cash register was patentable, when all it did was provide a new method to do things we've been doing for thousands of years.

An invention is using a discovery, usually several, to create a device that didn't exist before, and wouldn't have been obvious to someone working in the field before you invented it.

Right, and software is a device, pure and simple. After all, I could make a mechanical device to do what any program does.

I remember reading something about a "secret-service" that had discovered the idea of public key encryption...

Your analogy is flawed, since "public key encryption" is an idea, not an algorithm. RSA is an algorithm; ElGamal is an algorithm; "public key encryption" is merely an idea. I don't think you should be able to patent the idea of "public key encryption", but you should be able to patent the algorithm for it.

2 + 2 equaled 4

No; 2 + 2 equals 1 in a field of characteristic 3. Math is not like science: mathematical truth is not based on correspondence; it is based on consistency with certain assumptions.

A particular computer design is a device, an invention and so it is potentially patentable.

A particular computer design does nothing more than solve a problem. Why should it be patentable, and software not, since I can implement the computer design in software and run it in an emulator? This is why I maintain that software is also a device.

Hopefully you can see that discoveries are different than inventions.

Yes; Christopher Columbus discovered the New World; he didn't invent it. But an invention is still a kind of discovery, so -- discoveries may be different than inventions, because they're a more general idea. A software algorithm is a kind of invention. A theorem on the other hand, is a completely different kind of discovery.

You can't patent something that was already invented. Everything that can be discovered has already been created by God, so you can't patent it.

Now you're simply being silly. Think about it: I could thus argue that God foreknew photocopiers, or that God has his own method of making copies of documents, so it shouldn't be legal to patent them. Honestly, that's one of the most pathetic lines of reasoning I've read to date. I like to think that I am a deeply religious man, but I don't think that whether God foreknew something should determine whether it's patentable (or, for that matter, classified). Let's just leave God out of this, okay?

If you have a problem differentiating between a discovery and an invention, ask yourself these two things?

Now you obviously aren't paying attention to what I wrote. I did differentiate between a discovery and an invention: I said that an invention is a kind of discovery, and a software algorithm is a kind of invention. If you want to argue with me, please don't build straw men.

If it's a tangible object, if no one invented it, would it still exist?

The RSA algorithm did not exist before its invention. On the other hand, I don't agree with your distinction anyway. Many mathematicians (at least one of my advisors for example) refer to mathematics as the creation of knowledge, not its discovery. In other words, it doesn't exist before someone creates it. There is in fact strong evidence for this point of view. First, the note I made above that 2+2=1 in a field of characteristic 3. Second, depending on the postulates you pick, there are three mutually contradicting geometries (Euclidean, Lobachevskian, Riemannian), each one of which is enormously useful in some real-world situation, and each one of which is logically consistent, and hence "true", according to the rules of mathematics.

So, I still disagree with pretty much everything you say, before taxes no less :-) Tue, 22 Jun 2004 18:23:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "With this I understand you, as well as your point (quite salient) that the method ought to patentable in its own right."

Ok.

"That gives me pause: is that at all the theory behind how patent law works? If patent law works differently -- since if it works differently, there's a terrible weakness in my understanding & thus in my argument. What I wouldn't give to have a competent patent lawyer to talk to... "

I'm not quite sure what you mean here. But let me try something:

Actually, it's basically the law we're discussing here. What we're talking about is the difference between two versions of a directive that basically is a specification for how the law should be in the different member states.

Now there's two versions. That of the EP and the one of the Council. The one of the EP does not allow patents on "computer implemented inventions" unless the underlying method adheres to proper definitions of "technical" and "invention".

The version of the Council has such a weak definition of what is to be regarded as an invention, that a patent claim like "a computer running a program doing this and that" adheres to the definition of "invention" and is therefore patentable.

That means if you can somehow craft your idea^H^H^H^H method into the definition above, your method becomes patentable, regardless if the underlying method is patentable or not.

So, for example, you could patent a business method if you can apply your method to a webshop, since a webshop runs on a computer,... You get the idea. Tue, 22 Jun 2004 18:44:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Right, and software is a device, pure and simple. After all, I could make a mechanical device to do what any program does.


Computer scientists sometimes talk about software being a device, but that does not make software a device. It is just a colloquialism. And you can not build any software in hardware. For example, a software program can easily specify infinite memory usage, but no hardware device can provide infinite memory.


A particular computer design does nothing more than solve a problem. Why should it be patentable, and software not, since I can implement the computer design in software and run it in an emulator?


Because the conditions are different. If you do design a computer and run it in an emulator, you can violate any real world facts of nature of your pleasure. Thus you may do something really interesting, but you do not really advance the state of the art of electrical (computer) engineering.


I said that an invention is a kind of discovery


That is a stretch, in my opinion. Sure, an inventor will probably discover something. But an inventor will ask himself the question "How can I use this discovery to build some device?". Thus his aim is genuinely to create something new, not to prove (discover) that something new will actually work.


The RSA algorithm did not exist before its invention


Really? I tend to think that a mathematical system "creates" every possible operations which can be deduced using its laws.


Many mathematicians (at least one of my advisors for example) refer to mathematics as the creation of knowledge, not its discovery. In other words, it doesn't exist before someone creates it


I do not think this means much. In the German language, the word "science" translates to "creation of knowledge", or "the entity which creates knowledge" (the word reads "Wissenschaft"). So every science creates knowledge. So by that reasoning, everything scientists know should be patentable. Which, you have stated, you do not want. Tue, 22 Jun 2004 20:23:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? > A theorem is not a kind of algorithm, nor vice-versa. In
> mathematical journals, algorithms are published alongside
> theorems that state, and prove, the algorithm's correctness.

There are softwares, that is algorithms which prove theorems, how about those? The theorem itself is of course not the algorithm, but the proof of the theorem is an algorithm. Tue, 22 Jun 2004 21:15:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Computer scientists sometimes talk about software being a device, but that does not make software a device. It is just a colloquialism. And you can not build any software in hardware.

I don't agree. The original computers were programmed to do different jobs using actual hardware modification. Software (ie computer programming) was invented to get around that; software is simply redesigning the device, without physically rearranging the gates. The fact that today we do it at a high level with words and mnemonics and the rest does not for a moment exclude the fact that, at the low level, there is actual machinery involved (gates & the rest).

If you do design a computer and run it in an emulator, you can violate any real world facts of nature of your pleasure.

Not really. I can't pretend to have infinite memory, because the computer running the emulator does not, in fact, have infinite memory.

Thus his aim is genuinely to create something new, not to prove (discover) that something new will actually work.

This is because he doesn't have to: he simply demonstrates that his device works. Increasingly, computer algorithms are going the same way: many of them are unproven, but work on "good heuristics" which give validity to the method. I brought up the bit about theorems and algorithms merely because I wanted to distinguish between what I consider a genuine, unpatentable idea (ie a theorem) and a patentable method (an algorithm).

(The RSA algorithm did not exist before its invention)

Really? I tend to think that a mathematical system "creates" every possible operations which can be deduced using its laws.

Operations on the objects of mathematics (addition, subtraction, tensors, etc) are defined, not invented. Operations don't require proof. Algorithms to solve real-world problems are invented, not defined. You have to prove the algorithm.

In the German language, the word "science" translates to "creation of knowledge", or "the entity which creates knowledge" (the word reads "Wissenschaft"). So every science creates knowledge. So by that reasoning, everything scientists know should be patentable. Which, you have stated, you do not want.

I don't think that you understand my point at all -- by your logic, I would be advocating the patenting of theorems, which I have explicitly and repeatedly excluded. It's not knowledge that should be patentable; it's activity, it's implementation, it's method -- and not just any activity, but innovation that truly advances the state of the art.

The particular statement to which you were responding, was merely my illustration that many mathematicians believe that mathematics is not a discovery of something "real", but that mathematics is how the human mind understands real things. Again: we have three completely contradictory geometries, which are completely true in their appropriate contexts. Euclidean geometry (on a flat plane) is inappropriate when applied to the universe (which is curved), or to navigation on the surface of the earth. It won't correspond to the facts, whereas Lobachevskian and Riemannian geometries will. Which one is true? They're all true. But they contradict each other! So? They're simply systems of thought; they're not "real things" in some Platonic heaven where the forms reside. Tue, 22 Jun 2004 22:13:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? There are softwares, that is algorithms which prove theorems, how about those? The theorem itself is of course not the algorithm, but the proof of the theorem is an algorithm.

I don't have an inherent problem with patenting devices that implement such algorithms, but it simply won't happen. There's no point to it, because (as I explained in another post) the algorithms are useful only for proving that theorem. Once the theorem is proved, only a fool would copy the algorithm to prove it again. It's absurd.

The RSA algorithm to encrypt a message, however, can be used over and over again on many different messages. It's not a one-time device. This difference is not trivial.

Now, I suppose there are some cases where the device that implements the algorithm that proves the theorem, might be useful in other situations, without modifying the device at all. I don't know of any, and quite frankly if the device is in fact necessary, and if I can't simply do the effort by hand (as I could with the induction algorithm, say) then I don't have a deep moral problem with a patent on the device -- like RSA. I repeat: as I understand it, the RSA patent doesn't exclude me from encrypting messages by hand using the RSA algorithm, for pedagogical reasons or for research or even for private use -- after all, it was illustrated for years on blackboards in US universities, and toy problems were done with small integers. The patent only prevents me from distributing software or any other other device that implements the RSA algorithm, because the inventors of the RSA algorithm deserve a reward for their work, via profit from the device they've created. Am I wrong about what the patent permits/prohibits? Tue, 22 Jun 2004 22:37:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? Software is a document, perhaps a complex document, but it is nothing more.

Replace the word "software" with "document" in any argument and see if it makes sense. I have seen no case for software being anything other than a document presented.

e.g.,
The patent only prevents me from distributing a document that implements (explains) the RSA algorithm.

Because some separate device can interpret the language I used in the document and "do" something with it does not make the document mechanical, or anything beyond an idea. The fact that a computer can read the document and do something with it is completely irrelevant and does not change the fact that software is nothing more than a document. Wed, 23 Jun 2004 03:13:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The Dutch Minister of Economic affairs mr. L.J. Brinkhorst has placed misleading information inside a Letter to the European Council.

Inside that Letter he says that Dutch Parliament agrees fully with the Council's proposal on software patents. A proposal in which the Irish presidency reversed all complaints and arguments back to the original filed proposal on which we all came up in arms in the first place. So the Council actually was presented with false information just prior to a new vote on software patents. Minister Brinkhorst says it all was a Word processor typo, but in a less friendly environment this would have costed him his position, and be accused of forgery.

Well now, the original Letter including typo seems to be missing currently. Maybe there are some Council members who still have a copy (with the presumed typo) in their possession ?

regards,

Robert Wed, 23 Jun 2004 05:11:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? The letters the Minister and the Secretary of State sent to the parliament are publicly available. You might want to check out:
http://www.minez.nl/default_volg.asp?pagina=kamerbrieven Wed, 23 Jun 2004 08:46:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? But the RSA *algorithm* is not a document. The document should be copyrightable, but the algorithm itself *is* an invention and should be patentable. It's not obvious, there was no prior art and it has a specific scope, namely encrypting something in this specific way.

There is no difference in principle between a patent on a new type of rear axle, or on an algorithm. They're both ideas.

It's going to be a problem when the description of the patent is non-technical and overbroad. The rear axle will have a quite specific construction. So should the software patent have a pretty specific description.

Patents should not be able, and are really not allowed to, consist of a summary of general principles. The "One-Click" patent is a general principle. It should have been only possible to patent if a detailed description of the process was given, not the simple fact that you can buy something with just 1 click.
And that leaves a workaround. The patent might restrict the most efficient way, but it should not make it impossible to solve the same problem in another way. Just like you can design a different type of rear axle with about the same effects as the patented one, so you should be able to design a way to let people buy things with just 1 click, just by using text files instead of a database, or a different database scheme, or another order of transactions.

That also makes you wonder what the use is of software patents. Because it is usually not that difficult to find another way of doing thinks to achieve the same (kind of) effect in software, while in mechanical engineering it will cost you a lot more to design another solution for the same problem.

Because it is relatively simple to find multiple ways to solve a problem, I think the real problem with software patents is that to get a meaningful patent (read: monopoly on solving a problem) the only thing you can do is patent *solving the problem itself*.

It's exactly what you see happening in the US.
And that is frankly ridiculous. Wed, 23 Jun 2004 11:47:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? "The letters the Minister and the Secretary of State sent to the parliament are publicly available. You might want to check out:
http://www.minez.nl/default_volg.asp?pagina=kamerbrieven
"

These are _all_ Letters to the Dutch Parliament. I would be interested to see the actual Letter from the Dutch Minister such as the members of the European Council effectively received, preferably a printed out copy or fax with proof of Date and Time of receival. As we all now know, shortly after this, the original proposal on software patents (without the complaints and adjustments) was hammered away as a done deal.

Robert Wed, 23 Jun 2004 23:14:00 GMT donotreply@osnews.com (Anonymous) Comments http://osnews.com/thread? http://osnews.com/thread? > I don't have an inherent problem with patenting devices that
> implement such algorithms, but it simply won't happen.
> There's no point to it, because (as I explained in another
> post) the algorithms are useful only for proving that
> theorem. Once the theorem is proved, only a fool would copy
> the algorithm to prove it again. It's absurd.

I was talking about algorithms that prove theorems, not a specific theorem.

There's even a multitude of languages suited to such kind of tasks, and each of them implements the kind of algorithm needed for proving theorems. Ever heard of "prolog"? Fri, 25 Jun 2004 17:48:00 GMT donotreply@osnews.com (Anonymous) Comments