[Patents] DeCSS, source code and software patents

[Didier Metzen didier.metzen1@libertysurf.fr]

Xavier Drudis Ferran a écrit:

> Didier Metzen va escriure el Wed, 02 Aug 2000 00:52:14 +0200  :
> >
> >Computer programs = information that is "disseminated". I am surprised by 
> >this
> >concept.
> >Do you apply it to the source code, the binary... How do you consider a
> >firmware embedded in an appliance?
> >
> Well, I don't want to take away any merit from PILCH Hartmut, because it is 
> absolutely
> pertinent to point it out, but what he says, that computer programs  (in 
> any form) are
> information is such a basic principle of computer science that I (being a 
> computer engineer)
> wonder how anyone can be surprised (until I realise not everyone has to 
> have a computing
> background).

Of course in the theory of information, everything is information: programs, 
knowledge, datas... And they
are treated the same way by the machine....

Patents have nothing to do with scientific theory: it's business, politic and 
economy.
Information has not the same meaning.


>
> What is the main advantatge of computers?. That they are general purpose 
> machines, so that
> the same machine can solve an immense number of problems. How can this be?.
>
> Simply because you do not supply it with only the data it has to process, 
> but with other information
> telling it how to processs it (the program). The computer always follows 
> the same algorithm,
> as a Von Neumann machine, which interprets the information that is the 
> program on the
> information that is the data to produce other information which is the 
> result. The only difference
> between program and data is how you look at the information at each given 
> moment, but they
> are both information. A compiler is a program that uses as data other 
> programs, for instance.
> You can make a program that modifies itself because the program is just 
> data the program itself
> can manipulate.
>
> What is a patent?. Something by which you publish (disseminate) a 
> description of some
> useful process (or device to perfom a useful process) and get a temporary 
> monopoly on the
> device or process. The patent description is not different from a program 
> (only that is written
> for humans to interpret, not computers).  So, if programs can be patented, 
> what stops you to
> patent (with patent B) the published description of another patent on A, so 
> that you get the
> monopoly on A and nobody can read A's description (which is practising B) 
> without paying
> you royalties for B?. If that happens then you are not actually publishing 
> A's description in
> any useful manner, so you shouldn't be able to patent A.
>

Well a patent is information.
But more important a patent is granted by a community (states/nation in fact) 
for a business use.
Patents are not natural. The notion has been created by human communities to 
answer some problems some years
ago. (like a lot of laws)

One of the goal, for example, was to have the inventors explain their process.
Before patents, secret was the general rule. Patents have been a way to get 
the products and the way they
were designed published. (today we would say for software: publish the source 
code)
The counterpart for the community: give to the inventor the possibility to 
make money and a protection
against copy or imitation.

I think it would be interesting to understand more in depth how and why 
patents have been created.

Today the environment is different. And perhaps today patents do not make 
sense any longer and are not any
longer usefull for the community.

By the way, you won't get a patent on the description of a patent because 
this doesn't meet the criteria...

>
> To put in more computer science terms. You patent an interpreter for a 
> computer language you
> invent, and then patent an application written in your computer language. 
> The source code of
> the application should be a sufficient description to patent, because there 
> can be no more
> precise description. But even publishing that source code means nothing 
> without knowing the
> language, which is perfectly defined by the source of the interpreter (I am 
> not assuming, although
> I could, that the interpreter is written in the same language that it 
> interprets). So you get two source
> codes that everyone can read but nobody but the one patenting them can use. 
> The person
> patenting them can show you they are useful by running them, but nobody can 
> know how the
> application works, because they do not know the computer language you 
> invented, and the only
> way to know it is to execute (even if it is with pencil and paper) the 
> interpreter. If you say you can
> understand how the application works by merely analysing the source of the 
> interpreter, the patentor
> can claim that that analysis is allowing you to interpret (by hand) the 
> source of the application, and
> that is what is covered by the interpreter patent, so you can't do it 
> without a license.
>
> That is,
> with computer programs, there is no distinction between understanding its 
> code and running them
> (only that running them is usually easier if you have the means). That is, 
> they can't allow you to
> read a program and not to run it, just because the program is information. 
> So how can you
> publish it (for people to read) but monopolize its use if reading it is 
> basically the same as using it?.
>
> But I am very dumb in law, does any of these ramblings make any sense?.
>

By the way, trying to get a patent is expensive. Which means that you still 
need to have a financial
interest when you are trying to get a patent. You can patent an interpreter, 
but nobody will use it. So
what's the deal?
Your example doesn't seem realistic to me.

Another important point: all patent applications are not supposed to be 
granted. There are rules about the
importance for the community of the invention and the fact it is a real 
innovation due to the applicant
research and development.
Real innovation means also that nothing equivalent has been published or 
developped before.

I know that patents offices have a trend to grant patents without real 
review, and I think this is one the
real cause of the problems not only with software patents but also in general 
with patents in all area.
Trying to avoid patents on pure software is a way to solve part the problem 
but only in the software area.


>
> --
> Xavier Drudis Ferran
> xdrudis@tinet.org

Didier