🚀 go-pugleaf

In article <HANS.91Sep4094915@bilbo.es.ele.tue.nl> hans@es.ele.tue.nl (Hans Fleurkens) writes:

   on 3 Sep 91 23:02:56 GMT,
   kap1@icarus.spc.uchicago.edu (Dietrich Kappe) said:

   =>    => This question is perhaps answerable for less than one million dollars in
   =>    => legal fees, but I doubt it.

   =>    => A layman's reading of the law would seem to say that algorithm patents
   =>    => are not permitted.  But the Patent and Trademark Office disagrees.  So
   =>    => do many patent lawyers.

   =>    Probably because they can earn a lot of money with it. Finally, the
   =>    possibility to get a patent for an algorithm or piece of code will
   =>    kill the software industry (except for a few big companies) and they have
   =>    no job anymore :-).

   =>    Hans Fleurkens
   =>    Eindhoven University of Technology.


   => Distribution of source code should still be permitted, right, since that
   => is equivalent to the description of an algorithm rather than its
   => implementation? A text file is after all not executable by a computer
   => (under normal circumstances), and distributing source code amounts
   => to publication and (though wordy comments) discusion of the patented
   => material, which is what patents are supposed to encourage, right? 

   If someone wants to prevent his invention and make some money, just hide
   the implementation details from everyone. It will still take other companies
   lots of time to built an equivalent system, if they think it's worthwile
   to develop a similar program. Up to now I only heard of patents on trivial
   things a child of six years old could re-invent. If someone wants to use
   this technique, he has either to pay for it or work around it by using
   another (maybe less efficient) technique (if he's aware of the fact
   this technique is patented). Even worse, say you can patent look and
   feel in world which will be standarized soon (X, Motif, etc.). 

   What would happen if a car manufacturer would patent a steering wheel?
   Another one could either pay a lot of money or use two buttons to steer.
   A third one could ....

   Hans Fleurkens
   Eindhoven University of Technology.

My point is that in order to patent something you have to describe it
fully. Thus an algorithm would have to be completely specified. If
distributing source code were to be made illegal, we could draw a very
absurd conclusion: since we can design a compiler for any cononical
algorithm description method (making those descriptions in effect
source code), no two algorithms may be described in the same way
by anyone, or even described in such a way that they could be translasted
into a cononical description method by a well defined procedure.

Does this stuff seem loopy to anyone else?

--

---
Dietrich Kappe
kap1@icarus.spc.uchicago.edu