Re: [Patents] "Betrayal of physics"

[Erik Josefsson <ehj@cleosan.com>]

Xavi Drudis Ferran wrote:

> I think the mind-matter approach is needed, but not necessarily 
> as the first retorical device. You first explain why swpats must be avoided
> (because of economics, liberties, etc) and only when that is an stablished 
> need you look at what can be done to avoid them. And for what concerns
> the directive mind-matter is needed.

Another line of reasoning is that software SMEs need to know they own
what they write, and large corporations need not EOLAS and Acacia law
suits. Copyright instantly fulfills those needs. Software patents can
not by definition.

With copyright, the legal object is the code, and infringement questions
are solved based on interpretation of that object.

The infringement interpretation work is done

   1. when products are already in the marketplace
   2. only when called upon

With patents

   1. ownership is not instant and not guaranteed
   2. abstraction work must be done before the product is in the
      marketplace (not only when called upon)
   3. the legal object is the abstraction and the infringement
      interpretation work is based on that object (not the code)

With software patents, if you put disadvantages of the uncertainty of 1.
and the administrative cost of 2. aside, 3. is the most worrisome 
because the abstraction is necessarily broader than the code and the 
interpreation of the abstraction extends the scope of the exclusion furher.


One thing I would like to ask Greg, when you bust software patents, what 
is it that normally busts a swpat? Is it source code?

Swedish judge Lennarth Törnroth says published source code (e.g. on the 
internet) can be an infringement. Which implies source code can also be 
prior art. But to me that leads to an incoinsistency, because you 
compare two different legal objects: source code is not a patent document.

How does it work in the US? Can you claim a piece of source code is 
prior art? Perhaps there are examples in "IBM disclosure bulletin"?


With (normal) patents 3. is natural since you can't bring the
caterpillar into the courtroom. With software patents (an instruction to 
a person skilled in the on how to perform a computing process) you could 
argue that there is no "upper limit" to the abstraction. The abstraction 
can be arbitrarily abstract (broad).

Text-objects are also different from caterpillars. They can be 
translated*. The mere existence of translations of text-objects proves 
the need for the interpretation work in copyright infringement cases 
since byte-by-byte comparison would make any imitation an original.

You can't as far as I understand translate a physical process or object.

Perhaps another way to express mind/matter?

//Erik

*) see Tarski, Gödel and lambda calculus