ESP Wiki is looking for moderators and active contributors!

Difference between revisions of "Software is math"

(External links: * [http://arstechnica.com/tech-policy/news/2011/08/appeals-court-says-only-complicated-math-is-patentable.ars Does not compute: court says only hard math is patentable], Aug 2011,)
(Related pages on {{SITENAME}}: According to the EPO, as written in EPO EBoA referral G3-08 (''page 12 of 18''): <blockquote> computer programs were to be understood as a 'mathematical applic)
Line 38: Line 38:
  
 
Another approach would be the Curry-Howard isomorphism, which demonstrates that computer programs are equivalent to mathematical proofs.  If proofs are unpatentable, then computer programs must be too.
 
Another approach would be the Curry-Howard isomorphism, which demonstrates that computer programs are equivalent to mathematical proofs.  If proofs are unpatentable, then computer programs must be too.
 +
 +
==EPO says software is math==
 +
 +
According to the EPO, as written in [[EPO EBoA referral G3-08]] (''page 12 of 18''):
 +
 +
<blockquote>
 +
computer programs were to be understood as a 'mathematical application of a logical series of steps in a process which was no different from a mathematical method
 +
</blockquote>
  
 
==Related pages on {{SITENAME}}==
 
==Related pages on {{SITENAME}}==

Revision as of 17:49, 13 February 2012

Mathematical formulas are generally recognised as non-patentable because math is not patentable subject matter.

Since the logic (idea) of software can be reduced to a mathematical formula (idea) with Church-Turing Thesis, and because mathematical formulas (idea) are not patentable, patent applications for software ideas should be rejected.

Respected computer scientist Donald Knuth makes the argument:

To a computer scientist, this makes no sense, because every algorithm is as mathematical as anything could be. An algorithm is an abstract concept unrelated to physical laws of the universe.[1]

Math is not patentable

Case law in the USA

In the USA, math is unpatentable because it is a "law of nature", that is to say a "scientific truth", and as such it can never be "invented", only "discovered", and patents are not granted for discoveries.

The non-patentability of math was confirmed in the case Parker v. Flook (1978, USA):

Respondent's method for updating alarm limits during catalytic conversion processes, in which the only novel feature is a mathematical formula, held not patentable under 101 of the Patent Act.

Also, in the 1948 case Funk Bros. v. Kalo Inoculant:

He who discovers a hitherto unknown phenomenon of nature has no claim to a monopoly of it which the law recognizes. If there is to be invention from such a discovery, it must come from the application of the law of nature to a new and useful end.[2]

Ideas which use math can be patentable, but this is not controversial:

While a scientific truth, or the mathematical expression of it, is not patentable invention, a novel and useful structure created with the aid of knowledge of scientific truth may be.[3]

Church-Turing Thesis or Curry-Howard isomorphism?

There are two mathematical bases that can be used to make this argument.Can you help? This page was written by a non-specialist. Any help would be appreciated.


The Church-Turing Thesis is the more commonly used based. It is discussed by some documents linked in the #External links section.

Another approach would be the Curry-Howard isomorphism, which demonstrates that computer programs are equivalent to mathematical proofs. If proofs are unpatentable, then computer programs must be too.

EPO says software is math

According to the EPO, as written in EPO EBoA referral G3-08 (page 12 of 18):

computer programs were to be understood as a 'mathematical application of a logical series of steps in a process which was no different from a mathematical method

Related pages on ESP Wiki

External links

Counter arguments

References