Linked by Thom Holwerda on Tue 6th Apr 2010 18:32 UTC
IBM And thus our true colours reveal. IBM made a patent pledge in 2005, promising not to sue open source projects over a list of 500 patents the computer giant holds. Today, however, IBM has threatened to sue TurboHercules, a French open source software house which provides support for the Hercules open source s390 mainframe emulator. Some of the patents in question are on the 500 list.
Thread beginning with comment 417702
To view parent comment, click here.
To read all comments associated with this story, please click here.
saynte
Member since:
2007-12-10

The EU statement is essentially a special-case for computer programs. It does not disambiguate algorithms from processes by virtue of their definition, it just rewrites the definition of invention to exclude computer programs.

This doesn't answer the question of how a process and an algorithm are different. Algorithms are not the same as computer programs, btw.

Reply Parent Score: 1

lemur2 Member since:
2007-02-17

The EU statement is essentially a special-case for computer programs. It does not disambiguate algorithms from processes by virtue of their definition, it just rewrites the definition of invention to exclude computer programs.

This doesn't answer the question of how a process and an algorithm are different. Algorithms are not the same as computer programs, btw.


Is there a better way to define that which should be patentable, and that which should not? What exactly is wrong with naming things which should not be patentable?

If "mathematical methods" in general should not be patentable, why is there even a need for the EU to mention "programs for computers"? If the EU had not mentioned the special case of "programs for computers", then software would STILL not be patentable under the exclusion of "mathematical methods".

At least this way it is perfectly clear, software should NOT be patentable.

As for the more esoteric WHY of it, that comes down more to the case of the particular economic circumstances surrounding software development. These are, in part:

(1) patents have a chilling effect on software development,
(2) software development has a low cost of entry, and it makes no sense to build artificial barriers to entry into a market, as that would just inflate prices for no good reason,
(3) the term for patents is way too long for software,
(4) it is all too common in software for the same method to be independently arrived at by any developer who sets out to achieve a particular task, (i.e. true "invention" in software is rare)
(5) software development is accelerated by healthy competition, and it stagnates completely wherever there is an artificial monopoly of supply,
(6) in software, patents are used far more to eliminate competition, erect barriers to entry for new players in the market, extract exorbitant ongoing profits, and REDUCE the need for ongoing improvement to products than they are used as incentive for more development,

I'm sure there are other factors. All of these particular economic circumstances of software development make it highly unsuitable as subject matter for patents.

So why not just name software as being unsuitable for patents? What could be simpler?

Edited 2010-04-08 10:14 UTC

Reply Parent Score: 2

saynte Member since:
2007-12-10

If "mathematical methods" in general should not be patentable, why is there even a need for the EU to mention "programs for computers"? If the EU had not mentioned the special case of "programs for computers", then software would STILL not be patentable under the exclusion of "mathematical methods".


Why is an algorithm a mathematical method and a process is not? What happens if I hard-code my algorithm to a circuit instead of a computer (which is essentially a dynamic circuit) ? To me, this is an incredibly fuzzy area.

Edited 2010-04-08 11:32 UTC

Reply Parent Score: 1