OASIS patent policy sparks boycott

TalkBack 35 of 129:: Next »; « Previous

Thread View
Flat View

Why software is different: This apparently simple question goes to the heart of the matter, doesn't it?

Software has always (okay, until very recently) been regarded as too difficult to define, in order to be patented. As far as I know (and I have looked) there are no in-depth studies of why this is true - though commentators who believe it to be true appear to outnumber the rest by something like 1,000:1. But the basic premise for software is; it is too difficult to prove that it is a true innovation (i.e. 1. Completely different from anything that has gone before, and 2. Not obvious).

Of course patents are also, famously, awarded because the patented idea is not something: [not Artwork (some aesthetic creation, or presentation of informaion), not a Scientific Discovery (something that already exists but has not previously been described), and not a Process (which is suprisingly well defined as: a method of performing a mental act such as playing a game, or doing business).

For example:
- A new computer language could not be patented because languages already exist. Even if all current programming languages have limitations (which they do) this is not enough grounds to claim that a new programming language with new features is patentable, because it would fail at least three of the above, standard, global, tests for patents. This means that, if it were possible for the world to agree which one was first, only the first ever programming language would ever gain a patent and even then it is questionable. After all, didn't the complex and time consuming way in which machine code was written cry out for a programming language (i.e. was it not obvious that this incremental step would be taken eventually)? However, does a new function in a language count? So far, the patent office has chosen to believe the answer to this one is yes - but they have yet to be seriously challenged in court.

- A business program might appear, at first sight, to provide a whole new way of doing business. But in reality all such programs that exist today are derivatives of existing business processes - many of them centuries old. This is at the centre of the debate on software patents and OASIS. Software, by definition, is the formal representation (and/or presentation) of either state, or process. The Web Services standards at the core of OASIS' existence are designed to automate and emulate - through software - human interaction (i.e. processes). Any programme that provides Web Services (without patent protection) will be considered to be simply a way to define a process (of the non-patentable variety) so that it will never be patentable. In any case, one of the reasons that business processes were so long excluded from patent protection is that the costs, to society at large, of providing monopoly rights over tacit knowledge (the area from which, obviously, business process improvements spring) was far too high - and in any case were not needed (any advantage that might be gained through business process innovation would clearly be made redundant very quickly as all competitors would have the incentive to deal with that innovation).

According to Brian Kahin [Director, Center for Information Policy, College of Information Studies, University of Maryland]
- "In practice, software is created not by research but drawing from tacit knowledge as an original work of authorship, which is why code-level copyright is widely accepted as an appropriate form of intellectual property protection."

So software is different because it is too difficult to define as original ideas (under current patent rules and processes).

But Brian has said much more: "The complex, systemic, nature of software means that finished products are more likely to infringe patented proceses inadvertently and it may be difficult to remove such processes without rewriting large amounts of code. Portfolio holders, even individual patents, may wield power that is disproportianate to the role of the infringing process(es) in the software package.

[This] systemic failure is understood primarily by patent professionals [lawyers] who lack the incentive to fix it ... the problem is understood to a lesser extent by software professionals who lack the expertise and focus to critique the system effectively. The system [US software patents] maintains the trappings of a property regime but as patenting proliferates in industries where innovation is rapid and incremental, infringment becomes commonplace - as documented in the study of the semiconductor industry. The quality problem ultimately runs up against the notorious subjectivity of the nonobvioussnes standard [used to judge patent applications]. Close scrutiny [required for any new software patent regime] would open up a Pandora's Box of issues that are currently submerged in judicial doctrine ... that can only be evaluated after the fact ..."

Or to put that simply: Software is different because it is far more difficult to judge (if the same old patent rules are applied) what is original and what is not. But, worse than that, the current regime [now being applied to software without any thought as to differences] means that patents are being rewarded for software that is throwing a giant spanner into the ICT industry's works as just one patent could have the power to bring an engine as big as OASIS to a crashing halt.

Or Linux... or Apache... or ...

I could go into more detail, but I have some paying work to get back to...; Posted by: Stephen Wheeler Posted on: 02/23/05 You are currently: a Guest | Members login | Terms of Use