This little gem comes from NZ via Computerworld. Tuesday, 06 April 2010
Thumbs down for software patents in NZ Commerce Select Committee tips its hat to open source submissions
Open source software champions have been influential in excluding software from the scope of patents in the new Patents Bill.
Clause 15 of the draft Bill, as reported back from the Commerce Select Committee, lists a number of classes of invention which should not be patentable and includes the sub-clause ?a computer program is not a patentable invention.?
Site Map: http://www.dontronics.com/sitemap
What makes software different? If you can patent a novel circuit or mechanism, why shouldn't you be able to patent a novel software process? Having been, at different times in my life, a hardware designer and (currently) a software developer, I do not see a difference between the two, they are the same thing carried out by different means.
The main problem that I see is the way that software patents are issued, often for ill defined and questionablely novel ideas, but then that can also apply to other types of patents too.
Yay! Though, frankly, a more reasonable approach would have been to *greatly* reduce the period in which patent a patent on such "art" is granted -- and prevent the patents from being "renewed".
Given how quickly the industry moves, it's hard to imagine
*anything* being worthy of protection for more than a few (e.g., 3-4) years.
I think if that were the case, folks would avoid the patent hassle entirely, rely on trade secret, as needed, and just plan on being "nimble" in the marketplace instead of trying to live off old (artificial) glories...
so if I go in my basement and work on a new mousetrap for 10 years I can patent it, but if work on a new algorithm for 10 years it's somehow not worthy of protection?
a program is just a virtual machine
-------------------------- so if I go in my basement (were I to have one) and work on a new book for 10 years why shouldn't I patent that. No, software is much more like a book than a machine. Algorithms are much more like mathematics. More vague algorithms such as user interfaces and heuristics are much more like a novel. That's what copyright is for.
The purpose of patents, as I understand it, is to give the inventor protection to commercialise there idea while making the invention available for the public good in the long run. If you want to protect software in the same way then the code needs to be published in the patent application and be freely available for public use once the patent duration expires. Copyright, does require you to expose you code and the protection period is far longer.
There comes a problem with patenting software, consdiring the book analogy someone else cited.
Patenting software is like patenting a book of a specific size, type and colour of paper, cover photo, and binding combination. The copyright for a book exists mainly on the content of the book (and ancilliaries like illustrations, photos, cover layout.
Consider a novel circuit patent, and someone adds one resistor to improve it, they have improved upon the patent, not infringed it.
How are you going to patent the software in binary form or at source level in what language for which host. One could agrue using a different compiler has improved upon the patent because using a different compiler or compiling switches has speeded the software up hence improved upon a previous patent.
The algorithm used is best described without actual software as in
5 years everybody could be using it written in a different language for three new instruction sets of new cpus.
Just the byte oreding of the storage could make it non-identical,=20 let alone different instruction sets of the host that may improve it=20 again.
How are you going to patant software for any cpu, any language, any bus/register width, any memory size, any byte/word ordering. then prove the software is a copy.
--=20 Paul Carpenter | email@example.com PC Services Timing Diagram Font GNU H8 - compiler & Renesas H8/H8S/H8 Tiny For those web sites you hate
They're very different. It's akin to patenting say, the ripple sorting, or bubble sorting methods, (ignoring prior art for now) or any other newfangled sorting method that's perhaps incredibly faster.
This allows other vendors to develop very similar techniques to your superduper sort method, without stealing your code, or even being aware of your code. It does happen.
Under current law, there is no copyright violation. But with patents, in the event a very similar technique is found in a competitor's code, the competitor would be liable for costly payments for code they wrote years ago, and in good faith, never even knew it was similar to yours.
Or vice versa.
As Dr Phil would say, a good indicator of future behaviour is past behaviour.
The patent offices have proved to be full of idiots, who continually pass through patents that have been in prior art use for a long time.
Or, patenting some very vague description of some vague idea, that could virtually describe anything.
I'm not buying there's inherent built-in protection against that, ensuring all ideas are specific.
The proof is in the pudding, and the pudding smells funny.
AND, depending on code structure, optimizations, and CPU architecture, similar, but different, source code may compile to the same native CPU instructions. So, who is the patent violator there? The code writer or the compiler vendor? Stranger things have happened.....
Another bit of interesting news is that QUT and IP Australia is setting up an experimental crowd-sourced prior art web site. Register on the site and you get to review applications for prior art and send notes on what you find to the offical examiner.
Mark Zenier firstname.lastname@example.org Googleproofaddress(account:mzenier provider:eskimo domain:com)
That's been the wording in Germany since just about forever. Problem is that patent lawyers, as lawyers will, found a loophole to completely subvert the meaning of that clause. They did so mainly by submitting not the algorithm itself, but rather a piece of hardware based on it, for patenting. Once they got the patent, they applied it to other people's pure software works.
Which led to the _completely_ ridiculous situation where a piece of work was expressly denied a patent itself, but could still fall prey to other people's patents.
You could patent it but unfortunately you use stairs to access the basement and as I patented "stair algorithms" many eons ago and all works derived therefrom you owe me big time, of course I waited until you'd made a commercial venture out of your mouse trap before I informed you off this.
To use both the book analogy in both the patent and copyright sense. The result may be a far stronger protection for software. I know about quite a few software patents but I know of very few that generated enough revenue to pay for the costs of protecting and enforcing the patent. I only know of two that actually made real money.
Copyrights are a different matter. There are lots of precedents and the courts know how to handles cases of rewritten to circumvent as well as assign monetary awards. Depending on country the copyright protection can last a long time to a very long time.
There are several current very interesting open cases where copyrights and not patents may affect software. The most interesting one that I currently know about is a standard
40+ year old reference book of polynomial constants.
Dropping software patents may result in software protection with teeth.