Developers on the Windows compatibility project Wine have been forced to abandon an important feature due to the existence of a Borland patent.
The open source project Wine allows you to run unmodified Windows programs on Linux through a binary compatibility layer. It also allows you to compile Windows applications to create a Linux binary through a tool called Winelib. In a talk at the annual Wine Developer Conference in Stuttgart earlier this month, Wine developer Dimitrie Paun said that the development of Winelib is restricted by a Borland software patent.
A Borland spokeswoman said it did not have sufficient information on the patent or the open source project to comment on the situation.
Concerns about the Borland patent have prevented developers on the GCC project from adding structured exception handling (SEH) to the free software compiler. As SEH is commonly used in code that runs on the Windows platform, Wine users that want to use the Winelib feature must either modify their source code to remove SEH constructs or rely on a proprietary compiler, such as Microsoft's MSVC.
Paun told Builder UK on Thursday that relying on proprietary compilers is an unsatisfactory solution, and could impact the adoption of free software projects built for the Windows platform. "From a long term perspective, this is a problem because it means GCC is not even an option for people," said Paun. "They cannot adopt the native Linux toolchain, and they are stuck with using the Windows tools indefinitely."
"Having your FOSS project depend on a non-free tool can be a big problem in terms of adoption. Many distributions, such as Fedora [a free Linux distribution], have a principle of using only free software — they wouldn't be able to distribute such a package as they wouldn't be able to build it."
Paun said that software patents, such as the Borland patent, shouldn't be allowed, as they give large corporations an advantage over open source projects and SMEs. "Software patents are senseless and immoral," said Paun. "They slow down progress and cost everybody with the notable exception of a few very wealthy corporations that have signed cross-licensing deals."
Wine isn't the only open source project forced into an unsatisfactory work-around to avoid violating a patent. Last month, developer on the open source database PostgreSQL said that they had rewritten a piece of code to avoid violating an IBM patent.
Despite the growing number of examples of open source projects needing to rewrite code to avoid violating patents, the UK patent office has denied that the European software patent directive, which is due to go to parliament for a second reading in July, poses any threat to open source software.
Talkback
The UK patent office denies the patent directive is restrictive. Do they understand software or is it they are just influenced by govenrment/large companies?
Response from Borland's Chief Scientist Danny Thorpe:
http://blogs.borland.com/dcc/archive/2005/05/12/4294.aspx
Two paragraphs are of interest. The first is on the philosophy of patents themselves:
"Whether software patents are the root of all evil or merely a device in our intellectual economy is really beside the point. Patents suck, sure, but it's the system that we're given. You can choose to abstain on principle, and then be walked all over by those who use the system against you. Or, you can choose to defend yourself and use the system to make the schoolyard bullies regret bothering you. "
Finally, it would appear that no-one actually bothered talking to Borland; they simply went crying to the media:
"If you're seeking reassurance that some artifact won't be used to sabotage a project, just ring the doorbell and ask. There are much better ways to open a discussion than throwing a brick through the window."
This is how Borland and IBM want open source to work - only allowed through their grace and favour.
There is no reason developers of software should go around to all the patent holding fiefs and barons begging patronage, protection and permission to do their job.
Borland did not need to patent this feature, they could have done an IBM and released free for all to use, or they could have published in a journal.
Borland chose to be part of the problem rather than part of the solution.
did you ask Borland about this? your article seems a bit bit one sided?
Borland/IBM are not the enemies as such - they are (understandably) building up their list of patents to avoid being locked out of their own business by other companies that horde patents. The trick is to have a large enough list of patents then sign some sort of cross-licensing agreement at no cost to either party. Sure it sucks, and in Europe the software patents should never be allowed. However, if you're playing in the USA you have to play this silly game. Well, according to one article I read online the World Trade Organization is coercing countries like India to allow software patents. I haven't looked into the claim myself, but looking at the US-Australia 'free trade' agreement, Australia essentially has to support software patents and guess who's got a big say in what the WTO says? The best way to fight this nonsense (where software patents exist) is to a) flood the patent office with hordes of software patents and b) get the OSDL and other such groups to start cranking out software patents. If I'm not mistaken, Ingo Molnar has filed a few patents himself - not that he has any evil plans to suddenly ask people using certain features of the Linux kernel to hand him money, but to make sure that no one else steals his ideas, claims them as their own (via a patent) and then make claims such as those made by SCO... Don't think that you needn't bother because 'prior art' will win - sure it is likely to win, but that means mounting a patent challenge - and that's big $$$ - way more than merely filing a patent.
Borland refused to comment on the patent on the date the article was published, saying that it did not have sufficient information on the patent or the open source project to comment on the situation.