Of course they are right that they may write libraries which may be linked against OpenFOAM. This is what I expressed in the first sentence of my previous post.

But they may not distribute libraries including installation scripts or guides under a GPL incompatible license, if they are built for the single purpose to be linked against OpenFOAM.

>See here :
>http://www.gnu.org/licenses/gpl-faq.html#GPLPluginsInNF

Have you even followed that link? That FAQ entry discusses more or less the opposite of what Chris plans.

I wish him good luck, or better the wisdom to read and understand one page of terms he will depend on. Further discussion in this forum seems to lead nowhere.

Bye
Guillaume

[alberto]

Hello,

At the moment OpenFOAM is released under GPL 2: http://www.gnu.org/licenses/old-licenses/gpl-2.0.html

Under the same licence Linus Torvalds distributes the Linux kernel, and the use case proposed here is identical to those of accelerated video drivers (they need the kernel sources to create the module and they link it against proprietary code). What you can't do (even though some ditributions like ubuntu are doing that all the time), is to distribute the binary form of the module.

The only problem I see here originates when you want to distribute compiled/linked forms of the ported version of OpenFOAM/CrossFire or if you don't want to open the source code (you'd violate GPL), which would be at least unfair in all the cases.

There are two useful FAQ's here:

http://www.gnu.org/licenses/gpl-faq....oneyGuzzlerInc
http://www.gnu.org/licenses/gpl-faq....olledInterface

With kind regards,
Alberto

[deepsterblue]

My belief is - Why rely on nVidia/ATI proprietary software in the first place?
Use OpenGL - it's open-source and free. No legalese to deal with either.
I guess nVidia/ATI are good with the hardware aspect of things, but they don't comprehend the requirements of CFD too well. It's only fortunate that graphics applications have vector-processing requirements that are similar to CFD, but it's all just a random-access memory-bandwidth issue.

[mbeaudoin]

Hi,
>Have you even followed that link? That FAQ entry discusses more or less the opposite of what Chris plans.

Which was exactly the point I wanted to make...

> Further discussion in this forum seems to lead nowhere.

On the contrary, I think the discussion rapidly converged to the right conclusions, thanks to the contribution of many persons, yours included...

Cheers.

Martin

[alberto]

I agree with Sandeep about OpenGL. It's a well known platform, with a lot of documentation, and its choice could attract a lot more interested people from the community.

Regards,
Alberto