[reza]

Mark,
I am trying to model two blades counter rotating relative to each other connected to the main axis, and also axis has rotating, so two simultaneous rotations. TO my knowledge, OpenFOAM-dev version is the only one that has the capability of simulating such a flow, according to "Different ways to treat rotating geometries", Developed by Olivier petit, for the relative motion of the blades and fort the main rotation I have to modify the code.

[reza]

To be more specific, I am trying to get icoDyMFOAM working and to be able to run the tutorials in the moving mesh directory, the concept of sliding interface is important to me.

[hjasak]

Hi Mark,

You are assuming that copyright ownership is clear and resolved: in fact it is not. I personally hold copyright for all original author contributions I have done from 1993 onwards and have never assigned the rights to OpenCFD or anyone else. The complete story is too painful to write down (the message would have been deleted anyway). In versions after 1.0, list of contributors, the people who wrote the code has been quietly deleted, leaving the impression that OpenCFD actually wrote the code. - again not true. You may understand why I would be upset about that.

Therefore, your basic assumption about who owns the copyright does not stand up - this, together with the issue of adding contributions from the public to the OpenFOAM release are my basic problems. There's no-one from OpenCFD who would even talk about it - therefore, I have to look for another way.

Hrv

[olesen]

Hi Hrv,

Okay, I really do hope that your primary contention is just with attribution and not with copyrights on any particular bits of code since that won't be easy to resolve.

Besides the FSF warning about copyright fragmentation, the more immediate problem is not only deciding when code is worthy of being copyrighted (eg, a one-line patch probably shouldn't qualify) but managing the copyright/code correpondence as well!

As a simple example, suppose I write myClass (with my copyright). Some others add new methods or bugfix bits of it (with their copyright). Much later I realize that myClass would be more elegant if it used a different container. In the modified myClass, many methods and bugfixes were eliminated. Which copyrights are now relevant/irrelevant? I can't really keep all of them if the corresponding code no longer exists. Having copyrighting notices for code that no longer exists is akin retaining the copyright page when recycling the rest of the book. With a book the connection between content and copyright is quite clear, but who can possibly figure out which copyright corresponds to which bits of current code? A nightmare with an enormous potential for endless arguments.

Blindly retaining every copyright is not only senseless (following the above argument), but can also have a strange effect on people. From my own experience with maintaining and developing a bit of open source software ( http://en.wikipedia.org/wiki/Rxvt for people too young or with machines that didn't need it), I've seen people with a real copyright fetish, who make a two-line change and add their copyright. Seeing their copyright on lots of files seems to have been the primary motivation. With a list of contributors, this is generally much less of a problem, since they don't add their own name. It is also significantly easier to manage a single, alphabetically sorted list!

For your second point - about adding in public contributions into the OpenFOAM release. When I think about it more closely, this is completely the wrong problem statement. The issue is simply about who should maintain public contributions and ensure that they continue to work together with future releases. And how to ensure that the porting/upgrading is as painless as possible. When the code is part of the OpenFOAM release, then OpenCFD is maintaining the code. Otherwise the maintenance effort continues where it is, lands somewhere, or the code gets orphaned. If the add-ons are useful and well-written, it shouldn't be terribly important if OpenCFD maintains them on our behalf or the maintenance gets done elsewhere - eg, via the OpenFOAM-extend project.

The real problem as I see it is actually in the layout of the OpenFOAM-extend. At the moment, all the changes are mingled in with the full OpenFOAM source tree. As you probably noticed, it is really difficult to disentangle them from the original and determine which delta might need to be moved with to the v1.5, which ones might be redundant, or even just trying to see what has been touched compared to the original version and why is was touched is really difficult (for me at least). The brute force solution - build a filelist of both trees (1.4.1 pristine and 1.4.1 modified) and then cmp and diff the way through the files - was the only one I could come up with, but gave up when I saw the output. A distinct '-extend' source tree that could be installed parallel and independent from the OpenFOAM release would certainly ease the pain there. I think this is the real problem. I don't envy you in solving it.

/mark

[alberto]

Hi Mark,

I think you misunderstand the terms of the GPL here:

For everyone:
Anyone contributing should really assign copyrights to OpenCFD. Read the FSF links now if you haven't already done so, or read them again. Copyrighting it in your own name weakens the protection for OpenFOAM. Placing a derivative work in public domain is in clear violation of the GPL conditions since it changes the licensing conditions. Don't do that!
A non-derivative contribution (ie, independent of the OpenFOAM structures and libraries) can, however, be sold commercially or placed in the public domain. In this case you have your choice between "Copyright original author", "Copyright nobody" and "Public Domain" - the only real difference is in the number of keystrokes, not in the protection, which is namely none.

The GPL allows you to derive/fork from other's code and to improve it. It allows you to redistribute the code under GPL and eventually to sell it, at the condition you provide it under the GPL and you give the source code to your customer.

As a second point: the copyright has nothing to do with maintaining the code. If not transferred explicitly to someone else, it is common practice for authors to keep the copyright of their contributions in OSS projects. How can you recognize the merits if you don't do that?

Regards,
Alberto

[eugene]

One aspect of OSS copyright has always bugged me: How big a change to the source code do you have to make before you can claim copyright? I.e. can I open a file, add a comment, hit enter to add a new-line and then put my own name as the copyright holder? Or do I only have copyright over the added line? If not one line, then how many? 10%? 50%? 99%? When does the work become a "derivative work"? The GPL as far as I can tell says nothing about this and I cant find any firm guidelines online.

Until this is resolved to my satisfaction the copyright issue is one that I will stay well away from.

Also, I suspect copyright has quite a bit to do with maintaining the code, because:
1. Companies are less likely to maintain code they do not own the copyright to.
2. How many changes due to maintenance do you make before you should assign copyright to yourself?

From my perspective it is a veritable IPR minefield.

Hi all,

I am trying to install and compile OpenFOAM-1.4.1-dev on Ubuntu 8.04
and I am using the build instration in SVN.
Every thing goes well up to paraview section and when I do:

sarah@sarah-laptop:~/OpenFOAM/linuxSrc$ cmake -i ../paraview-2.4.4
Would you like to see advanced options? [No]:No
Please wait while cmake processes CMakeLists.txt files....

CMake Error: The source directory "/home/sarah/OpenFOAM/paraview-2.4.4" does not exist.
Specify --help for usage, or press the help button on the CMake GUI.


sarah@sarah-laptop:~/OpenFOAM/linuxSrc$

But paraview-2.4.4 exists.
I do not know what to do to get it working,any idea?
I am new to linux and openfoam, please flog me if you need more info.

Cheers

Hi,
The problem is that I put every thing in linux directoy and "/home/sarah/OpenFOAM/paraview-2.4.4 does not exist, where can I change the path to know, is it in CMakelists.txt? and where?

[reza]

Dear Sarah,
Have you tried to build dx before?
I had this problem, and I had tried to build dx. When I removed:
OpenFOAM-1.4.1-dev/applications/utilities/postProcessing/graphics/dxFoamExec

The CMake worked, however I still have problems in the next two lines; gmake and gmake install

[reza]

Dear all,
During the installation of dev version and after ./Allwmake I get some errors which cannot find -lfintievoulme and some others. I thought it might be related to libiberty and installed binutils 2.16.1. After :
export PATH=$HOME/OpenFOAM/linux/binutils-2.16.1/bin:$PATH
and
../configure --prefix=$HOME/OpenFOAM/linux/gcc-4.2.2 --exec-prefix=$HOME/OpenFOAM/linux/gcc-4.2.2 \
--enable-languages=c,c++ --enable-shared --disable-multilib

there is the configure error: cannot create executables. When reopening the terminal configure works, but which binutils show:
/usr/lib and there is no libiberty. Meaning the path given above is gone.

Does any one know how to do that?

Thanks

[noriaki_nishio]

HI Hrvoje

I installed OoenFOAM-1.5.

And , I want to use thin liquid film solver, using (FAM).

But I couldn't find that folder existing that program .

Where's directory is that program?

Regard..