Thanks for your answer...
Mmmh, you should explain the dependence relationship better. How do these executables interact with bamg? Do they invoke it by forking another process? Do they link against it? Do they incorporate parts of bamg source code in their own code?
The invoke it by forking another process (more precisely the system method of C++ is used).
What does he agree with? That bamg is distributed along with other files? Or that bamg is distributed under the terms of the GNU GPL v2?
He agreed on the distribution of bamg along with rheolef, we are allowed to include bamg's sources in our own.
This depends on your answer to the above questions regarding the relationship between bamg and rheolef...
Doesn't the fact that bamg's sources are now part of rheolef obliges us to make it GPL ?
Now it gets a little bit more complicated (I hoped we didn't have to go this far :). Bamg is undevelopped for more than a year now. Instead it's author is working on a new application (FreeFem++), which includes some of the bamg code, but creates a more complete binary (bamg is only a part of it). Now if he accepts to release the former code of bamg (the one I use) under GPL, won't he be forced to do the same for FreeFem++, since bamg code has been used in it. (in case this is revelant, FreeFem doesn't produce any bamg binary, it's using the sources directly) ?
I'll wait until I have a clear view on the issue before contacting him again to propose GPL :).
Alexis Papadopoulos