[Fauster]

Hello everyone,


I don't know exactly all the story but why all the standard solvers of foam-extend are locked on a version similar to OpenFOAM 2.2.0 (or even older)?



If it's Open source, why foam-extend couldn't benefit of improvements of other branches ?


For example and regarding VoF, semi-implicit MULES is not available in foam-extend package. This can make a huge difference for some cases.



Thanks to everyone,


Best regards

[Santiago]

Originally Posted by Fauster

Hello everyone,


I don't know exactly all the story but why all the standard solvers of foam-extend are locked on a version similar to OpenFOAM 2.2.0 (or even older)?



If it's Open source, why foam-extend couldn't benefit of improvements of other branches ?


For example and regarding VoF, semi-implicit MULES is not available in foam-extend package. This can make a huge difference for some cases.



Thanks to everyone,


Best regards

You are welcome to implement it! foam-extend is a non-commercial fork of OpenFOAM: you cannot expect these so-called "improvements" to be attached just for the sake of it, the community is small and narrow-focused. foam-extend has a very clear line of development towards block-implicit methods/solvers, immersed boundaries, Overset, and GGI: there's where most energy is put on. To make the counter-argument: Why the other releases of openfoam insist on their versions of Overset, despite the fact of being close to useless compared to FE?

Besides, what makes you think that solvers from the 2.2.0 era are wrong, and the newer are "better/good"?

[Fauster]

Originally Posted by Santiago

You are welcome to implement it! foam-extend is a non-commercial fork of OpenFOAM: you cannot expect these so-called "improvements" to be attached just for the sake of it, the community is small and narrow-focused. foam-extend has a very clear line of development towards block-implicit methods/solvers, immersed boundaries, Overset, and GGI: there's where most energy is put on. To make the counter-argument: Why the other releases of openfoam insist on their versions of Overset, despite the fact of being close to useless compared to FE?

Besides, what makes you think that solvers from the 2.2.0 era are wrong, and the newer are "better/good"?

Thank you for your answer. I clearly didn't say that solvers from 2.2.0 era are wrong. I asked why they are "locked" on a old version of the code. From your answer I understand that it comes from both size and focus of the FE community. As user I have done many tests for interFoam and newer versions are clearly better (both accuracy and robustness). For clarity, I didn't want to criticize FE. Just imagine all the capabilities of FE (solids mechanics, overset, immersed ...) merged with OF8 for example !

Best
F

[Santiago]

Originally Posted by Fauster

Thank you for your answer. I clearly didn't say that solvers from 2.2.0 era are wrong. I asked why they are "locked" on a old version of the code. From your answer I understand that it comes from both size and focus of the FE community. As user I have done many tests for interFoam and newer versions are clearly better (both accuracy and robustness). For clarity, I didn't want to criticize FE. Just imagine all the capabilities of FE (solids mechanics, overset, immersed ...) merged with OF8 for example !

Best
F

"Both accuracy and robustness (sic)" Is that a fact, or just a statement read here in this forum or in OpenFOAM's website? Seems to me like a rather arbitrary statement, considering that you can even implement MULES fully implicitly, if one wanted to, and still need to have Co < 1 in order to conserve mass (still VoF)... Anyway, in the times of Ray tracing, isoAdvector, and localEuler schemes, every VoF solutions "looks" like the correct solution, for those who don't know better.

About critizising FE: well, it really sounded like that; not that I care for either fork, but it really bothers me this discourse of XX version is "better" than YY because... NO ACTUAL REASON: just copy-paste of version control readme's and comments made here or there. If there is an actual reason, sustented by rigurous research, proving that a solver is better, then I can live with that, otherwise it's just noise.

[Fauster]

Originally Posted by Santiago

"Both accuracy and robustness (sic)" Is that a fact, or just a statement read here in this forum or in OpenFOAM's website? Seems to me like a rather arbitrary statement, considering that you can even implement MULES fully implicitly, if one wanted to, and still need to have Co < 1 in order to conserve mass (still VoF)... Anyway, in the times of Ray tracing, isoAdvector, and localEuler schemes, every VoF solutions "looks" like the correct solution, for those who don't know better.

About critizising FE: well, it really sounded like that; not that I care for either fork, but it really bothers me this discourse of XX version is "better" than YY because... NO ACTUAL REASON: just copy-paste of version control readme's and comments made here or there. If there is an actual reason, sustented by rigurous research, proving that a solver is better, then I can live with that, otherwise it's just noise.

Thank you for your very kind answer.