yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

Hi there,

The note:
Let me recall that Dem3Dof is unmaintained and links to many known and
unfixed bugs (see below).
If some of you are using this class without problems, good for them
(still, the fact that Yade doesn't crash doesn't mean that the result is
correct...).
But for long terms development, another geometry class should be
preferred, else the old bugs will re-surface one day or the other.

The question:
A long standing bug in Yade is the different force sign convention
between ScGeom and Dem3Dof. If Dem3Dof is unmaintained, there is no
reason to keep its sign convention as the default. Reverting the default
will let everybody run simulations using the maintained code without the
need to add "revertForce=True" in scripts.
The problem is - in turn - those still using Dem3Dof would have to
specify "revertForce=False" in their script, so one question is how many
users would be affected. I can see Jan impacted by such change. I think
Anton is also using Dem3Dof but not periodic BCs (?), which is the only
place where revertForce really matters. Is there somebody else using
Dem3Dof?

Another geometry class is L3Geom, which uses the same sign convention as
Dem3Dof, but it is (I think) not used currently. It is still possible to
revert the sign in L3 without much risk, so that there would be no more
sign conflicts in maintained code (assuming somebody will develop and
maintain the L3 framework).
What do you all think?

Bruno

https://bugs.launchpad.net/yade/+bug/585898 (concluding that Dem3Dof is
unmaintained and that we stop reporting bugs on it)
https://bugs.launchpad.net/yade/+bug/585771
https://bugs.launchpad.net/yade/+bug/399963
https://bugs.launchpad.net/yade/+bug/675955

-- 
_______________
Bruno Chareyre
Associate Professor
ENSE³ - Grenoble INP
Lab. 3SR
BP 53 - 38041, Grenoble cedex 9 - France
Tél : +33 4 56 52 86 21
Fax : +33 4 76 82 70 43
________________