Functional error identified in file mandels_problem.ipynb

[OmarDuran]

Attempts to employ catersian grids in the tutorial notebook produce incorrect approximations for the Mandel's problem. Possible source of the error is the combination of mpsa, cartesian grids and some kind of BC. The following MWE derived from the tutorial is exhibiting such functional error.
mwe_mandels_problem.py.zip

[IvarStefansson]

Can this be closed, @OmarDuran? I believe the explanation is inherent to the discretization method, right, @keileg?

[keileg]

Yes, there's no getting around this. The only question is if we should leave an explicit warning in the tutorial not to use such a combination of methods (I'm not sure if there is such a comment now). I'm not sure I am in favor of such a change.

[IvarStefansson]

Could you please take action as you see fit and then close the issue?

[OmarDuran]

My understanding of the advantages of a multipoint finite volume method is twofold: lowest order approximations and general meshes, including cartesian meshes. @IvarStefansson I am unable to understand how the explanation is inherent in the discretization method. Fixing this issue, however, is quite involved, and I would rather not take any action in this regard rather than leave this issue open for the record of users.

[IvarStefansson]

Sorry about the misunderstanding. My suggestion was for @keileg to add the warning if required.

[keileg]

Sorry about the misunderstanding. My suggestion was for @keileg to add the warning if required.

I'll take care of it, though it may take a week or so to find the time.