viscous semidiscretizations

[jlchan]

What's the current status for viscous semidiscretizations in Trixi (e.g, for compressible Navier-Stokes + BR1 discretization)? I only remember seeing a message in Slack a while back.

[ranocha]

There's some WIP in #378. However, it's not finished and it isn't clear whether that's the approach we really want to use in the future. The only way to find that out seems to be to invest some work into it such that it does the job for now, merge it, mark it as experimental, and see how it's going.

[jlchan]

Thanks! I was wondering how to cleanly prescribe gradient BCs. Seems like an open question still based on #378 (comment)

[sloede]

Yes, it is :-/ The only sensible way forward seems to be what @ranocha said above:

The only way to find that out seems to be to invest some work into it such that it does the job for now, merge it, mark it as experimental, and see how it's going.

[jlchan]

Should we close this now and update #891?

[sloede]

Should we close this now and update #891?

Sounds good to me! We should probably also close #378, since it's been superseded by the new approach and I don't really see us going back to that approach...

[ranocha]

Sounds good to me. We should also have another issue tracking this, e.g., other numerical fluxes for the second-order terms, don't we?

[jlchan]

Do you mean #1147?

[ranocha]

Yes. Let's close this issue here and continue with the others