Thomas Titscher

> Additionally I think we should agree on the interfaces of an integrand first. The result of about one year of discussion was to _not_ have a common interface. :)...

@joergfunger Could you briefly elaborate on why you prefer the decoupling? I am afraid that I overlook some issues with that. Maybe to clarify, I was just asking myself: -...

I like the idea of the shape class that may have some more geometrical information. And I like the idea that the actual creation of the interpolations/integrations is done in...

My argument for the _TriangleFactory_ was more like that it enables adding a whole new factory from the outside. Like adding `TriangleDiscontinuous` or whatever. But this may belong directly into...

You may be looking for something like: ~~~cpp auto TimeDepGradient = [&](double deltaT) { return [&](CellData cd, CellIpData cipd) { return momentumBalance.Gradient(cd, cipd, deltaT); }; }; // ... auto _...

1) I vote for the same repo - for the same reasons as listed above. 2) I am not sure about header-only or lib. I always have this little performance...

Developer documentation wishlist: [google-perf-tools](https://github.com/gperftools/gperftools)

> This can be solved much better with a dedicated integrand. Yup :smile: So are you fine with just a double value at the cell?

Ok, I see. I think in most cases we want a constant multiplier. Practically speaking, we will implement far more Integrands than cells. That's why I would put it at...

Forth option: `TimeIntegration` just _uses_ the `TimeControl` object, without modifying it. User creates a `TimeControl` object, modifies it as he wishes and copies/moves/references it to `TimeIntegration` which itself provides no...