operator id(u, c) and direct broken scalarplot!

[pjaap]

No description provided.

[pjaap]

The plotting logic is a bit messy. This adds another if/else in the code. We should clean this up "in the future" :)

[chmerdon]

I just had a look at the documentation and expected lots of discontinuous plots, but I only found few. For example in https://wias-pdelib.github.io/ExtendableFEM.jl/dev/module_examples/Example201_PoissonProblem/ the gradient is plotted still as a continuous function, since taking the Gradient from a continuous H1P1 element does not trigger the piecewise plotting. Is that fine or do we want to change that?

[chmerdon]

and I like the id(u,c) operator

[pjaap]

I just had a look at the documentation and expected lots of discontinuous plots, but I only found few. For example in https://wias-pdelib.github.io/ExtendableFEM.jl/dev/module_examples/Example201_PoissonProblem/ the gradient is plotted still as a continuous function, since taking the Gradient from a continuous H1P1 element does not trigger the piecewise plotting. Is that fine or do we want to change that?

The gradient is a vectorplot and this is currently not implemented. Only scalarplot for broken FE.

[chmerdon]

Yes, but I mean the abs-plot of the vector.

[pjaap]

I just talked to @Liameloy if he wants to refactor the whole plotting logic into vector/scalar/grid with and without abs and broken or not. He will have a look after next Berufsschule week.

[pjaap]

It is also hard to detect which operators produce broken functions. grad of H1P1 does, but grad of H1P2 not. Or we plot first order opearators grad, div, ... always broken? We should discuss a concept, I guess

[chmerdon]

grad of H1P2 also produces (L2P1) broken functions, we (unfortunately) don't have H2 elements yet :). I think we can base this on NeededDerivative4Operator(operator) > 0.

[pjaap]

Pls merge if it is fine. I want to release the new version for the pre-strain plots 😄

[chmerdon]

Looks fine

[pjaap]

grad of H1P2 also produces (L2P1) broken functions, we (unfortunately) don't have H2 elements yet :). I think we can base this on NeededDerivative4Operator(operator) > 0.

That's true. I had a mix-up. NeededDerivative4Operator(operator) > 0 is indeed the key. Let's come back to this thread in two weeks.