Stephan Rave

I don't use it but I have nothing against including it. However, when doing so, we should probable make it clearer in the `Coding style` section of `CONTRIBUTING.md` that falke8...

So, I am no expert here, but a few comments: - Before thinking about adding some fancy algorithm, we should think about the actual use cases for the algorithm. For...

pyMOR's current toolkit took quite a bit of effort to write. Still, it is rather quite limited in what it can do. Expressing finite element assembly via vectorized operations is...

Ok, so I guess the general opinion is, that there is currently no external solver package want want to use to replace pyMOR's discretization package, and a complete rewrite is...

We should start thinking about `discretize_stationary_XXX` methods, where XXX is our favorite PDE solver. However, this would probably require a more abstract definition of the data functions ..

With #1576, it should also be relatively easy to properly handle two-sided comparisons like `a < x[0] < b` or boolean operators.

We actually have the policy that `seed` parameters go along with `random_state` parameters as in `ParameterSpace.sample_randomy`. Grepping through pyMOR, I seem to be the only person following this policy. I...

> What bugs me a little bit with the latest version is the encapsulated redundancy of the assertions and generation of random states in the `rrf` methods and `VectorSpace.random`. I...

> It shouldn't be a problem really, just lets me think about whether its cleaner to manually do the seeding when necessesary and then only pass random_state to the functions...

Let's continue the big design discussion in #1577.