Joost van Zwieten

If you want to assemble matrices in parallel, then WSL is the way to go. Docker, which (nowadays?) uses WSL2 as well, is an option for production code. We don't...

```python >>> from nutils import element >>> import numpy >>> element.getsimplex(3).slice(numpy.array([ 0.12818456, -0.00017979, 0.06756571, -0.03730343]), 8) NotImplementedError ```

One way to go about this is to extend `function.Argument` with spaces. If an argument is bound to one or more spaces, it cannot be evaluated, but it can be...

Or if every `evaluable.Array` supports a variable number of leading pointwise axes, we can actually introduce the required axes with `evaluable.replace_arguments`.

I think I found the culprit. In `crates/store/src/query/sort.rs` method `Store::sort` there are three variants for sorting messages. The first applies to a single comparator and without collapsed threads, the second...

I think the problem is with the definition of `R`. The length of the basis is proportional to the number of elements. By appending a row to `R` for each...