Nick Papior

The plotly tests needs restructuring for easier maintenance. Instead of directly importing stuff it may be better to use `pytest.mark.parametrize` to leverage different possibilities. - [here](https://github.com/zerothi/sisl/pull/293#issuecomment-773199600) for details about documentation...

Xfail should only be used for to-be-fixed bugs!

**Describe the feature** DM + rotating spin + tiling

**Describe the feature** We want to track performance regressions. We should use [asv](https://asv.readthedocs.io/en/stable/)

Maybe there could be two categories, an "overlaps" for the overlap situation, and a "neighborhood" where R then should be specified? I think this would cover the relevant use cases....

Create an IPython siesta analysis tutorial

3

- [x] molecule analysis - [x] non polarised (k) - [x] polarised (k) - [ ] non-colinear (k) - [ ] spin-orbit (k) - [ ] grid handling - [...

Castep cell files

1

**Describe the feature** Enable reading castep cell files.

**Describe the feature** A way to implement SCF models in sisl: This involves: - [ ] *SCF classes which enables iteration with convergence criteria - [ ] A generic class...

We need a better way of handling `names` and other attributes. Having a common class to control attributes and changes to parent structures may be the way forward?

Tests currently have `class`es all over. These should be removed.