David Waterman

This has come up again: we don't have a way to modify the crystal in the same way that we do other aspects of the geometry.

this is somewhat trickier than the other models as we can't set arbitrary values in the **UB** matrix. But being able to set scale factors for the cell lengths and...

@elena-pascal did we talk about this at some point?

Looked a bit closer. The issue happens here https://github.com/dials/dials/blob/8aef2b1dc2a56df77f1fb4accabeb1ba45b5e99f/src/dials/algorithms/symmetry/laue_group.py#L788-L790 because the argument to the first `sqrt`, i.e. `self._n * self._sum_x_sq - self._sum_x**2`, is zero. This seems to be a corner...

For a sample size of $n=2$ the only possible values of Pearson's correlation coefficient are in ${-1, 0, 1}$ anyway. Performing the calculation probably only makes sense for $n \geq...

NB `dev.dials.simple_integrate` is the script I meant. See #2115

Ok, not really a "bug", but non-standard. Columns `H`, `K`, `L`, `M/ISYM` and `BATCH` are set here: https://github.com/dials/dials/blob/d8064546435e2db85f76e1d1675fedca6eb93ee5/util/export_mtz.py#L306-L312 After that it would be helpful to sort the MTZ data accordingly,...

of course what you are _supposed_ to do is run, e.g. `dials.integrate -h`

See also #1429 for a suggestion to change the behaviour of `dials.program -h` to add value over just typing `dials.program`

The scan-varying detector model does not work outside of refinement - that is, the varying detector is not part of the model used by the predictor that `dials.integrate` uses. I...