David Waterman

Sorry, DIALS is known to have higher memory requirements than other data processing programs because of its global modelling philosophy. We try to optimise this where we can, but 8...

> Beam at angle widget would give you option to reconstruct experiment at given frame number An alternative way of looking at this is that the beam vector display is...

This first requires a change to `Phil2Etree` in the i2 code to generate trivial `guiLabel`s when `short_caption is None`, then the explicit trival captions can be removed. This also applies...

BTW the user reports that doing things the other way round - that is spot-finding and indexing in DIALS, then _exporting_ to XDS - works very well and has been...

If there are multi-lattice images with overlaps on Zenodo, I think that would make a good addition for dials.data. We probably only need a fairly narrow wedge to test success...

Nice. Possibly a bit big for `dials data` though for writing a new test?

Guess we can take one of the tarballs, split to a smaller one and make a new Zenodo upload of just a small sweep

I'll look into doing that with a reduced version of `semisynthetic_multilattice_data_6.tar.bz2`

Right, I have integrated the 6 lattices in the semisynthetic set `acegik`, but according to flags, there are no overlaps: ``` >>> from dials.array_family import flex >>> rt=flex.reflection_table.from_file("integrated.refl") >>> rt.get_flags(rt.flags.overlapped_fg).count(True)...

These really should be overlaps, I'm sure: