Huw Jenkins

So I think the behaviour of XDS appears to be the complete opposite of DIALS where as you say "All reflections that touch an excluded frame are marked as ~integrated"

Yes `INTEGRATE.HKL` is also identical as long as the blank images either are or aren't excluded in the `INIT` step for both runs (otherwise the scales are different due to...

> Perhaps XDS excludes the blank images anyway? no they are read - the `INTEGRATE.LP` files differ: not excluded: ``` IMAGE IER SCALE NBKG NOVL NEWALD NSTRONG NREJ SIGMAB SIGMAR...

Going back to DIALS :grin: The behaviour when passing `exclude_images_multiple=20` to `dials.integrate` with respect to *summation* integration is to me quite surprising. This is dataset 1: no `exclude_images_multiple=20`: ``` Summary...

> I can understand the summation integration result being rejected if any pixel from the peak region is on the masked frame This does appear to be the cause of...

It looks like this is irrelevant. I tried increasing the number of decimal places in the `.ins` file but even in Olex2: (which re-formats the `CELL` and `ZERR` instructions) once...

This is very interesting! I looked at the thaumatin_i04 data with a slightly different approach. 1. Integrated data as usual with DIALS. 2. Scaled with `dials.scale` using: ``` dials.scale symmetrized.expt...

So of course I did: ``` dials.scale symmetrized.expt symmetrized.refl d_min=1.35 intensity_choice=profile a=0.80932 b=0.02742 ``` This gives: ``` Error model details: Type: basic Parameters: a = 0.80932, b = 0.02742 Error...

The reflections from dials with I=0 appear to be a xia2 thing. I cannot reproduce with 'normal' dials workflow. They have I=0 σ(I)=0: ``` >>> from dials.array_family import flex >>>...

For the small molecule ED dataset what does the table ``` +--------------------------+----------+------------------------+----------------------+ | Intensity range () | n_refl | Uncorrected variance | Corrected variance | |--------------------------+----------+------------------------+----------------------| ``` in the `dials.scale`...