Redefinition of `DIFFRN_DETECTOR`?

[rowlesmr]

Adding things as I work through powder diffraction chapter.

DIFFRN_DETECTOR is mentioned as a category from which data names should be used to give info on the detectors used.

DIFFRN_DETECTOR is a Set category, and PD_INSTR_DETECTOR is a Loop category.

Should we redefine DIFFRN_DETECTOR as a Loop category, and have a link to the _pd_instr_detector.id? either _pd_instr_detector.diffrn_detector_id or _diffrn_detector.instr_detector_id?

[jamesrhester]

This is, like, a crucial question for our raw 2D data presentation as well.

DIFFRN_DETECTOR is a tough category.

First of all, please enjoy the enhancements provided by the imgCIF dictionary, which describes a detector as consisting of a number of elements, and gives all the necessary info to turn a detector pixel into a reciprocal space coordinate.

Here is the definition of _diffrn_detector.diffrn_id - key data name. So we are essentially stuck with diffrn_detector being an outgrowth of diffrn and therefore a new detector means a new _diffrn.id.

In the powder diffraction raw data group, I'm proposing that a 2D detector would be described by items in the imgCIF dictionary - why reinvent the wheel?

So pd_instr_detector is for what's left - those powder-specific things that don't fit in diffrn_detector. Currently that's all the divergence, slit, analyser etc. information.

To answer your question: a particular _diffrn.id implies a detector setup via diffrn_detector.diffrn_id, so the 2D part of the detector is covered by the pd_diffractogram -> diffrn link. So we just need to "connect" the optics. If pd_instr_detector points to pd_instr, and pd_diffractogram points to pd_instr, then we're covered. What's not clear to me is whether or not the pd_instr - pd_instr_detector connection should be one-to-one (one incident beam config to one diffracted beam config), many-to-one, or one-to-many. The answer to that determines how the pd_instr - pd_instr_detector links work and therefore how to plumb them into pd_diffractogram.

[rowlesmr]

In the powder diffraction raw data group, I'm proposing that a 2D detector would be described by items in the imgCIF dictionary - why reinvent the wheel?

I like that. Don't do things twice.

.

I think it would be one incident beam to one or many diffracted beams; at least, I don't know of an instrument that has multiple incident beams, but I know of plenty that have many detectors.

In this, we also need to link the incident beam to the instrument; see #225
The link between PD_INSTR_DETECTOR and PD_INSTR is in #204

[rowlesmr]

We're not redefining DIFFFRN_DETEECTOR, but we are doing deep rethink - please put further discussion in #231.