How close should the structure and nomenclature of the schema be to XDI

[bruceravel]

There are things in the XDI dictionary that are not in this schema and vice versa. That's not necessarily a problem or even an issue. However, there might be merit to aligning them as closely as possible.

For example, the Spectra Table (why is it called Spectra Table rather than Spectrum Table?) has an entry called temperature while the XDI dictionary suggests nomenclature for 15 other extrinsic sample parameters in the Sample. namespace.

[newville]

@bruceravel there are definitely singular/plural form issues throughout. I think singular is preferable too, so it should be Spectrum.

I agree that this table should more closely match XDI. In fact some of the recent changes were to try to do that. For example, just have a dspacing entry per spectrum, and not have (in earlier versions) a Monochromator table. Some things map easily, and some (say, Facility) are completely doable, just a bit more complex in a database than a single spectrum file -- we want 1 entry per facility.

Anyway, yes, I agree.

[newville]

Trying to square XDI and the library, I think these are not so far off now. But I came across this:

https://github.com/XraySpectroscopy/XAS-Data-Interchange/blob/master/c/xdifile.h#L172 be

Should that be "beamline.xray_source"?

[bruceravel]

Hmmm... perhaps that makes more sense.

It is Facility.xray_source in every single example file, so it will have to be changed everywhere it happens.

[bruceravel]

On the other hand, the source is on the facility's side of the ratchet wall....