Corrected the definition of a* in _atom_site.B_iso_or_equiv and similar

[nautolycus]

Following correspondence with Massimo Nespolo (18-09-2025); He also provided a link to his paper [Nespolo, M. (2024). J. Appl. Cryst. 57, 1733-1746] for more detailed explanation:

Dear Brian,

Sorry for the late answer, semester has started here.
The problematic sentence is indeed the one where a*, b* and c* are called ‘reciprocal-space cell lengths’. They are the 'reciprocal of the direct-space cell lengths', which do not coincided with the ‘reciprocal-space cell lengths' unless the unit cell is primitive. The fact that the reciprocal of a unit cell is not a unit cell (unless the latter is primitive) has gone unnoticed for so long time that I was really surprised when I "discovered" it. But now that it has been pointed out, the imprecise sentences require correction.

Best wishes.

Massimo

[vaitkus]

Maybe the a* definition should also be updated in the templ_attr.cif file, e.g. in save_aniso_uij, etc.?

[nautolycus]

Maybe the a* definition should also be updated in the templ_attr.cif file, e.g. in save_aniso_uij, etc.?

Yes, that's already been done. Sorry not to make it explicit in the comments.

[vaitkus]

Sorry, I somehow managed to miss it.

[vaitkus]

This PR was created before the templ_attr.cif was moved into a separate repository. Sorry for not merging it before the split. The simplest approach now would be to close this PR and create two new PRs: one dealing with the templ_attr.cif in the new repository (https://github.com/COMCIFS/Attribute_Templates) and one dealing exclusively with the cif_core in this repository.

[jamesrhester]

This PR is now replaced by #592 and COMCIFS/Attribute_Templates#7 with no changes.