Relationship between the `coreCIF` and `symCIF`

[vaitkus]

What exactly is the relationship between the DDLm version of coreCIF and the DDL2 version of symCIF [1]? Is the intention for coreCIF to fully absorb symCIF?

I noticed a while ago, that some definitions from symCIF have been ported to the coreCIF (e.g. the SPACE_GROUP_SYMOP category), however, I only now realised that this is not true for all data items from symCIF (e.g. _space_group.transform_Pp_abc and _space_group.transform_Qq_xyz are not in coreCIF). I also see that there have been some other changes like making the categories only describe a single space group instead of a set of space groups (e.g. _space_group_symop.sg_id is no longer defined).

I am asking because the magnetic CIF (magCIF) dictionary [2] still references data items from the symCIF dictionary that are not defined in the curent DDLm version of cifCore.

[1] https://www.iucr.org/__data/iucr/cifdic_html/2/cif_sym.dic/index.html [2] https://github.com/COMCIFS/magnetic_dic

[jamesrhester]

This is an important question. The missing items you identified have been removed pending a way for more than one space group to be referenced, as they represent transformations between space groups, i.e. you need two. Now that we have ironed out the way that multiple blocks interact, we can reintroduce these data names in the multi-block dictionary, with a data name that points to a "transformed from" space group described in a separate data block.

[vaitkus]

Understood. But what about _space_group_symop.generator_xyz? It could potentially be used to describe a single space group, but is not defined in CIF_CORE. Maybe it could be added as an alias of _space_group_generator.xyz?

The other missing items which should probably go to the multi-block dictionary:

• _space_group.id
• _space_group.reference_setting
• _space_group.transform_Pp_abc
• _space_group.transform_Qq_xyz
• _space_group_symop.sg_id
• _space_group_Wyckoff.sg_id