Allow default values for site symmetry in the geometry categories?

[James H]

Dear Core DMG,

We have adopted as a rule of thumb that we do not provide default values
for data names when defining them in the dictionary. This is justified on
the grounds that a missing value doesn't mean that the most likely value
should be substituted, as the programmer may simply have not output that
data name. So, for example, if the temperature is missing, that doesn't
mean room temperature.

This blanket rule included the GEOM categories, which are used to describe
bonds and angles. Each atom participating in the bond/angle is identified
by providing the "site symmetry": which means in this case the symmetry
operation and cell translation that should be applied to the asymmetric
unit coordinates. If the atom in the bond just has its asymmetric unit
position, the site symmetry would be "1_555". A request has come through to
make this the explicit default site symmetry, that is, if no site symmetry
is supplied then it is assumed to be "1_555".  The justification for
allowing an explicit default in this case is that an error is easily
detectable, as the GEOM categories can be reconstructed from the supplied
atomic sites. In this way it is qualitatively different from the
temperature example above, where there is no way of recovering the true
value. Some further practical benefits of the request are outlined in
https://github.com/COMCIFS/cif_core/issues/362#issue-1652795179.

If this group has no objections to the above proposal, I further suggest we
adopt a general principle that default values can be provided in the
dictionaries if an error can be detected by derivation from values that are
known to be present. In this case, as the GEOM categories explicitly refer
to the contents of both ATOM_SITE and SPACE_GROUP(_SYMOP), the values from
which the bonds and angles are calculated should be present.

Please comment.

all the best,
James.

-- 
T +61 (02) 9717 9907
F +61 (02) 9717 3145
M +61 (04) 0249 4148
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.iucr.org/pipermail/coredmg/attachments/20230404/ce24056b/attachment.htm>