Andrius Merkys

I agree with the removal of actual property definitions. AFAIK, they do not contain anything which is not in #445, and there should be a single source of truth. >...

This is however blocked by the pending PRs introducing or modifying already existing properties. I would suggest resolving them first, or reworking them into PRs modifying property definitions directly.

It has been decided to move cheminformatics properties to a repository of its own, which has been done in https://github.com/Materials-Consortia/namespace-cheminformatics/pull/1 and https://github.com/Materials-Consortia/namespace-cheminformatics/pull/2. Closing this PR.

I had a somewhat similar idea when exploring the idea to store OPTIMADE responses or even serve them statically. I was thinking more about file tree layout than single file,...

I think the idea was to limit querying to the information which is normally returned within the response. But I do not see why optional support past that could not...

This issue has resurfaced today in the workshop in the same context of filtering structures on the results of their calculations. Thus I think it deserves to have its severity...

> I guess for OPTIMADE v1.x we can retain this special handling of `meta.description` of relationships, with minimal loss. In any case `files.description` is supposed to be a human-readable string,...

Good idea. I wonder whether we can re-use JSON:API relationships to describe every archive member in a same way the archive itself would be described. For sure we would need...

It seems that most of F/LOSS structure-to-IUPAC tools are based on neural nets nowadays. I did an extensive search of a rule-based solution, but was unable to find one.

This could be fixed by introducing a command line parameter to set a fixed seed for randomization. I see the use of reproducible validator runs.