Create individual files for every unversioned I-D instance

[ronaldtse]

From https://github.com/ietf-ribose/relaton-data-ids/issues/16 by @strogonoff

For every draft-xxx-nn, we create individual bibliographic items for every unversioned instance draft-xxx that contains relationships using the contains relationship to the individual versioned drafts (draft-xxx-nn).

This will provide an index for software to know the full series of an I-D without needing to find individual version files.

Related attempts to address similar discovery issues:

• https://github.com/ietf-ribose/relaton-data-ids/issues/6
• https://github.com/ietf-ribose/relaton-data-ids/issues/15

[strogonoff]

I don’t mind creating an unversioned item for each unversioned I-D and use it as series.

Note though it is not a requirement for series to be defined as a separate bibitem, we have plenty of series among Relaton datasets that do not have their own separate bibitems (and BibXML service converts those to links just fine):

• W3C NOTE — see https://dev.bibxml.org/get-one/by-docid/?docid=W3C+NOTE-imagemap&doctype=W3C

• NIST — see https://dev.bibxml.org/get-one/by-docid/?docid=NIST+CSWP+02042022-5&doctype=NIST

[ronaldtse]

we have plenty of series among Relaton datasets that do not have their own separate bibitems

W3C NOTE is not a series, it's a document type. W3C items are going to be separately encoded as versioned vs unversioned entries.

NIST CSWP is a proper "series" but it is not commonly cited in the way of versioned vs unversioned, dated vs undated.

[strogonoff]

I’m just setting the priorities. Specifying series can be done first, and unversioned bibitem YAML files could be created later. (Since there are precedents where series don’t have separate bibitems in current YAML data, and it works fine for BibXML service and does not violate Relaton spec.)

NIST CSWP is a proper "series" but it is not commonly cited in the way of versioned vs unversioned, dated vs undated.

We don’t have NIST CSWP as series, we have only global NIST series. (This is actually a problem, filed https://github.com/ietf-ribose/relaton-data-nist/issues/5.)

[ronaldtse]

@strogonoff While the "NIST CSWP publication series" (and the other 50+ publication series) can technically be citable, we don't need to make it citable at this stage because there is just no practice that uses it today. For sure IETF will not need them.

[ronaldtse]

Relevant to https://github.com/relaton/relaton-ietf/issues/74

[andrew2net]

For every draft-xxx-nn, we create individual bibliographic items for every unversioned instance draft-xxx that contains relationships using the contains relationship to the individual versioned drafts (draft-xxx-nn).

@ronaldtse contains isn't allowed as relation type in the grammar. In the relaton-iso we use instance.

[strogonoff]

I think rfcsubseries uses "includes", too

[ronaldtse]

@andrew2net instance is for a physical instance so we shouldn't use that. Then we use include/includedIn.