Governance Group minutes 2023-05-21

Tuesday 23 May 2023, 13:00

Attendees: Petr Knoth (PK), Mick Eadie (ME), Paul Walk (PW), Nicola Dowson (ND), Beverley Jones (BJ), George Macgregor (GM - Chair)

Minutes of last meeting

GM welcomed RGG members. The minutes of last meeting were reviewed and agreed as accurate.

In a review of actions arising, GM confirmed that the following actions were completed:

  • Action: GM/PK to raise the issue of ambiguity in the use of PIDs for AAMs with members of the UKRI Technical Requirements Group.

GM confirmed that the issue of community misunderstanding surrounding the application of PIDs for AAMs had been raised with Sara Ball (SB - UKRI). Community misunderstanding appeared to be traced to some loose wording within the 'technical requirements' section of the original UKRI Open Access Policy. Alternative wording was agreed with SB and SB confirmed that UKRI policy documentation would be updated and circulated to institutions. GM also confirmed that he had issued a clarification on the UKCORR member list, trailing UKRI's updates.

  • Action: ME and GM to explore the modelling of Enlighten data within the existing schema but also within the suggestions made within the Rioxx model documentation and those made within the meeting. Following the conclusion of testing and a modelling consensus, circulate findings to RGG for agreement and publication.

GM confirmed that ME, PK and GM had collaborated to resolve Rioxx modelling issues, as manifest in the documentation circulated to RGG members and the focus of the present meeting's purpose (to finalise Rioxx v 3.0 - below).

On the final action:

  • Action: Upon schema finalisation, GM to draft blog post for publication and thereafter community circulation.

GM noted that since the schema had not yet been finalised it was not possible to post a blog and organise any other dissemination. This action therefore currently remains outstanding but will be addressed soon after schema finalisation and publication.

Rioxx v3 - discussion of schema examples

Prior to the meeting GM had circulated documentation, detailing proposed schema changes arising from the work of ME, PK and GM. To open discussion GM summarised some of the principal changes proposed and indicated the substantive decision-making issue for the RGG was around the accommodation of multiple persistent identifiers (PIDs).

  1. Should multiple PIDs be accommodated? And, if so:
  2. How should they be accommodated?

Variant approaches in the treatment of multiple PIDs was suggested in the documentation. In instances where a repository system wishes to capture more than one PID to define a property, two main approaches were proposed: "whitespace delimiting", and; the creation of element hierarchies.

In addressing Q1, PW expressed concern about the potential for internal contradiction and therefore wanted there to be a sound use case for multiple PIDs and the. GM noted that the prospect of modelling mutliple PIDs had been discussed by extensively with ME and PK. Their conclusion was that some repositories were already storing and seeking to expose multiple PIDs, particularly in relation to authors. In further comments PK also noted that it also enabled systems not to make decisions about which PIDs to expose in Rioxx. While accommodating 'downstream' systems was not necessarily within the remit of Rioxx, GM commented that the ability to include mutliple PIDs would also indirectly satisfy this particular use case.

PW and other RGG members agreed that there needed to be treatment of mutliple PIDs within Rioxx.

RGG members considered the PID modelling approaches provided in the documentation.

Members agreed that modelling based on "whitespace delimiting" was undesirable and agreed that hierarhical approaches would be preferable.

RGG members reviewed Example 8 in the documentation. GM thanked PW for provided written feedback on an earlier draft of the documentation. He noted that Example 8 incoporated some of PW's suggestions, including shallowing of XML element hierarchies, changing the rioxx_version attribute to coar_version to avoid confusion, and changing accessRightsURI to access_rights to follow the established snakecase convention of Rioxx.

A lenthy discussion between RGG members ensued, particularly around the use of dc:relation and the introduction of rioxxterms:resource as a subproperty. Covered within this discussion was agreement that:

  • dc:creator should remain as a subproperty of rioxxterms:author, as in Example 8;

  • dc:identifier be used as a subproperty of rioxxterms:author instead of the proposed rioxxterms:pid;

  • the type attribute label should be re-labeled to coar_type, owing to feedback received by PK from H. Van de Sompel that type was too closely associated with MIME types and ergo could cause confusion to both humans and machines;

  • rioxxterms:resource be 'promoted' to a 'root' level property instead of as a subproperty of dc:relation, with dc:relation retained for other expressions, related items (i.e. dumb(er) relations); and,

  • that the pid attribute be re-labelled as cite_as.

  • Action: GM to prepare revised Rioxx examples incorporating the above noted adjustments and circulate for RGG refinement and eventual approval, after which the schema documentation is be finalised and also circulated for approval.

AOB

No substantive business was raised.

DONM

To be confirmed.

GM 25/05/2023