2019-11-04 Technical WG Agenda and Notes

revisiting new inflection, ?info, as landing page – new document summarizing discussion so far

Proposed: for any ARK X, X?info should lead to an HTML-formatted "landing" document (page) with metadata embedded as JSON-LD. The metadata, in human- and machine-readable form, includes

1. The ARK X
2. Descriptive metadata:
   1. who
   2. what
   3. when
   4. where
   5. how (metatype, similar to resourcetype)
   6. domain-specific elements (eg, publications vs physical samples vs vocabulary terms)
3. PIDs to first-level variants (versions, formats, change history) and components of X, if any
4. PID to the first logical ancestor of X
   1. eg, if X is a PDF variant of a document object, this points to the logical object ARK listing X along with its sibling HTML and MSWord forms
5. PID to the last logical ancestor of X
   1. eg, if X is a section of a chapter of a book, this points to the book logical object
6. Change history, if any
7. Licensing and accessibility information
8. How to cite, including "cite-as" header
9. Persistence statement

A great example to follow would be the A data citation roadmap for scholarly data repositories.

JK gave a walk through of the proposal sketch

TC: there's a distinction to be made between aggregation (UML sense) vs containment (ARK sense)

GJ: overall like the proposal, especially in providing a list of defined slots to populate that are mostly optional; suspect that most people won't have all the fields
GJ: the citation roadmap paper doesn't commit to metadata; perhaps we should recommend something? this proposal differs from others in that the ERC recommendation doesn't distinguish Publisher, while the publishing community recommendation does distinguish Publisher
JK: the Publisher distinction can be made in the domain-specific core elements that follow the kernel/ERC elements
GJ: should have a "how to cite" element
SM: yes, it would be very useful to have citation information
SM: how do we thread the needle between what we want and making things work with standards?
JK: hopefully we can use something like yamz.net to move things forward
SM: recommend to be loose enough in language we use about metadata and not let the ARK spec get bogged down in endless metadata discussions
TC: what if I only want the JSON and not the HTML, eg, what if I use the Accept-Header to ask for JSON-LD? that shouldn't return HTML + JSON, right?
GJ: there may already be standards and conventions on how to do this – I will investigate