Date
Attendees
Discussion items
Time | Item | Who | Notes |
---|---|---|---|
Announcements | JK | Personally identifiable information in the NAAN registry. John is looking at Github actions to automatically strip. Pre-generated 10,000 NAANs to remove need for running eggnog to mint. | |
Continued discussion of "?info" inflection proposal (latest at the end) JSON encodings/formats: JSON-LD, geoJSON?
| (JK) Proposing ? and ?? be left untouched, "reserved for future use," but ?info would be required. ?info return metadata would not be detailed in the ARK spec per previous meeting. Looking for couple volunteers over next month to help brainstorm over what ?info might return. KH, CM, GJ willing to help. General discussion on JSON, JSON-LD, etc. (TC) If intent is for machine readability, then JSON is the answer. TC's group has adopted JSON-LD extensively. (GJ) JSON-LD is the mainstream practice on the web, and Data Citation Roadmap paper (Fenner et al) recommends it. Is there any reason to do something different? (GJ) JSON-LD is implementation/serialization of RDF. Defined by W3C standard, which specifies content negotiation and application/ld+json content type. (TC) Good reference on understanding JSON-LD, schema.org, etc.: https://developers.google.com/search/docs/guides/intro-structured-data
| ||
Status check on milestones; remaining to-do list, and rough estimates for objectives and deliverables
| All deliverables delayed. Deliverables page, charter need updating. Will visit topic next meeting. |
Action items
- subgroup will meet to work further on ?info response (notes below)
- John Kunze revise deliverable schedule for next time
2019.11.26 subgroup notes: Greg, Karen, John
GJ: conneg is invisible to resolver? because repo should handle it?
GJ: is ?info handled differently?
JK: we need to define this better: when does first resolver vs last resolver
GJ: ** Which Agent handles ?info conneg request – must be defined
KH: like to have configuration option for NAAN entry to say whether to pass thru or have first resolver
GJ: geojson seems difficult place to start compared to json-ld
KH: like the idea of having required core plus optional stuff; prefer minimal json-ld; believe you can use other vocabularies just use multiple @contexts, eg, for geo and dublin kernel
KH: * should test whether json-ld extensions would break features
KH: could be recommended, with optional
KH: like the idea of: if you do anything, do this