Notes on meeting 27-abr-2026

[marcelobianchi]

Some points to take into attention from today discussion:

• We all agree that the JSON should encode objects.

• We all agree that a relationship information should be available, and that the metadata relation cannot be purely generated from the ID itself. It must be supplemented from a kind of relation information attached to the object.

• The object can be reduced to its relation alone. I.e. making a very simple example, or can have its full information inside.

• We should be able to have a ONE document as today XML approach if needed.

• While would be nice to allow to reduce the redundancy, we do not have solutions to that (yet).

Keep in mind:

• Ids should be more like URL like. In a resolvable way.

• Use of haspart and partof, to be LD consistent. this is a matter of changing the word "relationships" -> to a LD consistent word haspart-partof.

• net,station,location,channel are not uniquely resolvable, just in a context of one datacenter maybe, from a URL to point to. But we all agree that many datacenters can have a different view of the same network

  • Javier, why not enforce that the "data" part of a network to be same between datacenters while relationships know, could be different.

• Chad bring this to be considered: https://ulid.org/.

  • How does this remove the timestamp from the metadata?

[crotwell]

This json-ld best practices document may be useful:
https://w3c.github.io/json-ld-bp/