@Lydia_Pintscher, @LucasWerkmeister, just to confirm(or the opposite) what we discussed in email thread - only thing required here is the doc update?
@Cparle - this is probably of interest to SD team as well.

I think so, yes. Maybe an announcement email, though I’m not sure which mailing lists would be interested (we send Wikidata changes to wikidata-l, wikidata-tech-l, pywikibot-l, and Wikidata:Project chat).

No issue from my side :)
In the past the SDoC team decided not to yet have their work covered by the stable interface policy Wikidata follows. I guess that is a question that should be given some thought again if we want to call the RDF for MediaInfo stable now.

The doc should be updated with the new predicate introduced in T277665 first.

In T288230#7304754, @Zbyszko wrote:

@Cparle - this is probably of interest to SD team as well.

I have no objections, but I do have some questions:

• The stable interface doc makes specific reference to wikidata - for example "The change will be available for testing at least two weeks before deployment on https://test.wikidata.org/". Does the stable interface doc itself need to be updated? Do we need a https://test-commons.wikimedia.org/ going forward? Do we need to add commons mailing lists to this?
• Who is responsible for administering the stability of the interface? How is it done in wikidata? Say for example in 5 years someone wants to make a change to the MediaInfo RDF - how are they going to know that it's covered by this and announcements need to be made?

@Cparle See my previous comment :)

@Zbyszko I think the MediaInfo RDF format could be considered stable now, but we cannot guarantee anything because

1. there is no team to oversee its stability (we're the structured data team, not the commons team)
2. the stable interface policy is only for wikidata, and if it's to be used as a guarantee it either needs to be updated to include commons, or we need a new doc for commons

If we feel like our users need a guarantee then we need to have a chat about what procedures we need in place.

In T288230#7333958, @Cparle wrote:

1. there is no team to oversee its stability (we're the structured data team, not the commons team)

I think that the team owning WikibaseMediaInfo should be the team responsible for the stability of the format.

2. the stable interface policy is only for wikidata, and if it's to be used as a guarantee it either needs to be updated to include commons, or we need a new doc for commons

As I understand it the stable interface policy applies for wikibase components (even-though the page is hosted on wikidata). Relatedly commons is offering wikibase APIs (wb* actions) and is de facto benefiting from this policy. I think it makes sense to see this from a software point of view rather than a site-specific approach and thus for a wikibase extension like WikibaseMediaInfo to apply to the same policies than wikibase itself.
I have no strong objections making this separate but I believe there needs to be a guarantee that the format remains stable and that any future changes will have to be discussed.

Based on discussion with @Cparle and @MarkTraceur, we will create a new stable interface policy for Commons MediaInfo RDF specifically. I created a new subtask for that (T290671). @dcausse when do you need this done by so that it's not a blocker to ongoing WCQS work?

@CBogen I would not block WCQS work because of this, esp. because we agree on the substance that this format must be stable. I would perhaps clarify quickly (before going live) what the user documentation for the RDF format will look like, esp. because MediaInfo was added to https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format but I feel it should now be removed from this page and only maintained under its own page in https://www.mediawiki.org/wiki/Extension:WikibaseMediaInfo/RDF_mapping to avoid future confusions with the two policies.

https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format#MediaInfo still claimed the MediaInfo format was unstable, I’ve removed that note now. You might want to add a pointer to the Commons stable interface policy though.