Should we use @container:@set to ensure arrays?

[azaroth42]

As demonstrated by @workergnome here:
https://gist.github.com/workergnome/e6e5bc219c7349ad64190b18e551e051

The trade-off is between having to test for the type and having lots of singleton arrays everywhere.

The equivalent discussion in IIIF is: IIIF/api#1131

[azaroth42]

Tagging as priority as it will break all of our JSON. We should make it before it gets even harder to change.

Thankfully the cromulent library can take care of the serialization invisibly ... but anyone with other code / pipelines might have a lot of work to do.

[workergnome]

None of my pipelines will have problems, since they're native RDF and use framing to take care of this. And the read scripts need to accommodate arrays OR objects, so they're already prepared for this. It will make it easier for everyone I the future, though.

I'm a strong yes on this.

[azaroth42]

My main misgiving is type. It seems painful to always have an array of one, other than the single case of [EndOfExistence, Activity] caused by the crazy inconsistency of CRM.

Other questions are:

• format -- if a resource has content negotiation, should it be able to have multiple formats?
• Should label and description be language maps?
  • value can't be a language map, as it can contain integers ... so LinguisticObject.value would be inconsistent with label, requiring the full {"@value":"", "@language":""} construction

I'll try and implement in crom as an optional flag and turn it on in a branch.

[azaroth42]

azaroth42 marks this issue as being discussed in slack

[azaroth42]

My first cut at a profile for whether properties should be multiples or not:

Property	Multiples?
P1_is_identified_by	yes
P1i_identifies	no
P2_has_type	yes
P4_has_time-span	no
P7_took_place_at	yes
P9_consists_of	yes
P9i_forms_part_of	yes
P12_occurred_in_the_presence_of	yes
P12i_was_present_at	yes
P14i_performed	yes
P14_carried_out_by	yes
P15i_influenced	yes
P15_was_influenced_by	yes
P16_used_specific_object	yes
P16i_was_used_for	yes
P20i_was_purpose_of	yes
P20_had_specific_purpose	yes
P22_transferred_title_to	no
P22i_acquired_title_through	yes
P23i_surrendered_title_through	yes
P23_transferred_title_from	no
P24_transferred_title_of	yes
P24i_changed_ownership_through	yes
P28i_surrendered_custody_through	yes
P28_custody_surrendered_by	no
P29i_received_custody_through	yes
P29_custody_received_by	no
P30i_custody_transferred_through	yes
P30_transferred_custody_of	yes
P32_used_general_technique	yes
P32i_was_technique_of	yes
P43i_is_dimension_of	no
P43_has_dimension	yes
P45_consists_of	yes
P45i_is_incorporated_in	yes
P46_is_composed_of	yes
P46i_forms_part_of	yes
P50i_is_current_keeper_of	yes
P50_has_current_keeper	no
P52i_is_current_owner_of	yes
P52_has_current_owner	no
P55i_currently_holds	yes
P55_has_current_location	no
P56i_is_found_on	no
P56_bears_feature	yes
P62i_is_depicted_by	yes
P62_depicts	yes
P65_shows_visual_item	yes
P65i_is_shown_by	yes
P67_refers_to	yes
P67i_is_referred_to_by	yes
P72_has_language	yes
P72i_is_language_of	yes
P75i_is_possessed_by	no
P75_possesses	yes
P81b_begin_of_the_end	no
P81a_end_of_the_begin	no
P82b_end_of_the_end	no
P82a_begin_of_the_begin	no
P89_falls_within	yes
P89i_contains	yes
P91_has_unit	no
P92i_was_brought_into_existence_by	no
P92_brought_into_existence	yes
P93_took_out_of_existence	yes
P93i_was_taken_out_of_existence_by	no
P94i_was_created_by	no
P94_has_created	yes
P95_has_formed	yes
P95i_was_formed_by	no
P104_is_subject_to	yes
P104i_applies_to	no
P105_right_held_by	no
P105i_has_right_on	yes
P106i_forms_part_of	yes
P106_is_composed_of	yes
P107i_is_current_or_former_member_of	yes
P107_has_current_or_former_member	yes
P108i_was_produced_by	no
P108_has_produced	yes
P120i_occurs_after	yes
P120_occurs_before	yes
P128_carries	yes
P128i_is_carried_by	yes
P129i_is_subject_of	yes
P129_is_about	yes
P134i_was_continued_by	yes
P134_continued	yes
P138i_has_representation	yes
P138_represents	yes
P140i_was_attributed_by	yes
P140_assigned_attribute_to	no
P141i_was_assigned_by	yes
P141_assigned	no
P143_joined	no
P143i_was_joined_by	yes
P144_joined_with	no
P144i_gained_member_by	yes
P145_separated	no
P145i_left_by	yes
P146_separated_from	no
P146i_lost_member_by	yes
P148i_is_component_of	yes
P148_has_component	yes
P180i_was_currency_of	yes
P180_has_currency	no
rdfs:label	no
dcterms:conformsTo	no
dcterms:relation	yes
pi:paid_amount	no
dcterms:subject	yes
pi:paid_to	no
pi:paid_from	no
skos:exactMatch	yes
rdf:value	no
dc:format	no
dc:description	no

[azaroth42]

Re @type as @container:@set: json-ld/json-ld.org#512

[azaroth42]

Done, and explained in the context design section.