> > Mostly, relabeling `@id` as `identifier` and tucking it inside `metadata` > seems very confusing. It look more like rel="identifier" or > schema:identifier--but both of those can have multiple values and would be > used for different purposes (especially within a WP). > rel=identifier is simply a proposal at this point and schema:identifier is specific to schema.org and nowhere nearly as useful as @id for an RDF parser/client. We avoided most properties prefixed with @ at this point for two reasons: - identifier is more consistent with EPUB (dc:identifier) than @id - we minimize our usage of @ properties as much as possible, mostly because we don't necessarily want all the powerful but complex stuff that JSON-LD can handle (I'm thinking mostly about local @context definition and @graph) As I already told Ivan earlier in this thread, JSON-LD is always a balancing act between pure JSON and its RDF output, and in our case (Readium) JSON and hypermedia are clearly the priorities. Keep in mind too that the goal for RWPM is slightly different than WP: - we want maximal compatibility with the EPUB 2.x/3.x infosets, not just WP - it's meant to be a generic hypermedia format that can be used for specific publications profiles (comics, audiobooks) as well as distribution (OPDS 2.0) This could easily be revisited, we'd simply avoid mapping identifier to @id in the default context document.
Received on Wednesday, 10 January 2018 10:21:04 UTC