Remaining issues for 1.1

Just a refresher on the primary changes added for JSON-LD 1.1 to date:

* Updated Framing algorithm
* Scoped Contexts,
* Node Identifier/Node Type indexing (@id maps and @type maps),
* Nested Properties

There remain several open issues tagged for 1.1, not all of which I intend to complete (some, I think, we should close). The following are those I’d like to complete:

* issue 293 “Framing shouldn’t introduce new blank node identifiers”. I agree that they’re horrible, and as framing was never really completed, entirely reasonable to address.
* issue 333 “Support JSON values that aren’t mapped” I interpret this as calling for a JSON datatype which the content used to create a literal of an appropriate datatype. It also satisfies some concerns about use in GeoJSON.
* issue 405:Prefer simple terms to complex terms when creating Compact IRIs. Who here hasn’t been bitten by this? It’s really annoying when you have things like “resource” and “resources”, which can cause a term such as “resource:s” to be genrated! I would probably go further to say only use terms who’s value is an absollute IRI, rather than a compact IRI.
* issue 435: “Provide means to get framed output without @graph”. Indeed, I always have to write code to extract this, and I’ve heard this come up many times.
* issue 446: “@version announcement for enabling 1.1 features”. I hate it, but don’t see a way to avoid it. Without explicit anouncement (either through API processingMode, or @version in context) a 1.0 processor may silently ignore these features and run anyway.
* issue 463: “abstract model …”. We’re most of the way there now, and this doesn’t affect cognitive load, IMHO.

Gregg Kellogg
gregg@greggkellogg.net

Received on Wednesday, 15 February 2017 20:09:09 UTC