Add compact forms of language literals about json-ld.org

[lanthaler]

Feedback:

Thomas Steiner: +1 for no. 1
Personally I'm also a big friend of HTTP headers (remember the @context discussion?) to keep the actual payload clean, so I could also imagine an HTTP header like "Content-Language: en".
The overall opinion with @context was that HTTP headers are not considered the most suitable solution for a number of reasons.

Ivan Herman: +1 for no. 1

Gregg Kellog: +1 for no. 1, +0 for 1B, -1 for no 2

Markus Lanthaler: +1 for no. 1
+1 but perhaps we should move @language into the context description. The "meaning" of the document is described in the context so I think that's the place to be for the @language tag as well.

[lanthaler]

Ivan Herman:
I would also feel it more natural if this was part of the @context. It would also make it clear/clean that children of a node would inherit the language setting, just like any other contextual settings.

Thomas Steiner:
Oh, I instinctively assumed it would be part of the @context. -1 on any proposition different from that. +1 for putting @language in the @context.

Alexandre Passant:
+1 for having it in the context.
And for also supporting separately in the headers - using Content-Language.

[gkellogg]

+1, I had mistakenly read the example as putting it @context, this is where it should go,

Also, as with other stuff in @context, it applies through all sub-nodes, unless overridden by another definition. (@language = null might clear a setting.

[msporny]

@language in @context is accepted. @langmap is rejected.

[msporny]

http://json-ld.org/spec/ED/json-ld-syntax/20120122/#default-language

[niklasl]

Great. Just to be clear, although it is said that "apply a particular language code to all string values", I expect it is still the case that a @language in the context does not apply to coerced values, right?

I.e. the language should apply to all string values except those which are coerced, including coerced to xsd:string (which is needed to exclude certain expressions from improperly having a language applied).

[lanthaler]

Niklas, I think we should discuss this in issue #64.