Re: Sect 5.1 of future; Recommended JSON-LD context

+1 for stable, versioned JSON-LD context documents

Versioned contexts are useful when JSON-LD annotation data structures reference a context. If the context content changes, but the JSON-LD annotation data structure hasn't been updated to point to a different URL that holds the prior context that was used when building the JSON-LD data structure, then the JSON-LD might not represent the original annotation graph/concept/intent.

So if we publish JSON-LD contexts for OA, we need to do so in a way that ensures they are stable so that people can reference them.

A stable, versioned URL for a JSON-LD context is also useful when doing initial development/testing and you don't want to incorporate a full JSON-LD context processor into the application right away. Eventually, a proper JSON-LD application will need to process the context instead of keying off of the context URL.

In some ways, the JSON-LD context can act like a namespace in XML, but it's more of a reference to a schema than a unique name. It's a mistake to treat the JSON-LD context as an XML namespace. As long as we're clear that we're not acting on the JSON-LD context URL as we would an XML namespace, I'm good with establishing well-defined, reference JSON-LD contexts for OA and what we might call OA application classes or modules.

-- Jim

On Jan 27, 2013, at 11:10 PM, Bob Morris <morris.bob@gmail.com> wrote:

> Yes, exactly.  I do note that in the Provenance section there is an
> Editor's note deferring modeling the version of OA itself.  That could
> apply to this too, but that's not clear to me.
> 
> In the case of the serialization as JSON-LD, there is a need for a
> good contract between producer and consumer as to the actual
> @context. The spec itself is adequate to that, since the @context is
> listed in the spec, but whatever is arrived at for versioning the spec
> has the chance of needing to change that, hence change its name.  A
> scenario of slightly less importance is that as the OA spec itself may
> continue evolve toward its possible state as a W3 Recommendation,
> there are likely to be changes in the recommended @context.
> 
> 
> On Sun, Jan 27, 2013 at 7:48 PM, Robert Sanderson <azaroth42@gmail.com> wrote:
>> Hi Bob,
>> 
>> Do you mean a version in the URI so that systems "know" that they can
>> use their cached copy?
>> So:  .../context-20130128.json  to distinguish from any future version?
>> 
>> Thanks!
>> 
>> Rob
>> 
>> 
>> On Sun, Jan 27, 2013 at 1:44 PM, Bob Morris <morris.bob@gmail.com> wrote:
>>> Sooner or later it might be good to assign a version to the
>>> recommended context document and/or some other way stable designation
>>> that does not depend on http dereferencing at  the time of
>>> serialization or deserialization.
>>> 
>>> Bob
>>> 
>>> --
>>> Robert A. Morris
>>> 
>>> Emeritus Professor  of Computer Science
>>> UMASS-Boston
>>> 100 Morrissey Blvd
>>> Boston, MA 02125-3390
>>> 
>>> IT Staff
>>> Filtered Push Project
>>> Harvard University Herbaria
>>> Harvard University
>>> 
>>> email: morris.bob@gmail.com
>>> web: http://efg.cs.umb.edu/
>>> web: http://wiki.filteredpush.org
>>> http://www.cs.umb.edu/~ram
>>> ===
>>> The content of this communication is made entirely on my
>>> own behalf and in no way should be deemed to express
>>> official positions of The University of Massachusetts at Boston or
>>> Harvard University.
>>> 
> 
> 
> 
> -- 
> Robert A. Morris
> 
> Emeritus Professor  of Computer Science
> UMASS-Boston
> 100 Morrissey Blvd
> Boston, MA 02125-3390
> 
> IT Staff
> Filtered Push Project
> Harvard University Herbaria
> Harvard University
> 
> email: morris.bob@gmail.com
> web: http://efg.cs.umb.edu/
> web: http://wiki.filteredpush.org
> http://www.cs.umb.edu/~ram
> ===
> The content of this communication is made entirely on my
> own behalf and in no way should be deemed to express
> official positions of The University of Massachusetts at Boston or
> Harvard University.
> 

Received on Monday, 28 January 2013 14:34:19 UTC