Re: thing descriptions

> On 20 Jul 2015, at 16:27, Peintner, Daniel (ext) <daniel.peintner.ext@siemens.com> wrote:
> 
> Hi Dave,
> 
> I think there is some misunderstanding here. Please let me try to make my points clear.
> 
> JSON-LD is perfectly fine with me, if it fulfils the needs.
> 
> My argument is more about the actual serialization format. JSON-LD defines the syntax based on JSON and that's where I think we are looking for a "better" representation.
> 
> Let's suppose you can use an ID to define xsd:boolean for highly constrained devices. You wouldn't need the "table of strings to identifiers" anymore that seem to have caused issues in your tests.
> It is the job of the serialization format to transparently map IDs to "textual" type-names and vice-versa.
> 
> Hope these clarifies what I tried to say.

I think so.

Consider a situation where a resource constrained device wants to proxy a remote “thing”. For this the constrained device will need to get hold of the data model for that “thing” in order to construct the corresponding virtual object.  In principle, we could use content negotiation (in the sense that HTTP uses) so that the constrained device can download a more compact version of the thing’s data model in which the strings have been precompiled to identifiers. This suggests that the charter for the proposed working group should be scoped to allow for such content negotiation, and even for work on associated serialisation formats. However, as a matter of principle, I would want to avoid using the charter to dictate to the WG what technology it must use.

p.s. we are using the same principle for the charter for the propose web payment WG.

Best regards,
—
   Dave Raggett <dsr@w3.org <mailto:dsr@w3.org>>