RE: Yet another serialization format?

On 06/30/2011 1:35 AM glenn mcdonald wrote:
>> "Out-of-band annotation" as you call it doesn't
>> prevent every node to have an ID in any way.
>
> I was talking specifically about the idea that you could
> take arbitrary current non-graphy JSON and make it into 
> Linked Data by applying external annotations...

Well Linked Data means, at least for me, that you have out- and/or ingoing links to other data nodes/records. If that's not the case, it's not Linked Data. It doesn't matter how those links are technically extracted from of a (JSON) representation. The reason why I propose to look at out-of-band annotations simply lies in the nature of JSON. It is mostly used in Web APIs and is thus normally well defined and described, even though those descriptions currently are just human readable.

What are the arguments against using external descriptions? The only one I can see is one additional round-trip. Is that really so bad? Why is CSS then mostly served as external file for instance?


--
Markus Lanthaler
@markuslanthaler

Received on Thursday, 30 June 2011 10:20:58 UTC