[Home] [By Thread] [By Date] [Recent Entries]

  • From: Dan Brickley <danbri@d...>
  • To: Paul Denning <pauld@m...>
  • Date: Tue, 1 Aug 2017 16:36:21 +0100

On 1 August 2017 at 16:14, Paul Denning <pauld@m...> wrote:
Agree JSON is fit for purpose where the coupling is tight, i.e., the _javascript_ running on the browser is typically sent from the server.  The server code and the _javascript_ sent to the browser are managed as a tightly coupled pair.  If the server needs to change, the _javascript_ (sent from the server to be run on the client side) is updated at the same time. The independent evolution (loose coupling) of the client and the server is not really an issue in this case. 

Using JSON may not be optimal for true machine-to-machine exchanges where each side needs to evolve independently (yet both side are dependent on a good specification of the exchange).

FWIW at Google we have gradually moved our emphasis from Microdata/RDFa structured data markup towards JSON-LD embedded in HTML pages. Both have strengths and weaknesses, but people seem to appreciate that the JSON(-LD) representation is more compact, and not full of terms like "itemprop" (Microdata) or "property" (RDFa)  that aren't from the problem domain. Both flavours use the same underlying (schema.org) vocabulary, since that is defined in terms of types and properties rather than being hard-bound to any particular notation's structure.


Dan


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index]


Site Map | Privacy Policy | Terms of Use | Trademarks
Free Stylus Studio XML Training:
W3C Member