- From: Andy Seaborne <andy.seaborne@epimorphics.com>
- Date: Wed, 06 Apr 2011 20:14:57 +0100
- To: Manu Sporny <msporny@digitalbazaar.com>
- CC: RDF Working Group <public-rdf-wg@w3.org>
On 06/04/11 01:50, Manu Sporny wrote: > I really liked Nathan's proposal a few weeks ago: > > http://lists.w3.org/Archives/Public/public-rdf-wg/2011Mar/0565.html > > Tom's serialization work is also excellent, and is a must read before > diving any further into this e-mail: > > http://www.w3.org/2011/rdf-wg/wiki/JSON-Serialization-Examples > > I'd like to see if we can come to some sort of consensus on a starting > point based on Nathan's proposal. I'm going to remove things that raised > issues w/ some people and see if we can all agree if the result could be > the starting point for the JSON work. > > Note that this proposal is imperfect by design - it is only here to > capture the things that the majority of the group seem to agree upon. > It's merely meant to put a stake in the ground so that we may start > building on top of it. If we can get agreement on these 5 principles, > then we can add on features as the group discusses them: I don't quite understand how this list of questions arises as being the key questions, or maybe just I don't know which user segments we are addressing. If it's 3/4/5+C, that segment assumes a library, and we have two forms of that: a one step specific-parser to produce a JS structure that the app can use, or a full-blown library where access is always via a library call. Either way the format on-the-wire isn't directly visible to the application. Which is this addressing? On-the-wire or JS structure? I'd answer the questions differently for these 2 cases. Andy > 1: Constrain JSON [1] to be an (optionally nested) sequence of one or > more objects (where one, no enclosing [] is needed). > > 2: constrain object keys to be strings with no white space. > > 3: add recognition for a special "@id" property who's value is an IRI > (sets the subject of the object when present). > > 4: add recognition for a special "@type" property who's value is a > simple string. The value is looked up in the @context. > > 5: Support a "@context" property that allows for a set of mappings from > JSON keys to IRIs. > > { > "@context": > { > "Person": "http://xmlns.com/0.1/foaf/Person", > "name": "http://xmlns.com/0.1/foaf/name", > }, > "@id": "http://jondoe.example.org/#me", > "@type": "Person", > "name": "Nathan Rixham" > } > > That's it - please +1 below each number if you support the general > direction of the feature. -1 if you don't, please explain if you don't. > It's been around 2 weeks, so hopefully some of us have had time to let > these ideas kick around in our heads for a while. I'll try to setup a > Doodle poll to have a discussion about this proposal later on in the > week as well as discuss some of the serialization work that Tom has done. > > -- manu >
Received on Wednesday, 6 April 2011 19:15:36 UTC