W3C home > Mailing lists > Public > semantic-web@w3.org > September 2007


From: Keith Alexander <k.j.w.alexander@gmail.com>
Date: Wed, 05 Sep 2007 19:08:44 +0100
To: "Ben Adida" <ben@adida.net>
Cc: semantic-web@w3.org
Message-ID: <op.tx60oud763ayaz@keith-alexanders-computer.local>

On Wed, 05 Sep 2007 18:15:41 +0100, Ben Adida <ben@adida.net> wrote:

> Keith,
>> The structure of an RDF/JSON document looks like this...
> I like this idea.

Cheers Ben :)

> In fact, I have an ad-hoc implementation of something
> much like it in a development project,

That's good, it probably suggests  the structure is pretty sensible.  
Benjamin Nowack also commented that he had been thinking of something very  

> though not as complete in terms
> of datatype and lang support. I would suggest taking a look at
> potentially nesting the data structures, especially for bnodes which you
> don't really want to have to name. (We went through this same evolution
> with RDFa.)

Initially I tried using a nested structure, but I reverted to a flat  
structure, as it is more consistent, and you always know where to find a  
resource (on the top level). Consistency is very important here so that  
the data structure can be easily used straight away.

I can see there might be an argument for making an exception with bnodes,  
and this would be somewhat similar as SPARQL's  behaviour with DESCRIBE,  
of including related bnode resources in a resource's description. (Though  
personally, I think I probably prefer the uniformity of treating bnodes  
the same as URIs). Can you think of any use-cases where nested bnodes  
would be preferable?


Received on Wednesday, 5 September 2007 18:11:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:41:59 UTC