W3C home > Mailing lists > Public > public-linked-json@w3.org > April 2012

A personal review of JSON-LD by Zhe Wu (of Oracle)

From: Manu Sporny <msporny@digitalbazaar.com>
Date: Fri, 06 Apr 2012 21:14:54 -0400
Message-ID: <4F7F950E.50302@digitalbazaar.com>
To: Linked JSON <public-linked-json@w3.org>
A personal, non-official review of JSON-LD by Zhe Wu of Oracle:

> Hi All,
>
> As I have got quite a few deadlines coming, I was not able to find enough time to
> perform a thorough review. However, I did manage to read through the
> Syntax document twice and here are some comments. Note these comments
> only represent my personal opinion.
>
> - Overall, the document is very well written and I can see that the authors have
>    put a lot of thought into it.
>
> - I like very much the design goals of achieving simplicity, zero edits, and one pass parsing.
>   (Compatibility is a given I guess.)
>
>   There is a Note before 2.2.2. We might want to use stronger language to mandate
>   that context definition to be at the top. If we believe it is a good thing to do, why not make
>   it a requirement?
>
> - Relationship between this document and RDF.
>  The first reference to RDF is in Appendix A. I assume it is intentional. The first reference to "TRIPLE"
>   is in Section 3.2.  After reading through the document, my understanding is that the primary goal of the document
>   is to define some convenient syntax so that RDF triples (semantics) can be extracted from JSON objects.
>   (I must confess I am biased because I have been working on RDF for quite a while :) )
>   However if this is indeed the case, then RDF concept should be properly introduced much earlier in the document.
>
> - As much as I love the "zero edits" goal/rationale, I don't think this document truly makes it possible.
>   For example, we need to identifying different subjects using different @id values.
>
>   As another example, when multiple vocabularies are referenced, it seems that the compact prefix:suffix
>   notation is the way to go. This requires not only specifying a context, but also touching the data.
>   (See the example in 4.1)
>
>   Referring to multiple vocabularies is quite typical in my opinion. After all, we want to encourage users to
>   reuse well defined terms/concepts defined in FAOF, DC, GoodRelations, SIOC, SNOMED, etc.
>
> - Is it possible to consolidate
>   "@value", "@type", "@language" with their counterparts defined in RDF-JSON (RDF 1.1 JSON Serialization)?
>
>
> Some very minor editorial comments.
>
> In 2.2.1
>   These Linked Data terms are typically collected in a context document that would look something like this:
> ==>
>   These Linked Data terms are typically collected in a context document that would look something
>    as follows. Assume this context document can be retrieved at http://json-ld.org/contexts/person.
>
>   Without the above assumption, it is hard to see how that person URI gets related to that example.
>
> In 4.13
>    The second example seems to be missing a few commas in the context definition.
>
> Thanks,
>
> Zhe

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
President/CEO - Digital Bazaar, Inc.
blog: PaySwarm Website for Developers Launched
http://digitalbazaar.com/2012/02/22/new-payswarm-alpha/
Received on Saturday, 7 April 2012 01:15:25 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:36 GMT