W3C home > Mailing lists > Public > public-rdf-wg@w3.org > February 2011

Re: [JSON] Initial comments

From: Thomas Steiner <tomac@google.com>
Date: Fri, 25 Feb 2011 10:25:30 +0100
Message-ID: <AANLkTiktQpJyQksa3MWpTFbJic479qz1CQg3mW3498=a@mail.gmail.com>
To: Ivan Herman <ivan@w3.org>
Cc: nathan@webr3.org, Pierre-Antoine Champin <pierre-antoine.champin@liris.cnrs.fr>, RDF WG <public-rdf-wg@w3.org>
Hi all,

Randomously responding to different people, sorry if this should be
confusing, but trying to be clear whom I address.

>> Does anybody actually want to write RDF, by hand, in JSON?
@Nathan: Yes, I'd definitively say so! I can see the workflow
1) read triples from RDFa
2) do awesome stuff
3) store/modify them in/for local/remote storage
For 3) all JavaScripters I know would simply build the (RDF) object
literally in code. Note that I treat JavaScript literal objects and
JSON serializations somewhat equivalent here.

> There are multiple ways to approach the subject side of things
>  { "id": 123412, "foo": "bar" }
> where you also provide a map to say you construct the subject with a
> template "http://example.org/users/@id"
@Nathan: Not sure I understand the map thing. Looking at the example
further down in your email, I guess if I get it correctly, that then I
don't really like it that much at the first sight. I'm pretty much in
favor of having one format to rule all use cases, no need for goggles,
maps, or human/robots versions.

> You don't actually need namespace/prefix/CURIE functionality, there's no
> reason why a property name can't have a colon in it..
>  { "ex:name": "mr bob", "foaf:name": "bob" }
@Nathan: Of course properties may have a colon in their name, however,
you kill what makes JavaScript so nice: the dot syntax. You can no
longer write obj.foaf:name, you have to write obj['foaf:name']. The
JavaScripters I talked to hate this.

> strings :p
@Nathan: LOL, I knew something was missing, but was too lazy to check
what it was.

[email switch]

> Thomas: I have heard that argument against RDF already: too many syntaxes
> confuse the potential users. This has been used to argue against the
> endorsement of Turtle by the WG. Although I see the point of this argument,
> I think the multiplicity of syntaxes emphasizes that RDF is *not* a file
> format, but a data model. So I think we can live with that -- but surely, we
> need to get better at explaining *why* we have so many syntaxes, and *when*
> to use which syntax...
@Pierre-Antoine: I have tried to illustrate with the exaggerated
pseudo-dialog in my email that I don't see this as an neglectable
point. Semantic Web newcomers have at least to deal with RDFa, and our
RDF JSON serialization(s). I think we should try to avoid to have the
plural 's'.

[email switch]

> So, do we focus on giving people a way to view simple objects as RDF, or
> focus on trying to get them to forget simple objects and work with RDF via
> APIs, or try and provide RDF in such a way that you don't always need APIs
> and can work with it as if it's objects?
@Nathan: in my humble opinion we should try for the latter: make our
RDF JSON so easy that people can work with it directly, without the
need for an API or simple objects (which is somewhat a consequence of
an API). I think the abstractions always come from the (external)
communities, not because the standards suck, but because people love
abstractions. I emphasize the plural 's' here, because I think there
is space for more than just one abstraction on top of our RDF
serialization (each JavaScript library might come up with a slightly
different one, for doing RDF the jQuery way, or the Prototype way

[email switch]

> not afaik, the closest I've seen is people writing javascript objects then
> JSON encoding them, but that's only for simple demo data / examples. Perhaps
> better said as, if they do, it's a minority barely worth considering, and if
> they want to write RDF by hand, point them at turtle. (imo of course)
@Nathan: Agreed to the prior: JavaScript literal objects, then
converting them to JSON (almost the same, as stated above in many
cases). Many do! Most API publishers do. I don't think it's a
neglectable group.

We got many, many opinions on the topic, which is a good thing. Maybe
even worth a F2F where we agree on something that makes most people
happy enough to continue to work on the topic. I'm happy to be
convinced :-) Let's get the discussion going...


Thomas Steiner, Research Scientist, Google Inc.
http://blog.tomayac.com, http://twitter.com/tomayac
Received on Friday, 25 February 2011 09:26:20 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:04:02 UTC