Re: Proof: Linked Data does not require RDF

On 06/17/2013 08:11 PM, Manu Sporny wrote:
> On 06/17/2013 06:21 PM, Luca Matteis wrote:
>> This still doesn't answer my initial question "How do you produce
>> Linked Data without RDF?".
> Here's the first way (plain 'ol JSON object):
> {
>    "id": "",
>    "type": "",
>    "name": "Luca Matteis"
> }
> The document above is interpreted as Linked Data using the following rules:
> 1. The thing you're talking about is identified via 'id'.
> 2. The type of the thing you're talking about is identified via 'type'.
> 3. All keys, except for 'id' and 'type', are appended to 'type''s value,
>     with a '/' separator.
> That's Linked Data. It has no formal relationship to RDF.

No, it is data that is linked.  I would not consider it Linked Data (the 
term of art) because there is no standards-based way to interpret it as 
RDF.  You are using private knowledge to interpret its meaning.

Bear in mind a document does not have to *look* (overtly) like RDF to 
*be* RDF -- i.e., to be standards-based interpretable as RDF.  Arbitrary 
XML documents that use GRDDL are a good example.  If there were a 
standards-based equivalent of GRDDL for the above plain old JSON, then 
IMO it *would* qualify as Linked Data (assuming the URIs are 
dereferenceable to more Linked Data) *because* it could be interpreted 
via standards as RDF.

> Here's the second way (Microdata):
> The base Microdata spec has nothing to do with RDF. However, if you are
> going to assert that RDFa is Linked Data, then you have to assert that
> Microdata is Linked Data. Since Microdata has no linkage to RDF, what
> are you left with? Answer: A syntax that allows you to express Linked
> Data without using RDF in any way.

I don't know the status of that spec, but if there is a standards-based 
way to interpret it as RDF, then it qualifies as Linked Data.  If not, 
it doesn't.  It is as simple as that.  (Well, assuming it also makes 
URIs dereferenceable to other Linked Data, etc.)

> Here's the third way (RFC-5988: Web Linking):
> The Web Linking RFC defines a typed connection between two resources
> that are identified by Internationalised Resource Identifiers (IRIs) and
> is comprised of:
>     o  A context IRI,
>     o  a link relation type (Section 4),
>     o  a target IRI, and
>     o  optionally, target attributes.
> Isn't this Linked Data as well?

If it expresses an RDF triple (and if the URIs are dereferenceable to 
more Linked Data), then yes.  Does it?  I'm not sure.  In section 3 I see:
    A link can be viewed as a statement of the form "{context IRI} has a
    {relation type} resource at {target IRI}, which has {target
So although that looks somewhat RDF-ish, it is not clear exactly what 
RDF it should represent.  So it looks to me like it is data that is 
linked, but not Linked Data.

> Here's a fourth way (HAL - Hypertext Application Language):
> """
> HAL provides a set of conventions for expressing hyperlinks to, and
> embeddedness of, related resources - the rest of a HAL document is just
> plain old JSON or XML. Instead of using linkless JSON/XML, or spending
> time developing a custom media type, you can just use HAL and focus on
> defining and documenting the link relations that direct your clients
> through your API. HAL is a bit like HTML for machines, in that it is
> generic and designed to drive many different types of application.
> """
> HAL is also Linked Data.
> To assert that Linked Data requires RDF requires you to make compelling
> arguments against at least these four pieces of evidence.

They are all data that is linked, but I (and apparently most others in 
the Semantic Web context) would not consider them Linked Data, because 
IMO Linked Data should support the goal of the Semantic Web, and as I 
explained at
that requires RDF, because RDF is the chosen standard universal 
information model for the Semantic Web.


Received on Tuesday, 18 June 2013 04:44:51 UTC