W3C home > Mailing lists > Public > public-html@w3.org > December 2009

Re: ISSUE-76: If we fixed namespaces, does RDFa still have problems?

From: Henri Sivonen <hsivonen@iki.fi>
Date: Tue, 15 Dec 2009 17:20:32 +0200
Cc: "public-html@w3.org" <public-html@w3.org>
Message-Id: <61F166FE-07DA-4A26-B4EC-5AFD30289C48@iki.fi>
To: Manu Sporny <msporny@digitalbazaar.com>
On Dec 14, 2009, at 03:29, Manu Sporny wrote:

> Microdata:
> 
> <div itemscope>
> <p>My name is <span itemprop="name">Aryeh Gregor</span>.</p>
> </div>
> 
> RDFa:
> 
> <div about="#me" vocab="myvocab.html">
> <p>My name is <span property="name">Aryeh Gregor</span>.</p>
> </div>
> 
> or
> 
> <div about="#me" xmlns:myvoc="http://ficticious.url/vocab#">
> <p>My name is <span property="myvoc:name">Aryeh Gregor</span>.</p>
> </div>

Are these RDFa examples complete without an id="me" somewhere? That is, is baseuri#me supposed to be an imaginary URL that doesn't dereference to a node but is just talked about or is it supposed to point to a node?

On Dec 11, 2009, at 08:27, Manu Sporny wrote:

> RDFa 1.1 will allow for the use of RDFa without declaring any namespaces
> and using vocabularies and keywords instead. For example:
> 
> <p vocab="http://example.org/foaf.html" about="#robert" typeof="Person">
>   My name is <span property="name">Robert Ennals</span>.
> </p>

Can sites host their own copy of foaf.html to avoid failure in communication when example.org is down?

>> ·         RDFa doesn’t have a DOM API – but that can be added
> 
> It will be added in RDFa 1.1, which will hopefully be at REC by Dec 2010.

This is an interesting development, considering the previous position of RDFa not introducing any processing requirements to browsers.

> Here's the short list of things I've heard:
> 
> * Use of xmlns: is confusing for authors
>  (being addressed in RDFa 1.1)
> * "prefix:value" syntax in attribute values is bad
>  (being addressed in RDFa 1.1)
> * Markup rules are "too complex for authors"
>  (being addressed in RDFa 1.1)
> * Data model is "not simple enough"
> * RDF has failed, let's move on
> * Not allowing full URIs in @about, @rel/@rev, @property, etc.
>  (being addressed in RDFa 1.1)
> * Isn't tightly integrated with HTML5 (<time>, <cite>, <title>, etc.)
> * There is no browser API
>  (being addressed in RDFa 1.1)
> * Namespaces are too confusing for authors
>  (being addressed in RDFa 1.1)
> * Indirection is too confusing for authors
> 
> Any others that I missed?

 * Generating more than one triple per element is confusing.
 * Conflict with existing rel token processing (due to not requiring Safe-CURIE there).
 * Relying on namespace mapping context from the underlying layer is a layering violation anti-pattern (see QNames-in-content even though CURIEs are not QNames).
 * Mappings based on inheriting along the tree are brittle under DOM mutations. (Yes, @lang is similarly brittle.)
 * Declarations away from point of use make mappings brittle under copy and paste. (Yes, @lang is similarly brittle, but dropping @lang is has less severe effects than dropping namespace declarations.)
 * Designing a language by overlaying an RDF graph over HTML as opposed to starting with use cases is backwards and leads to a more complex language and to the use of RDF jargon in attribute names.
 * Explicit datatyping is an inconvenience when the datatype could be specified by the vocabulary that the property is part of.
 * Namespaces are a complication for implementors.
 * Namespaces are confusing to definers of RDFa themselves.
 * The processing model is not well defined in terms of the Infoset or DOM Level 2.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Tuesday, 15 December 2009 15:21:14 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:55 UTC