W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2004

Re: Is promoting RDF+XML a lost cause?

From: Arjohn Kampman <arjohn.kampman@aduna.biz>
Date: Tue, 23 Nov 2004 20:39:38 +0100
Message-ID: <41A391FA.1060204@aduna.biz>
To: Phil Dawes <pdawes@users.sf.net>
Cc: www-rdf-interest@w3.org

Phil Dawes wrote:
> Hi All,
> 
> I've been reading a lot of XML vs RDF heat recently, and am thinking
> that we've got a bit of a unsurmountable problem when it comes to XML.
> 
> I'm arriving at the opinion that we'll never be able to convince the
> majority of developers and hackers to use RDF/XML instead of XML. It's
> just too complicated, even in a cut down form. I suspect that even a
> striped XML format is too confusing (the team I work for had problems
> with this, and they're bright people).
> 
> Why? I think it's because the RDF model isn't obvious in the
> serialisation.
> 
> The XML infoset is palatable because it corresponds to what people see
> when they read XML - i.e. a tree with attributes. 
> Unfortunately people don't see triples (or a graph) when looking at
> RDF/XML - they see a tree, with additional nasty RDF syntax.

Add to this that XML does not appear to able to encode all Unicode
characters. See also the following posting from 2 months ago:
http://lists.w3.org/Archives/Public/www-rdf-interest/2004Sep/0183.html

> I'm not sure what the solution is. Ideally we'd be pushing a simpler,
> terse, more graph-friendly syntax (maybe a cut down version of
> turtle). The problem of course is that most developers hearts and
> minds are already committed to XML for data interchange.

IMHO, XML is a great format for communicating units of "hierachical"
data. It's usefulness for encoding RDF is doubtful. A much appreciated
property of XML is its easy processability. Main reason for this is the
availability of tools like parsers, data structures (DOM) and query
engines (e.g. XSLT) which allows one to skip the low-level details. A
requirement for most of these tools to be useful is a certain amount of
regularity, for example in the form of a DTD. RDF/XML lacks all of this:
the data is not hierarchical in nature but graph-like and it's
serialization in XML is far from regular.

> Maybe pushing turtle more is a good idea. What do people think?

I like Turtle: it's a simple yet convenient language. It might need some
extensions though, most of which are already covered in the "Possible
Extensions" section of the Turtle document: support for other character
sets so that documents can be written in a platform's default charset,
and a default language tag would be nice.

Arjohn

-- 
arjohn.kampman@aduna.biz
Aduna BV - http://aduna.biz/
Prinses Julianaplein 14-b, 3817 CS Amersfoort, The Netherlands
tel. +31-(0)33-4659987  fax. +31-(0)33-4659987
Received on Tuesday, 23 November 2004 19:39:43 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:56 UTC