W3C home > Mailing lists > Public > www-rdf-logic@w3.org > May 2003

RE: Alternative RDF/XML serializations

From: Jimmy Cerra <jimbobbs@hotmail.com>
Date: Wed, 28 May 2003 18:52:59 -0400
To: "'Pierre-Antoine CHAMPIN'" <rdf18@lisiperso3.univ-lyon1.fr>, <www-rdf-logic@w3.org>
Message-ID: <000001c3256b$e32e0c00$1a6afea9@picard>

> I like your proposition, since it does not make such a mixing : the r:
> namespace only contains keywords, while the standard resource names
can
> still belong to the usual rdf: namespace. By the way, in your first
> example, r:type should be rdf:type...

Ya I agree. It is annoying the semantic syntax and structural syntax of
RDF are in the same namespace; although, I admit I also messed up in
that regard (type was in the RDF namespace).

> I also like the idea of using qnames for identifying resources
> themselves (and not only classes or properties).

Using qnames simply seems more natural; that's why I proposed it (I
doubt I'm the first to do so, however).

The only problem with using qnames (in XML 1.0 at least) is the fact
that they have an annoyingly limited set of allowed characters and not
common escaping method.  Although an URI only uses an even more
restricted character set (a subset of ASCII, I think), there is a
standard escaping sequence (and an URI also allows characters, like the
'%' sign, that are illegal in XML 1.0).  For instance, say I want to
name a resource 'Hello world'.  That is impossible using qnames;
however, it can be encoded into an URI as 'Hello%20World'.  So a syntax
(I used one similar to Tim Bray's RPV) has to be added to accommodate
these cases.  Not a big deal, but something that has to be considered
netherless.

Since three of us (you me and Dick McCullough) like the idea, perhaps we
should brainstorm up some formal language or schema and then submit the
alternative XML serialization to the W3C for consideration?  There ought
to be more people would benefit from an alternate serialization.

--
Jimmy Cerra

] "I have learned these days, never to limit
]  anyone else due to my own limited
]  imagination." - Dr. Mae C. Jemison


> -----Original Message-----
> From: Pierre-Antoine CHAMPIN [mailto:rdf18@lisiperso3.univ-lyon1.fr]
> Sent: Wednesday, May 28, 2003 3:56 AM
> To: jimbobbs@hotmail.com
> Cc: www-rdf-logic@w3.org
> Subject: Re: Alternative RDF/XML serializations
> 
> One of the most diffult things in the 'official' RDF/XML syntax is
that
> the rdf: namespace mixes language keywords (like Descriprion, about,
> parseType...) and resource identifiers (like Property, type, Bag...),
> which is a mess.
> 
> I like your proposition, since it does not make such a mixing : the r:
> namespace only contains keywords, while the standard resource names
can
> still belong to the usual rdf: namespace. By the way, in your first
> example, r:type should be rdf:type...
> 
> I also like the idea of using qnames for identifying resources
> themselves (and not only classes or properties).
> 
>   Pierre-Antoine
Received on Wednesday, 28 May 2003 18:53:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:46 GMT