Re: [xml-dev] RDF syntax

[Dave Beckett <dave.beckett@bristol.ac.uk>]

On Sun, 02 Nov 2003 21:00:52 +0200
Toni Uusitalo <toni.uusitalo@luukku.com> wrote:

> I read RDF primer stuff, at least first part of warning commonly
> printed on instruction manuals: "you must READ and UNDERSTAND the
> instruction manual" is now true in my case when it comes to RDF ;)
> 
> But I was shocked when I started browsing through RDF syntax spec 
> http://www.w3.org/TR/2003/WD-rdf-syntax-grammar-20030905/

That's the doc I edit about RDF/XML.

> and came to
> 2.5 Property Attributes
> where it says:
> "When a property element's content is string literal, it may be
> possible to use it as an XML attribute on the containing node element.
> This can be done for multiple properties on the same node element only
> if the property element name is not repeated "...

Yes, I wrote that too.

> compared to section 2.4 Empty Property Elements' Example 5
> what's gained with this "abbreviation" but maybe that "the most
> horrible markup usage award" or something?!

"horrible" is a form of emotional response to XML rather
than a technical critcism.   I find qnames used in attribute values
a disaster, but it seems "good" XML style for several XML syntaxes.

I scare-quote "good" also since that is also a value-laden measure of
XML style.

> I'm not saying RDF hasn't got a future since I'm not really 
> using/understanding it fully yet but isn't the syntax just too
> liberal?

Yes, the RDF/XML syntax has too many abbreviated forms.
So the obvious answer is to not use them all.   

Personally I'd say there is no need to use property
attributes - stick with just property elements form.  It's got other
advantages too, such as being able to write down human languages
(xml:lang) and datatypes (rdf:datatype) on the property elements.

Dave