[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: content tagging requirements and rdf/a
Re: the micro-metadata I mentioned in the previous note, I was looking
again at the RDF/A documents for the content-tagging use case.
<http://www.w3.org/TR/xhtml-rdfa-primer/>
A good practical example that could relate to office documents is this
one of embedded vCard metadata:
<p class="contactinfo" about="http://example.org/staff/jo";>
My name is
<meta property="contact:fn">
Jo Smith
</meta>.
I'm a
<meta property="contact:title">
distinguished web engineer
</meta>
at
<a rel="contact:org" href="http://example.org";>
Example.org
</a>.
You can contact me
<a rel="contact:email" href="mailto:jo@example.org";>
via email
</a>.
</p>
If you extract this as a set of triple statements, you get:
<http://example.org/staff/jo>
contact:fn "Jo Smith";
contact:title "distinguished web engineer";
contact:org <http://example.org>;
contact:email <mailto:jo@example.org> .
This is (more-or-less) the N3 syntax for representing RDF. An untyped
resource with four properties, two of them literals, and two of them
resources (identified with URIs).
So what's to note about the mechanism for making these statements and
associating it with document content?
For the most part, they use existing element (a, p, div) to carry the
content, and a small handful of attributes to convey the semantics.
Plain literals are represented by text nodes, and linked resource
objects with an href attribute (a URI).
There's some controversy about their use of QNames in attributes, but
they did this clearly to a) make authoring easier, and b) to make
validation and extension easy.
For the longest time I struggled with imagining how to practically
implement this is ODF applications, and wondered whether it really
offered any advantages to storing the metadata separately, but the more
I think about, the more I could imagine it.
Like, I could imagine highlighting a paragraph and having a contextual
menu to enter the "about" URI, and then applying properties to spans
using styles (or maybe a dedicated "meta" tag).
There is also an example from the HXTML 2 spec (which uses RDF/A) of
the sort of extrinsic linking Rob was mentioning:
<http://www.w3.org/TR/xhtml2/mod-meta.html#s_metamodule>
<html xmlns:dc="http://purl.org/dc/elements/1.1/";>
<head>
<link about="#q1" rel="dc:source" href="urn:isbn:0140449132" />
</head>
<body>
<blockquote id="q1">
<p>
'Rodion Romanovitch! My dear friend! If you go on in this way
you will go mad, I am positive! Drink, pray, if only a few
drops!'
</p>
</blockquote>
</body>
</html>
So there the paragraph gets an id, which external metadata (in this
case the link in the head) can then reference. I think in that case I'd
do it the other way around (have the link embedded in the quote to
reference outboard metadata), but it's not that different.
What this says to me for this use case is we need to add the following
requirements:
1) need to define mechanisms to associate content with metadata
descriptions
These mechanisms may include attributes and/or elements which can
assign metadata properties or URIs to document content, or which can
serve as anchor points to make extrinsic statements about document
content. Metadata statements may be embedded in the content file or
stored externally.
Bruce
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]