RDFa

So, after some beta-testing, the W3C Markup Validation Service is now available in a stable version. This validator checks the markup validity of Web documents in HTML, XHTML, SMIL, MathML, and: XHTML+RDFa … try it out!

In case you successfully pass the validation, you’ll be rewarded with an icon stating that you have produced valid XHTML + RDFa.
You can then choose between two versions:

or

PS: In case you need some material to play around, have a look at the RDFa Test Case repository.

What can I say…this post has everything.

First, it shows how Egon has used RDFa to mark-up HTML documents with information about molecules–something he started doing last year. But then it shows how he uses the latest version of Operator (0.8) to parse the RDFa, and generate menus which will search chemistry databases.

What’s most interesting about this post for me, is that Egon actually went through the process of doing all of this by using a GreaseMonkey script first, before he then used Operator to do the heavy lifting. GreaseMonkey is great for general purpose processing on a page, but the important thing about extensions like Operator is that they deal with your page at the semantic level. It then makes it very easy to do the type of thing that Egon has done, and add a menu item that searches a specific database. By talking us through the ‘long-hand’ technique we start to see the real benefits of Operator.

I’m looking forward to the next instalment, because I think Egon’s on a roll here!

Bruce D’Arcus announced in semantic-web@w3.org that the OASIS OpenDocument Metadata Subcommittee works on a proposal for enhanced metadata support in their file format:

A quick summary is that we’ve adopted RDF as the model for metadata, which can be:

• embedded as RDF/XML in the file package to describe either the document, or embedded content
• used in an RDFa-like syntax to tag content as triples (where the content is the literal object)
• hooked up to a new generic field (text:meta-field)

Note: For further discussions on this topic see also http://lists.oasis-open.org/archives/office/.

Cédric Mesnage has just announced a plug-in that provides Ruby on Rails programmers with an API that enables them to easily publish RDFa. More details are in his log post RDFa on Rails, the Semantic Web made simple. The RDFa on Rails project itself, is hosted at RubyForge.

A typical use would be:

<% rdfa_post url_for(post) do %>
    <% rdfa_title post.title do %>
    <h1><%= link_to post.title, {:controller => :posts, :action => :show,
        :id => post.id} %></h1>
    <% end %>
    <h4><%= rdfa_creator post.author %> the <%= rdfa_date post.created_at %></h4>
    <div><%= rdfa_post_content post.content %></div>
    <h5>Tags :
        <% for tag in post.tags %>
            <%= rdfa_link_to_tag tag.term, url_for(tag) %>
        <% end %>
    </h5>
<% end %>

Following on from the announcements that Jena supports RDFa and Firefox Plugin Operator supports RDFa this is great news, showing a rapid growth in awareness of RDFa amongst some key development communities.

Harry Chen from Image Matters LLC gave an invited lecture in Tim Finin’s semantic web class. His topic was an introduction to Geospatial Semantic Web technology. Harry discussed the problems related with unstructured geo-data and showed how to resolve some of the issues using RDFa.

RDFa started out as a project within the XHTML2 community. Oftentimes, people think that means it can’t work at all without XHTML2. Importantly, because RDFa is effectively only extra attributes in the body, you can start using RDFa right now in any version of HTML. You won’t “validate”, but you’ll still be “compliant” with the HTML specification, which says that extra attributes should just be ignored by the browser, and that’s exactly what happens.

That said, we want RDFa to be clean from a web architecture standpoint. In the words of Dan Connolly, that means that you should be able to “follow your nose” from a document that contains RDFa to a specification of RDFa. That’s why we’re building RDFa as an XHTML 1.1 module, so you’ll be able to specify that your page contains RDFa with a module declaration, and your document will validate. Importantly, the latest HTML Working Group charter of the W3C, now has an “encouragement for extensibility”:

The HTML WG is encouraged to provide a mechanism to permit
independently developed vocabularies such as
Internationalization Tag Set (ITS), Ruby, and RDFa to be mixed
into HTML documents. Whether this occurs through the
extensibility mechanism of XML, whether it is also allowed in
the classic HTML serialization, and whether it uses the DTD and
Schema modularization techniques, is for the HTML WG to
determine.

This means that, if the HTML WG does the right thing and follows this encouragement, we’ll also be able to make RDFa validate in HTML5. RDFa is, after all, just about agreeing on a few tweaks of syntax to enable interoperable structured data on the web. The really exciting stuff will be the applications, of course.

This is a Call for Proposals for RDFa utils and services. There is already an impressive amount of RDFa-software out there (see below), but still there may be a blind spot what is needed to improve one’s everyday work. What can you imagine might be worth to have? A plugin for a blog system, an editor, a semantic aggregator, etc.?

So far the following utils and services are known - at least to me:

• REST-based RDFa services, as a Elias’ extractor, or Ruben’s RDFa monkey
• DOM-based RDFa highlighter, as the RDFa Bookmarklets
• Libraries, as RDFLib
• RDFa editors, as the TopBraid Composer
• Plugins, as the MozCC plugin for Firefox, or the wordpress plugin, Daniel Lewis is working on

Please, post your whish list here, tell us about your implementation, or simply state that you are happy with all we have so far

Incidentally, there is a draft of the RDFa Test Suite available, including an alpha-version of an automated RDFa Test Runner.