Web, meet Semantic Web

By Simon St. Laurent
August 13, 2008 | Comments: 2

A common theme at this year's Balisage Conference has been integrating Semantic Web technologies with more traditional web interfaces. So far, it hasn't been the more common microformats approach, but rather approaches that let users add assertions to documents in the course of their normal wiki, blogging, and syndication processes.

I already mentioned Murray Altheim's use of wikis to collect assertions in yesterday's talks, and this morning's sessions returned to similar subjects.

Sam Hunting led off one track this morning, examining creating topic maps through Drupal. Hunting's emphasis was the combination of lessons learned from his topic maps work and his blogging experience: "The blogosphere influenced me. I'm out there every day getting hits, watching my memes propagate."

The key point of Hunting's experience, which emphasized letting users do what they wanted to do, valid or not valid, was that "People really do care about tagging - they really do tag - when they get an immediate positive result." The key phrase there is "immediate positive result." Hunting showed examples of the kinds of features that users could add easily if they were willing to take the time to add some Topic Maps markup to their documents.

There was also an interesting contrast between Hunting's approach and Altheim's approach of the day before. While Altheim used a very flexible wiki approach, letting users do what they'd like to do with the markup, while insisting that the key content had to match his structural expectations. Hunting, working in the more structured Drupal, was a lot calmer about users damaging things: "They're going to anyhow.. if they want to load bad data into their database, they can."

Alex Milowski stepped a little further back from the users, working with Atom feed categories to generate queryable and processable RDF. Starting from a messy desk that combined ordered lists, somewhat ordered stacks, and no longer ordered piles, Milowski described his path toward establishing stacks and lists out of what often looks like an overwhelming pile.

While Milowski noted that his Atomojo work isn't aimed exactly at the average blogger. There's a Firefox plugin, a server, some creative use of RDF and SPARQL... Nonetheless, his work can fit neatly with pretty much any system that's generating Atom feeds and using the atom:category element.

Unlike Hunting and Altheim, however, he doesn't expect feeds to be neatly categorized with RDF directly - as he put it, he "likes RDF, but never liked its intrusion into markup." The atom:category tag has just enough content in its scheme and term attributes for him to assemble RDF out of it that he can usefully query it. (He noted that XQuery could have done similar processing, but that wasn't his chosen path.)

Milowski's processing felt to me to be a step above simple tag aggregation, but still - from what I saw this morning - a little less ambitious than the others. Somehow, I suspect, that balance is coming to something useful. As Milowski described it:

"Specialized vocabularies still have a heavy cost to them... Using a simple vocabulary and a simple term mechanism, I and others can 'get organized'. Finally!"

You might also be interested in:



Thanks for the kind words. I should add that my Drupal topic map application does indeed allow users to add "bad data" -- for some definition of "bad" -- but only at administrator option.

The administrator can also impose (like Altheim) the requirement that only "good" data can be added, where "good" means checked by one of the several validation plugins that ship with the distribution, or by custom plugins authored to meet the site's requirements for topic map creation and content management.

All this is permitted by the tremendous modularity and flexibility of the Drupal platform, which is a joy to administer.


Great coverage throughout the conference. Speaking for those of us who would have liked to attend but couldn't, thanks.


Popular Topics


Or, visit our complete archives.

Recommended for You

Got a Question?