Montreal, August, Markup

By Simon St. Laurent
August 12, 2008 | Comments: 3

A small but high-powered core of markup specialists converges on Montreal every August. A decade ago there were XML conferences everywhere - everyone wanted to know everything they could about XML. Today, I think it's fair to say that most people have slaked their thirst for XML and markup generally - but while markup may be everywhere now, there's still a huge set of possibilities yet to explore at Balisage.

As Tommie Usdin noted in her opening talk, the conference isn't calling itself an "XML" conference, but rather a "markup" conference. XML has grown old and stale, right? It's ten years old, widely partially implemented, and just kind of there. Some of its predecessors also called themselves "knowledge" conferences, but that's so broad, even in an age where markup is a key part of the Semantic Web conversation, that it doesn't have much magnetism. Is markup still cool enough to draw an audience? Since more than half the attendees are here for the first time, apparently it is. Or, as Usdin would suggest was better, it is useful.

C.M. Sperberg-McQueen picked on "extremely cool Service-Oriented Architecture" in introducing XML.com editor Kurt Cagle's talk on RESTful Services. Cagle emphasized that the web is not about processes - 'transactions', for instance, are a concept alien to the architecture of the web, working "against the fundamental grain". Instead of thinking about the web as a delivery mechanism for byte streams, it can be better thought of as a storage repository, containing resources, things you can point to and do things with.

For Kurt, XQuery is the easiest path to RESTful bliss, naturally creating a connection between XML data on the server and the expectations of clients. XML provides abstraction, while XQuery provides the aggregation, integration, and processing needed to bring together data from disparate sources. And presentation? You want HTML? JSON? XML of some flavor? It's all possible. Kurt recommends XForms (of some kind, not necessarily the W3C XForms) for an interface to the XML data, AtomPub for the RESTful foundation, and XQuery as the glue, in a mix called XRX.

Murray Altheim, the "SGML Greasemonkey" (in)famous for his work on the XHTML Modularization DTDs, came to talk about semantics and ontologies - but using an informal approach built on wikis. Altheim marveled at knowledge representation's lack of interest in epistemology since Wittgenstein, pointing to a few different alternative approaches. The Semantic Web has suffered from its foundation in a particular technical world and a failed epistemology - and that may have real effects on the New Zealand issues Altheim is working on.

Instead, Altheim called for let there be jazz, opening the creation of faceted classification through wikis. While Altheim is still working with basic subject-predicate object logic and Topic Maps, it's a pleasure to see an approach that steps back a bit from the deeply formal "we must structure this precisely" systems while still supporting explicit triples beyond just tags. As he concluded, "it's just people talking."

(I like tagging - don't get me wrong - but I've wondered for a long long time if there could be something useful between tagging and the more formal systems. Altheim gives me hope that there may be more approachable and more flexible intersections between human and computer than what I've seen of past Semantic Web systems.)

That was just the first morning - there will be more!


You might also be interested in:


3 Comments

I hope there is talk about how XQuery kind of sucks for content conversion. I am not talking about pulling pieces out to put somehwere else. I am talking about converting from one schema/dtd to another (usually with a good deal of mixed content).

typeswitch makes me feel dirty.

Rob,

That will likely be discussed either tonight or tomorrow. I think most XQuery developers would agree with you (indeed, I prefer working with XSLT WITHIN XQuery, a la eXist in order to handle the schematic mapping).

Thanks for giving a perspective for interpreting the proceedings (looking forward to that). Regarding the midpoint between tagging and formal triples, is SKOS (http://www.w3.org/TR/skos-reference/) an option? Or do we need more flexibility? Or more formality?

Popular Topics

Archives

Or, visit our complete archives.

Recommended for You

Got a Question?