Semantic Web at work?

by Uche Ogbuji

I'm still getting my Weblogger profile here updated, but this year I transitioned from one company I co-founded to another. Zepheira provides data architecture solutions, with a focus on semantic technology. I was early on the Semantic Web bandwagon, and I almost fell off at one point because I felt the useful, modest ideas at the core had been overrun by an academic brand of technological megalomania. This year I felt the timing was right to not only renew my interest in the technology, but to stake my livelihood on it. Part of it was timing: I was starting to see the more useful underpinnings of semantic technology take hold in corporations. Part of it was people: I found a group of professionals who I believed were capable of building practical semantic technology solutions, and, more importantly, selling them.

One of those people, Eric Miller, former chair of the W3C Semantic Web Activity, is especially well known for describing the benefits of semantic technology in terms executives can appreciate, and he's featured in a new InternetWeek article "The Semantic Web Goes to Work". The article says:

"You['d] better figure out what the Semantic Web is and soon, because its concepts have graduated from academia and are starting to contribute to your competitor's bottom line."

I'm hearing a lot of that sort of thing, lately. The pundits, having written off semantic technology as so much pipe-dreaming for so long have switched into a level of hype overdrive. The reality is that as Eric puts it in the article, a consistent, universal system of identifiers and a layer of technologies for mapping these identifiers is the sweet spot of semantic technology. Semantic Web technology is the specialization that builds these identifiers on Web technology, and in particular URIs. This opens up the benefits of REST architecture, and for me that the third pillar is the universal writing system provided by XML. These are all, individually, modest technologies. Hardly nanotech, quantum mechanics or genetic engineering. But take these three and combine them with a skilled data architect and I do believe very special things are possible. There's a large crowd of folks who still make the free association from "semantic" to "metacrap", but that presents nothing but a ripe opportunity for others who know how to keep it simple, and thus get real work done.

The article also mentions Eric's keynote at the Semantic Web Strategies conference, which is chaired by fellow XML-meets-semantic-tech pragmatist Bob DuCharme this October. I'll also be on a keynote panel, and I'll be co-presenting with Kristen Harris, long-time collaborator at Sun about how we improved content architecture for Sun's mail Web sites using Semantic technology and REST.

This conference, organized by IT industry watchers Jupitermedia, is just another indication of how seriously folks are starting to take this stuff. I almost cringe that the stampede could end up ruining the crop, but that's a test every worthwhile technology must endure at some point.