The Semantic Web

Two good articles from Kendall Grant Clark (first article, second article) on Web Ontology or the “Semantic Web”: a proposal for a systematized markup that would make Web content structured and machine-readable (and object-oriented, but that’s another discussion).

Steven Johnson in Emergence notes — correctly — that the Web, as it is, is undifferentiated, and, insofar as organization goes, kind of stupid. Though search engines make somewhat usable maps of the landscape, and though TCP/IP quite elegantly allows machines to send information to one another, there is really no innate knowledge structure here. An ant knows what ants are around it, and ant colonies have a remarkable systemic sophistication as a result. But a Web site doesn’t have a granular, bottom-up way to understand the data or connections around it or related to it.

There are all sorts of solutions that humans have implemented in order to develop our own understanding of these connections — clickstream reports, referrer reports, log rollups, as well as search engines — and there’s software like Alexa and TrackBack which assemble aggregate traffic and connection data. But OWL would blow this all wide open, expanding the connection categories way beyond traffic trends and string matches. It would create a formal knowledge structure for the Web, would allow the automated processing of structured Web data, and would inevitably facilitate the emergence of new and immensely useful patterns and structures from the Web.