I consider myself a metadata geek. I write about it, I work for a company that helps manage it, I track and play with the related standards, and I have a network of friends who fall into several of these categories.
Last week I went to a conference devoted to metadata and I didn’t know a soul at this large, well-attended conference. I had heard of only one speaker, John Zachman, because a former co-worker was a big fan of his framework.
It turns out that there’s a whole metadata industry separate from the metadata world I’d been following. The world that I know is focused on describing HTTP-addressable resources (the “RD” in “RDF”), whether on the public internet or private intranets. We think of schemas as metadata, of course, but the management of huge amounts of schemas across an enterprise is not a research issue as much as the association of metadata with resources in order to get more value out of those resources.
At this conference, tracking of schemas across an enterprise was a big issue. Data Warehousing and Data Governance were the context for it all; as one attendee told me at a reception, organizations like his financial institution want to be able to know where a report’s numbers came from.
Presentation titles like How to Perform Information Stewardship within Business Process Redesign and Using Metadata to Support Compliance and Accountability Efforts give the flavor of what was on attendees' minds. I don’t think of my metadata interests as being limited to publishing applications, but if we take the word “publishing” in its broadest sense, maybe it is. The standards I use are typically about tracking the metadata associated with a resource that will be used to convey information to a user—or, to put it differently, a resource that will be published. (It’s not the first time that I’ve found another whole world of metadata separate from the kind I pay close attention to.)
In the Venn diagram showing the concerns of the HTTP/publishing metadata world and the enterprise data governance metadata world, the overlap is OWL and semantic technologies, because both groups are getting more interested. My conference talk was on the basics of RDF, and I was definitely not preaching to the converted. I found myself explaining issues that I thought we’d all moved past, such as why the URIs that represent resources may look like web addresses but aren’t necessarily.
It was particularly odd for me to see so much interest in OWL from people who weren’t that interested in RDF. I saw great coverage of both in a three-hour seminar by Dave McComb and Simon Robe on Enterprise Ontology: Designing Your Core Model Using Semantic Web Technology. Unfortunately, I only saw the first hour because I had to leave for the airport. In a tag-team format, they covered many important issues very well considering how much of the audience was learning these concepts from scratch. (And now I finally understand the difference between Tboxes and Aboxes!) As with any such presentation, much of it was about how OWL models data, and I wish I could have stayed to see how people apply these models to their world of metadata. (Now that I think of it, some of my own experiments may relate more to the enterprise data world than the publishing world.) Any suggestions on good background resources for the world of enterprise metadata?