Still enthusiastic about RDF


It took me a long time to “get” RDF, and I read through a good few introductions. (For well-written takes, have a look at Rahul Sing’s, Stefano Mazzocchi’s, Naveen Balani’s and the W3C RDF Primer.) Every time I read through another one, I’d pick up on a few key concepts (it’s about URIs and graphs and metadata, right?), but never quite make a practical connection.

Most introductions take the same tack – start with RDF-as-graph, and show a simple example:

page → was-written-by → user → whose-name-is → Example Name

Graphs are a critical implementation detail, and a wonderful illustration for a certain mindset, but the best summary I saw was in an article by Aaron Cope:

“RDF, by constrast, was created to help manage disparate collections of metadata across a large information-space whose boundaries are unknown or in flux.”

It’s probably still a little vague to secure that venture capital funding, but importantly it’s about the purpose, not the implementation. (“Sell the benefit, not the product.”)

I’ve been thinking about barriers to entry. There’s a great Stephen O’Grady post with a clearheaded roundup of the ways that users can be dissuaded from learning to work with a new platform. (It seems like Ruby on Rails’ success is due to a large degree to its tearing down of all possible barriers to entry.)

It seems like the barriers to RDF are still too high. Toolkits aren’t ubiquitous, and there isn't really an obviously critical mass of data to consume. So what if you went the other way? Rather than telling people to bootstrap, get a desktop application out there. Make it simple, but compelling, and base it totally around open RDF data. (Did someone already suggest this?)

On a practical level, it gets RDF onto people’s desktops. Want to check out SPARQL? Now you have something to query. Want to write a twenty-line script against RDFLib, or produce RDF that something can present to you? Go for it. Want an incentive to fix up Drive, the only lightweight RDF parser for .NET that I could find? Sure.

Does this sound disingenuously vague? Shelley’s post on marketing RDF is almost inspirational:

“This is just going to revolutionize our lives. I am not joking — the next generation of the web is here, and I’m just so excited! It is going to be big, babies! Big!”

but ends up being more thought-provoking as her nerve goes.

“must we exaggerate that passion–emphasize it so it can be seen at a distance [...] ?”

RDF as all-encompassing metadata framework seems to have, as its most immediate rival, microformats. Norman Walsh articulates something I’d considered but never put so succintly:

“Microformats rely on the existing extensibility point, the role or class attribute. As such, they consume that extensibility point, leaving me without one for any other use I may have.”

I’ve already completed some interesting work on this, really without much reference to prior art. Danny O’Brien ruminates on the perils of keeping work private: “I’m out of practice at being open.”

In conclusion:

(Music: Fugazi, “Do You Like Me”)
(More from this year, or the front page? [K])