On XML.com, Kendall Clark gives a clear and accessible review of the semantic web transition, then criticizes our own RDF metadata strategy, specifically. It’s useful and insightful feedback, so we’ve taken the time to respond at length here.
(If you’re not familiar with RDF or the semantic web, or why they’re important to our mission, read more. )
We want to get RDF out there. We want people to use it on their sites even if they don’t fully understand it. (Why should they need to?) RDF needs momentum more than perfection to become useful; we’re working to provide some of this momentum. We fully recognize Kendall’s argument that our recommendation is not the purest from a technical standpoint. It is, however, a practical approach to getting users to integrate RDF.
On our site, we use content negotiation to send RDF/XML to programs that request it and HTML to normal browsers.
But for our users, we recommend the quickest and easiest way to have RDF somewhere. We think it’s better to have RDF in HTML comments than no RDF at all.
Kendall is correct to point out that we haven’t provided enough guidance on more advanced ways to connect HTML pages to their RDF descriptions. We’ll work on that, with the help of the community (including folks like Kendall). Hopefully, we’ll be able to put together a document with the best practices for including our RDF in all formats, include various versions of HTML and RSS.
Our goal is to get RDF out there. We’re trying to maximize RDF presence by adapting to our different types of users.