Raw

Ian Davis questions Elliotte Harold’s prediction of 2007 being make-or-break and derives a possible time frame for the Semantic Web to "cross the chasm" of 3-9 years, based on what happened with CSS. I'm sure Ian would be the first to admit that this is only guesstimating, but I reckon he picked a good analogue.

CSS is involved in separating content from presentation on the web, but from a document point of view at least there's at least one more layer: semantics. The boundaries of these terms are blurred, but I'd suggest that most of what's commonly known as "semantic HTML" is on a level of structural syntax. Microformats do separate out certain domain-specific semantics, but don't provide a general model in which those semantics can be interpreted (and integrated), which is what RDF and the Semantic Web technologies are designed for.

The separation of semantics could be seen as a much bigger job than that of CSS. But this may be offset given that most of the prerequisites are already in place (e.g. the specs). There's also significant drive from places that aren't traditionally so involved in publication of web markup (the sciences, inside the enterprise). In many cases they are starting with data, the raw material of semantics.

So taking Hofstadter's Law into account, but balancing it with the compression of time that happens on the Internet (perhaps driven by the Law of Accelerating Returns), Ian's sticking of a pin in 5 years for chasm-crossing seems reasonable to me.