I bet you knew this already, but still.

Important thing about the semantic web being lowercased, is that there were two groups of people dealing with it: W3C comittee members, and a mere web developers.

First were saying: “HTML does not contain any semantic structure in it, only the markup, so let’s create a parallel universe, where the real semantic structure for every web site is encoded in the corresponding RDF”.

Second answered: “It is only part of the truth, that HTML does not contain any semantic structure; indeed it was developed with the structure in mind, while the CSS was responsible for how things looked.”

The meaning here is that when a mere web developer uses ul/li tags combination this is almost always a list. And when you have a CSS class with name ‘address’ assigned to it, the list is almost always intended to represent an address. So, on one hand you may define the CSS version of the ‘address’ class to render HTML representation of address as you pleased, and on another you may extract (tada!) semantics from the ul/li combination (namely, the value of someone’s address). This is what microformats (or lowercased semantic web, but NOT the Google’s search) are all about.

The Semantic Web is a thing that got too much publicity. Along with semantic nets, frame hierarchies, and description logics, it tried to deal with the problem of knowledge representation and processing. (Up to date) it was unable to solve any major real-world problem, but it’s OK for ongoing research initiative! As well as it is OK for a simple technical solution to the complex problem to work almost always (I am talking about semantic web/microformats now).