There is a great article from Adam Bosworth in ACM Queue online magazine titled Learning from THE WEB. He makes many great points. I agree with most of them.
The wisdom of the crowds works amazingly well. He mentions examples of how Flickr and del.icio.us allow people to tag resources freely. We once tried to pre-define a set of tags in a collaborative portal. That feature was rarely used. It is a top down approach and did not work well. The bottom up approach of free tagging works well. Blog search engines like Technorati can use tags to locate blogs.
He then goes on to say why Semantic Web may not work well but does little to explain why. I am not sure I agree with his assessment. RDF, a language for describing resources can enrich search engines like Google and Yahoo. Have you ever tried Google alerts or searches with terms like ESB and RDF? Google can certainly do better with some contextual information or help from RDF.
I think RDF is where XML was about 5 years ago. Since XML is just a syntax for describing data, we need something to describe the relationships in a flexible manner. That is what RDF does. However, it does not follow the KISS rule. The RDF is serialized in XML format and that makes RDF statements hard to read. Some of the alternate serialization formats like N3 make it a bit simpler. If you really want to spend a few minutes to understand RDF try Joshua Tauber’s Quick Intro.
The article covers several topics including growing complexity of XML. He urges database vendors to learn from the lessons of the web and step up to the plate.
Overall, it is a great article and definitely worth reading and thinking about.