{quote:MANOWAR^}
 (...) just for simple document it would take way more processing power than the finished product would be really worth.
{quote}

Was he trying to extract relevant words on a text or something? It's because the RDF generation isn't that heavy, snipsnap for example does it and even though it's not optimized it's fast. 

My belief is that normal generation of XML with chunks of date won't really take much time, most of the time the biggest problem is that the chunks are just useless.. But I could be wrong.

{quote:MANOWAR^}
At the moment I am more inclined to believe that customer tracking/history is where we will be headed for a long while (...)
{quote}

Indeed that's why, for example, some of my google's cookie only expire on 2038, they want to keep a record of who I am, a click-stream...so they can learn about me. It might be handy, yet spooky!