Devastating critique of the FOAF-ian echo chamber. Even if he weren't right (and he is) and there were some adequate ontology to express the nuance of human interactions, there's this slight knowledge elicitation problem.
Memo to readers: Do not bother sending me 'semantic web' plans that reprise the delusions of big symbolic AI, having only swapped the LISP costume for new XML clothes. On the other hand, if you want to spend your own time and money hacking FOAF, I'm all for it. Train wrecks can be highly instructive. Just try making some fresh new mistakes, rather than repeating the old ones.