Anyway the highlight of this talk for me at least, was his one slide about XML which was titled “XML is like Cardboard” (an analogue of course) for which he gave the following reasons;
I thought that this was a great analogue. Pity he did not try to think of one for XML schemas which could only be some complex validation system before the cardboard box is handed over (DNA sampling of the receiver?) or for XSLT which is like an IKEA instruction manual - a hundred steps to transform this pile of bits into the thing that you wanted, only after plenty of pain and suffering.