Stefano writes about structuring data and when to do it. It made me think about the strong versus loose typing argument in multi-computer-language-land, and an anonymous quote I grabbed at a seminar once: "People are generally not interested in the types, just in the values."

Other (unstructured) associations:

• The way R. Pirsig describes the Platypus effect in his book Lila. Which is very nicely quoted in these rawnotes:

...that when the Platypus was discovered, scientists said it was a paradox. But Pirsig's point was it was never a paradox or an oddity. It didn't make sense only to the scientists because they viewed the nature of animals according to their own classification, when nature did not have any.

• The web usability expert articles advising us a long time ago to replace all those 'advanced search' forms with the simple one-line fulltext-search-box.
• The one-big-bag approach and the facetted browsing in Daisy of course.
• My kids.