Mathemagenic

There is a couple of follow-up discussions for Facts on archiving, search and retrieval. As usual in weblog conversations it's fragmented, so I'll do some capturing for myself. 

Anjo Anjewierden suggests that 60% overlap in classifications is not so bad

What is the problem? I would be very happy if Google and myself would overlap 60% of the time.

The rest of discussion is on what/when/why ontologies can support retrieval and a bit on mental models. 

My position (originally in the comments to Anjo's post)

I'm very much biased towards "all things distributed" - simply because mental maps of people are different. Of course, this suggests only that a centralised solution (e.g. ontology) should work if you can find a case where people are likely to have similar mental maps (I can't help thinking about workflows :)))

Andy Boyd takes another angle:

To me this is an important fact why users get so dissatisfied with stored categorised information, i.e. they do not follow/agree with how it is categorised.

One thing automation may do is give us consistency, even if we do not agree with the classification at least by the fact that it is consistent we will have the chance to learn the automata’s logic.

Me (in the comments):

I'm thinking why Google works better for most people than any hand-crafted "knowledge base"... I guess it's not only because of automatic categorisation, but also because:

- it's hidden: internal rankings and categories are not revealed, so we can't see how far from perfect they are

- we know that it's done by machines and we forgive them lack of precision and completeness that we wouldn't tolerate from a human "expert"

See also: What's Your Idea of a Mental Model?

Side note: why TypePad guys do not add comment permalinks to the default templates? why I have to dig them out manually?