A came across a Jeffrey Zeldman post about tag clouds (via something that was via Kevin Marks).

I agree with many of the comments that others have left, the gist of which being that Jeffrey is wrong to assert that tag clouds are supposed to replace taxonomies.

I do also agree with Jeffrey though that the tags aren’t nearly as interesting in isolation as they would be if some sort of taxonomy of tags could be created.

Since tags are user defined, the taxonomy would also have to evolve on its own.

I wonder if it might be possible to do this by looking at two things.

1) the relationship between one tagged item, and another linked tagged item

2) items which have multiple tags

Both of these situations suggest some kind of linkage between tags. I’m not quite sure whether it’s enough, but it would be interesting to see what could be done by looking at these linkages.

Reminds me of an article I read in Scientific American about using google searches as a way of inferring the semantic ‘connectedness’ of words.

I wonder whether it would also be possible to look at the relationships between the people who use tags, the tags that they commonly use. Perhaps people who are at the centre of hot topics of conversation would somehow allow some of that heat to rub off on other tags that they use.

That might be a way to combat the problem of obscure tags withering and dying. The analogy would be - “well, that’s an obscure topic, but Ms X is talking about it and I like what she normally says so maybe I should take a look”.