The flat “all tags are equal”-structure leaves no way to describe the relationships between tags. Unless one creates some ad-hoc-system there are no forms of qualification. A lot of context is peeled away, so I have to rely on my memory to tell if my weinberger tag signifies an item written by Weinberger or about Weinberger.

Having to deal with a lot of fields might alienate users, but for me adding a few optional facet fields would be pretty handy. I’d like to be able to do see some simple facets: A “what” field (article? Blog post? Web site?), and a related source field (if it’s a news site, which one?), a “by” field. These aren’t always relevant or interesting, but I’d like to be able to sort content from my favourite writers or sources. (Ideally, these data should be automatically retrieveded from the document itself, as we move on to more structured websites). The regular tag field could then be left for subjects and descriptive words, much like today.

As you point out in another post, each tags “weighs” just as much, which for most people will be a good thing because of the saved mental effort in having to calculate “tag weight”. Personally I’m using a primitive numeral ranking system in order to sort the more from the less important items.