There’s an interesting article on We Break Stuff about tagging and social bookmarks. Frederico seems to be a bit sceptical about the benefits of tagging, mainly because – and rightly in my opionion – he thinks the average user is not motivated enough to use tagging in the way it needs to be used.
How does it need to be used? Well, for example, I could save this post with the tags “social” and “bookmarks” and, if you’re a geek like me, you’d know what it was about. But what if I saved the post with the tags “tags” and “benefits”, some people may think it’s something to do with care labels in clothing. Rubbish example, but you see the problem.
So, my idea is this. To use an online thesaurus web services (an API, in other words) to search for close matches for the tag being searched/looked at, that would then in turn look for items stored with those extra tags.
A simple example may be something like this:
User searches for the tag “bookmark”. The system brings back all items (these could be blog posts, links etc) associated with that tag AND also runs a thesaurus search for related tags.
Related tags returned are “clip, flag, flap, holder, logo, loop, marker, slip, sticker, stop, strip, tag, ticket” (from http://thesaurus.reference.com/search?r=2&q=bookmark) and then runs a second search on these related tags, offering the results with a relevance rank of some kind.
Obviously this wouldn’t be a foolproof system, but I think it may have advantages for some people using tags with obscure or obtuse meanings. Another idea would be to scan the text of an item looking for the related words, perhaps that would give better results.
There may also be some interesting application with regard to Dave Winers latest OPML suggestion.