I’m collecting lots of tags from lots of tagging services such as Flickr, del.icio.us, and Technorati, etc. I’m not just referring to the tagged items, I mean the tags themselves. This is a good thing because I am able to find my stuff and other people’s stuff that’s like my stuff easier as a result of these tags. The downside however, is that I am losing track of all my tags because there’s no one place to put them all. I have a tag cloud on Flickr that’s different to my del.icio.us bookmarks tag cloud, and so on. It’s understandably different in some ways because the objects being tagged are different, but in other ways the differences are a result of the inconsistent way I may create tags and groups of tags. I often forget what I have called something in the past and may, especially over time, change the way I refer to things, or tag them. Is this a bad thing? I don’t know yet because I probably haven’t got enough stuff for the inconsistencies to show up as being a problem. But is it a problem generally?
I’m looking for a ‘suggest a tag’ service. One that works across all other service so that when I’m trying to think of a good tag the service will suggest some for me. Del.icio.us and Flickr already do this by allowing me to choose from my existing tag and it helps me be consistent but it’s not quite what I’m after. What I’d like is some kind of service that suggest as I type what other people have tagged either the same object or what other tags are related. Something like Google Suggest, but for tags.
We could call such a tag aggregator a taggregator. Some already have, such as that described in Alan Levine’s post from a year ago but sadly the service mentioned seems to be offline, plus it only worked with Flickr and del.icio.us (the usual suspects, but between them responsible for a tiny amount of content relative to the web).
I still think that there should be an extension to the HTML spec that allows you to put tags in HTML links. That way everything gets tagged. For example:
<a href="http://www.foo.com/" tag='foo, bar'>link<a>
If every link on the web was potentially tagged then it’d be easy (especially if you were Google) to create a service to harvest these tags and create rich resource discovery applications. Presently the way to refer to other people’s tags and tagged content isn’t very efficient. For example, at the foot of this post I use 6 hypertext links to refer to ‘tagging’ etc using two different services, and those tags/links are purely arbitrary. There must be a better way.