The concept of tagging is very simple – to take a resource and attach words that describe it. The motivation behind the desire to tag can be varied, with the most common reasons being to capture data that is not included in the resource itself (often the case with things like photos) or to aid retrieval of resources (as with bookmarks).
Sometimes it can be desirable to capture more than a simple or compound word in a tag. Early last year, Rev Dan Catt posted on the subject of Triple Tags, tags that encapsulated a name-space along with a name-value pair. This format is probably familar to many Upcoming users, who are already used to tagging Flickr photos of their Upcoming events with an upcoming:event=12345 type of tag. Flickr seems to have now adopted this style of tagging more formally and named them machine tags. Take a look at a photo with machine tags and you can see that they’re now sectioned out into their own tag list.
Whether or not machine tags is the right name for these or not (I personally think not, as even though they’re of a fixed syntax, they are easily written and read by humans, not just machines), I think we’re going to see a lot more of them. Indeed, I’ve been using them for a few months on one of my own projects – check out the photos that belong to this Living Generously action. The great thing about services like Flickr formalising their use is that it’s now possible to search component parts of the tags via their API.
Of course, this is going to bring its own challenges for things like tag clouds. If your site currently supports tagging, expect to have to deal with this sooner or later.
For rel-tag users, you just need to remember that standard practise is to url-encode tag names to keep them legal.