I like tags. I like the ideal of being able to quickly tag a link for easy retrieval later, or being able to find what other people have tagged using the same tags. But despite the popularity of tagging and the frequency with which new tagging services pop up, I’ve been getting frustrated by the way tags just aren’t quite working. Well, not quite working for me as I’m sure they’re working just fine for many.

The trouble is, at least as it seems to me, is there’s no consistent tagging protocol. No consistent way of using tags to discover new resources. Every time a new tagging service pops up there’s a new method of assigning and using tags that has to be learnt that’s just a little different form the existing ways.

Take three examples, Technorati, Flickr and del.icio.us. It’s quite understandable that each of these highly popular services has their own proprietary way of assigning tags, that’s just down to user interface issues. But it’s using each service’s tags and in particular the way they interoperate, or don’t as the case may be that’s causing my frustration.

In each of these services there exists resources that are tagged with ‘wordpress’. How do you find these resources? Well you could specify a URL that contains the tag e.g

http://www.technorati.com/tags/wordpress

http://www.flickr.com/photos/tags/wordpress

http://del.icio.us/tag/wordpress

Now for starters each uses a different URL structure so you have to know how each service constructs its URLs to retrieve their tagged content. A bit of a pain if you want to create a new service that aggregates tagged resources from each of these popular existing social services. Plus if a fourth or fifth or nth new tagging services comes along chances are they’d use a different method too. Pity. There are other ways of accessing tagged content via these services such as using XML-RPC or SOAP but the methods are again different.

Because of these differences it’s really difficult to scale tagging as a human activity and to aggregate tagged content. Is this because this is still a new activity? Probably not as tagged content has been circulating for some time now. What’s needed I think is for some standardized ways of specifying and using tags. That way I predict we’ll quickly get much richer applications with novel applications.

As tags are most frequently applied to materials that are also defined by a URL, would it be realistic to incorporate tags into the URL link markup itself? I wonder. It could be inferred from current tagging activity that content itself isn’t inherently taggable, at least not by the owner, that is to say tags are not in the metadata to the content itself. Only tagged links to the content are meaningful. You put up some content and I’ll tag my links to it in the most meaningful way to me, and so on for the links that other people create. Now if the tags were specified in the links themselves, it’d be a cinch to consistent extract tag information about any resource.

For example:

<a href="http://www.foo.com" tag="wordpress">link</a>

and:

<a href="http://www.foo.com/" tag=”blogging”>link<a>

etc.

In one step everything on the web can be tagged and not just via a handful of divergent services. Search engines and any kind of application you care to think of that knows about HTML links can now extract and use tags. Of course there’s nothing stopping anyone using tags in this way now because adding a tag attribute to an HTML link won’t break browsers but unless it became part of the HTML spec or at least more widely used then it’d be pointless.

Hmm. Would that work? Isn’t that something like what the Semantic Web guys were trying to do?

Comments are closed.