Technorati tags (updated)

Time to try out the Technorati tags I think.


Update 30 Jan 05:  Like Tim Bray, I think there is something potentially really interesting about the tags concept (see David Sifry's overview).  But I'm a bit confused.  I get the concept, but it seems clumsy.  Also, it doesn't seem to work at the moment.  To be fair, Sifry does explain in his disclaimer from his post on 17th Jan:

"This is still beta software. It has bugs. There are kinks to be worked out, and missing features, too. For example, we still miss some tags here and there, especially for certain weblogs, due to some post detection issues. We're working on that. In addition, we will be providing RSS feeds so you can subscribe to your favorite tags. We're also working on ways that anyone can participate and be included in the system, keep your eyes on this space in the near future."

I've search around and found a few references.

Technorati's Kevin Marks (I think he works a Technorati) tries to clarify a couple of matters in his post 'Tagging is for posts'

"Over the past few days I've seen a lot of reaction to Technorati's new feature. I hope to deal with the wider social networking, spam and ontological issues at a future date, but I'd like to have another go at clarifying the details of rel="tag" again, as if Tim Bray, Roland Tanglao and Kevin Burton are all puzzled by it, we're obviously not conveying the idea clearly enough."

Kevin goes on to provide some detail on the following:

  • It is not for tagging a URL
  • Tags do not require Technorati-proprietary URLs
  • Tags are not invisible metadata

I'll keep digging.

Comments (2)
  1. Tags are about the stupidest way to make this happen. Tags entrust the end user to be honest. We all know 95% will, but the rest will use the tags for malicous purposes.

    If you want to extract metadata from blog postings, just apply baysien technique. It’s been effectively classifying text for years.

    What a dumb idea.

  2. David Sifry says:

    Thanks for the comments, I don’t think anyone really knows how all this stuff will end up, but we’re taking it one day at a time. As for bayesian (note the correct spelling) techniques, yes, those are one way of identifying information, but not the only way. And yes, I guarantee there will be spam, but we’re already working on a variety of filtering techniques to deal with those issues when they arise.

    So watch this space, we’re building this thing while we’re flying it…


Comments are closed.

Skip to main content