Michal Migurski's notebook, listening post, and soapbox. Subscribe to this blog. Check out the rest of my site as well.

Dec 7, 2005 2:57am

tagging on the rocks?

In How Tagging Could Gain Ground, Philipp Keller laments a lack of "new, visionary, inventive articles on tags."

Keller argues that he's missing a "middle-distance" view of information, somewhere between folksonomic free-for-all of tags and the hieratic temple of DMOZ. His graph puts Google at the far-left, on the assumption that the search engine rewards recognition rather than novelty searches:

I have a few resopnses to this argument:

First, I think the results of Keller's quick survey of Google vs. Del.icio.us (using the search term cryptography) aren't representative. Del's audience skews heavily geekward, so it feels natural that links about math, technology and secrecy would be represented well. Picking another topic out of a hat, the results aren't so clean-cut. Google's results for organic farming lie at about the same level of generality as its results about cryptography. There are a lot of them, many fairly basic. Narrowing searches by adding terms (e.g. politics, how-to) helps squeeze the volume down a bit. Meanwhile, Del returns just 10 links, total. Hope they're good. Conversely, the quality of results for terms like web or blog in Del.icio.us is also generic and fairly undifferentiated. Google maintains a middle-distance from the user in more situations, while Del.icio.us swings wildly based on the composition of its user base.

Second, I'm losing faith in the potential for staying informed from tagged web content alone. I've been subscribed to several Del tag searches via RSS (e.g. attention, maps, infographics), and today I unsubscribed from the last of them. The data was far too general: I was either getting a very slow trickle of information, or a torrent. Attention in particular picked up sharply from fewer than a half-dozen items per day a few months ago, and accounted for a substantial percentage of my RSS volume. Many of these items were repeats or pointers to the Attention gang's own writings. Interesting, but I get this stuff via other channels, and wasn't learning enough new stuff.

Third, I think these two points show that a social bookmarky whatchamacallit like Del.icio.us has a population sweet spot. The quality of the links was highest when the system was populated by hardcore early adopters, and dropped when middle of the bell curve moved in. In particular, the quality drop was characterized by echoes and repetitiveness. I would argue that Folksonomy Helps Me Stay Informed On a Given Topic best when the diversity of a given tag/user/resource population is high and the headcount low.

Fourth, Keller is seriously on-target when he says that the "missing in-between view can be won by analyzing tags" - his love cluster example is like a textbook example of aspects in linguistics, showing the nuances of social meaning in a term as general as love. I expect that Del.icio.us could do something like this too, but the API terms are too limiting for external experimentation, and the math is hairy. If anyone had the gall to do this on a large scale, it'd be Google - they're already treating link text as a de-facto tag on a page, so they're arguably the oldest tagging outfit on the block.

In general, I don't think that tags as a concept are sliding into irrelevance, as Keller seems to suggest. I think the bigger picture is that tags are (currently) one kind of intentional choice that can be expressed digitally. There are other such choices that may aid in search, and I imagine they won't be called "tags". There is also the approaching time when tags stop being reflective of human choices, as automated others-tagged-this-with schemes become prominent, and spammers decide that Del.icio.us Popular is a good place to be.

October 2017
Su M Tu W Th F Sa

Recent Entries

  1. planscore: a project to score gerrymandered district plans
  2. blog all dog-eared pages: human transit
  3. the levity of serverlessness
  4. three open data projects: openstreetmap, openaddresses, and who’s on first
  5. building up redistricting data for North Carolina
  6. district plans by the hundredweight
  7. baby steps towards measuring the efficiency gap
  8. things I’ve recently learned about legislative redistricting
  9. oh no
  10. landsat satellite imagery is easy to use
  11. openstreetmap: robots, crisis, and craft mappers
  12. quoted in the news
  13. dockering address data
  14. blog all dog-eared pages: the best and the brightest
  15. five-minute geocoder for openaddresses
  16. notes on debian packaging for ubuntu
  17. guyana trip report
  18. openaddresses population comparison
  19. blog all oft-played tracks VII
  20. week 1,984: back to the map