Category Archives: Tagsology


Tagsology: an information system built with tags, but with an ontologist and some technology moving quietly in the background holding the fort together.

    1. Start with a low barrier to entry: tags
    2. Mix in some structure, e.g. narrower terms, broader terms and related terms: ontologist
    3. Layer on some clever indexing, algorithms, clustering, auto-classification, content classifiers, etc: technology

The key is to keep the tags coming en masse. The ontologist should never change people’s tags, the ontologist should simply make sure the tags are communicating with each other. It’ll take a light touch. Also, everyone should be allowed to contribute to the ontology. Rather than centralized control there will be centralized oversight (again, light touch).

Same with technology: tag co-occurence and other neat little tricks will group similar concepts together, but only when the user wants that (so again, the light touch). There will be no forced structure.

This is where I see Flickr, Furl, YouTube,, Yahoo!’s My Web, Wikipedia, etc. heading. Tagging by itself allows too much information to fall through the cracks for searchers, while controlled ontologies miss the benefit of group knowledge. By mixing the two together you get individual expression that is easier to navigate. Easy creation, easy navigation: a tagsology.

One last thought: in the long term all the various tagging platforms need to communicate with each other. Imagine how much more powerful it will be when my tags stretch across all tagging systems and search engines know how to handle those tags.

Technorati tag:

Tags are a step on the way

Tagging isn’t the final destination for social search. Tagging is a step along the way. The beauty of tagging is that it reflects real-life. I turn to my friends and family for a lot of my information needs and with social search I can search or browse within my community of friends whenever I want to, rather than pushing emails to them. This could be through reviews they’ve written, web pages they’ve saved, photos they’ve taken, online communities they are part of, etc.

Search isn’t there yet, but it will be.

Obviously tagging in its current form has three big problems:
1. not everyone does it.
2. it’s susceptible to spam.
3. it’s not a controlled vocabulary so two people might use different tags for the same thing, such as –ToyotaPrius– vs. –Prius–.

My responses:
1. OK, I don’t need everyone to do it. But if some of my friends do it just a little bit, then algorithms can learn from their preferences and expand it to a larger set of data. It’s the search cycle. Start with human created data and make machines learn from that.
2. Yes it is. That’s why people and software need to constantly police it. Also, if I’m mainly concerned with my community’s interests, then I’ll trust them not to spam me.
3. No cataloging system is perfect, but a combination of tagging/folksonomies and catalogers/taxonomists/ontologists, can allow for individualism within a semi-controlled environment. See my post about tagsologies.