How Important Are Tags to You?

June 1, 2008 Dennis D. McDonald

How important are the tags you assign to the things you write and publish online? And how important are tags to your finding useful information online?

I thought about these questions when I read Sarah Perez’ Semantic Tagging with Faviki. Faviki, a new social bookmarking service, suggests what tags to use based on structured data extracted from Wikipedia into an online database called DBpedia.

Unlike services such as del.icio.us, Faviki-suggested tags incorporate a structured tagging vocabulary that has been created and maintained through a communal effort of experts. For example, Faviki prompts you to use a tag such as “coca-cola” instead of “cocacola.” The former term has been incorporated into the DBpedia database along with references to explicit Wikipedia locations for this reference, and this is the term that is recommended.

It’s easy to see the supposed benefits of such an approach to tagging: consistency in assigning tags should make it easier to share information with other people, and relating tags to a structured area of knowledge should make the tag maintenance process easier to perform for those responsible for updating the tags.

While controlled indexing vocabularies and classifications schemes have existed for as long as indexes, catalogs, and information retrieval systems have existed, the benefits of such controlled vocabularies have been somewhat limited to professional and specialized communities or other organizations that already have a vested interest in standard ways of referring to concepts and ideas. Once authorship and usage extend beyond such communities — which happens very easily online — it’s possible that the advantages of standardization, specialization, and specificity of tags might start to break down as profession- and knowledge-based borders are crossed.

This is sort of a “chicken and egg” situation where, it seems to me, you need to understand enough about a discipline in order to make the best possible use of a set of structured tags that relate to that discipline.

On the other hand, a service like Faviki explicitly relies on a tool (DBpedia) that is based on a cross-disciplinary tool like Wikipedia. That suggests that there could exist an underlying cross-disciplinary framework that makes it simpler to cross over community and discipline based boundaries.

But will this be obvious to casual users of Faviki tags? Maybe, maybe not. But I certainly intend to try it out.