Last Friday we welcomed Thomas “Folksonomy” Vander Wal to Nature. We spoke about various areas of join interest, including of course tagging and social software. Thomas was also kind enough to give a talk to assembled staff. Here are my impressionistic and rather inadequate notes from that session.
What is a tag? A simple piece of metadata externally applied to an object. Used for sorting and as a hook for aggregating. Acts as a personal marker.
History of tagging: Lotus Magellan indexed content and allowed annotations to be added. Also Bitsy (sp?) Bitzi, but this wasn’t collaborative. Del.icio.us got this bit right.
Folksonomy: the result of personal free tagging of pages and objects for one’s own retrieval. Usually done in a social context, but done by the person consuming the information not a specialist. The value is that it derives from people’s own understanding. Not really categorizing but rather connecting items and providing hooks so that they can be aggregated. Rashmi Sinha: Tagging taps into an existing cognitive process without adding much cognitive cost.
Folksonomy triad: Object, metadata (tag) and identity (user). Sometimes combined with community where different users settle on similar tags.
Folksonomies allow new objects to be discovered — things I haven’t tagged that have been tagged by other people who use the same terms as me.
Folksonomy versus taxonomy: Organisations create taxonomies to provide structure and make information easier to find. Users create folksonomies. Taxonomies and folksonomies overlap, but folksonomies go further — in one [presumably representative] case 70% of tags were not in the taxonomy. Folksonomies allow us to validate taxonomies, identify gaps and provide new terms to put into the taxonomy. Taxonomies provide formal structures and efficiency, and also the foundations to apply new understanding.
Taxonomy is structured, efficient, and provides a sold foundation. But it’s resource-intensive, non-emergent, and difficult to validate. Folksonomy is messy, provides poor ‘findability’ (but good ‘re-findability’), and is slow to emerge. But it’s relatively inexpensive, emergent, and enables continual validation.
Who tags? Each day, 7% of Americans on the web use tags. Following the same trend as RSS — not yet mainstream but growing.
Social vectors: Different groups refer to the same thing by different terms.
Business value of tagging: Tension between control, in-house, known (usually high) value, and consistency versus decentralisation, outsourcing, unknown value and emergence. Compared to approaches like abstracting, annotation, ratings, etc., tagging delivers relatively large value compared to the work involved.
Cold-start problem: Social software starts out as just software with no ‘social’ because there are no other users. Del.icio.us got around this by providing an incentive to individual users while at the same time enabling social behaviour.
Phases of interaction: Tagging, re-finding, exploring, searching, interacting.
Spheres of sociality: Personal, selective, collective.
The future: Portability and interoperability between services and devices.
[There followed a Q&A session during which I wasn’t able to take notes.]
Leave a Reply