Riddle, P. (2005). Tags: What are they good for? INF 385Q Knowledge Managment course. unpublished.

Prentiss Riddle comes at tagging from the perspective of knowledge management as an M.S. student in Information Studies, the University of Texas School of Information. He briefly states that the strengths and weaknesses of tagging that have been shown with del.icio.us and flickr which are the most known and heavily used systems. He also looks into other domains drawing on a number of unanswered questions about tagging that were not known at the time (May 2005) while offering directions for the future.

Tagging is good for:
1. Personal Recall (e.g. re-finding something that was tagged)
2. Supporting Social Effects (e.g. emergence of flickr group tagged “squared circle”
3. Serendipity (e.g. like discovering a good thesis book in the library that catches your eye while looking for another)
4. Novelty (e.g. spikes in popularity of tags such as AJAX in del.icio.us)

Tagging is not good for:
1. Precision and recall regarding information retrieval (e.g. multiple tags to an concept or a tag being associated with multiple meanings can return queries that will be noisy (low in precision) or incomplete (low in recall)
2. Ontologies in the sense of hierarchical system for classification (e.g. no formal system such as Dewey Decimal System for creating a controlled vocabulary)

Unknown Domains tagging might be good for
1. People (e.g. social relationships such as consumating dating service
2. Products (e.g. amazon.com)
3. Places (e.g. geographic locations such as museums)
4. Music (e.g. iTunes shared libraries)
5. Filesystems (e.g. apple spotlight model if shared)
6. Tasks (e.g. organisational intranet)
7. Scale — Does tagging perform better at some scales, and under what circumstances?
8. Organizations (e.g. factors of organisational culture that come into play)
9. Normalizaiton (e.g. desire for UI that helps with consist retagging)
10. Ease of Use (e.g. UI tools that emphasise the low cost in terms of user time and effort for tagging)
11. Information Retrieval Methods (e.g. unknown if can be usefully clustered)

Directions for Research
Unclear how IR tools could best be used to study tagging which are more social and user-centered. Feels difficulties with Usability research which is single user and sees some usefulness with ethnographic systems although they may be limited in addressing hypothetical questions. Sees some hope in the open-source community.

Interesting References Cited:
Quote from David Weinberger (2005), Taxonomies to tags: From trees to piles of leaves:

There is a simple solution, however, to all of these issues: Create the tags and
experiment. Tags are becoming a new layer of infrastructure. They will enable
yet another round of creativity as we figure out, collectively, what variety of
things we can do with this metadata.

Thesis Relevance:
Looking at tagging from an IR and knoweledge management perspective

Technorati Tags: , , ,