I'm at the SIG-CR workshop on before the ASIS&T Annual Meeting in Austin TX, listening to information scientists discuss Social Classification. Papers on-line at http://www.slais.ubc.ca/users/sigcr/events.html

Jonathan Furner: Introduction

Why social classification? What's the big deal?

What are the implications for traditional systems and roles, for infrastructures (new and developing) and for the structures of traditional classification

[If tagging is about indexing, and indexing is about retrieval...]

Can we expect taggers to think like searchers, and should we?

Joseph Busch

"Tagging: It's the interface, stupid!"

- while the introduction focused on the linguistics of tagging, he's going to explore inter-relationships between language and use (in the interface).

- joseph's background is eclectic, and he's worked in a lot of spaces (public, private, corporate, ngo) always helping people find things; "we organize stuff". Using combinations of metadata specifications and vocabularies, Taxonomy Strategies' services specify the context- [or corporation] inter-realtionships between these in specific applications.

- he'll explore relationships between tagging, and tagging interfaces (words and the context in which they are assigned) and content organization (faceted classification) and its role in retrieval.

- tagging is better than just the words that appear in a piece of content, and all tagging is useful.

- tagging is an ongoing process that happens throughout the content life-cycle [document once is a librarian's urban myth] and the tags assigned help understand how the purpose and role of content changes over time.

- tools to assist the creation of metadata are critical to its successful creation and deployment.

- much web metadata is now pushed back to the content-creator (blog tags, flickr, del.icio.us), so tools that are successful are easy and intuitive (without training). learning takes place in a social discourse space, not a training space. Almost instantaneous feedback is important to maintaining motivation (allows people to tag/re-tag, and allows language to evolve), it's a form of real-time user-testing and feedback, that allows the development of emerging categories (new ways of seeing old things).

- popular tagging applications

His Four Tagging Rules for end users

1.Use specific terms

- apply the most specific terms when tagging content. specific terms can always be generalized, but generic terms cannot be specialized.

2. Use Multiple terms:

-use as many terms as necessary to describe what the content is about and whi it is important

3. Use appropriate terms

-only fill-in the facets and values that make sense. not all apply

4. Consider how the content will be used

- "anticipate how the content will be searched for in the future and how to make it easy to find.

How to support tagging / requirements for interfaces

- don't ask people for anything that you can derive in another way [e.g. Word's metadata, CMS applications]

- show precedents - what people have already done [del.icio.us}

- controlled vocabularies (pull-down lists)

- multi-value tags [flickr

- geo-tagging (drag and drop) [flickr]

- group tagging [flickr]

- clean-up tools [flickr]

- batch editing [flickr, del.icio.us, jt/iphoto]

- share / don't share [del.icio.us]

- explict ownership

- feedback [support workflow, cms - review and improve cycles, alerts, popular/tag cloud, email, subscriptions/news, interestingness, following individuals, recommendations]

jt: what about the tension between emerging semantics and controlled vocabularies. Doesn't the latter mitigate against the former?

Ways of Organizanizing Things.

- wurman - information anxiety [5 ways of classifiying]

- business processes / records management

- business products / classifications of stuff and services vs market segmentation

- editorial / content lifecycle, production processes

facet theory and implementation

- multiplier effects 4 term lists with 10 terms each = 10,000 combinations (easy to get variety/specificity from managable lists)

His big question: is faceted indexing the future of social tagging?

Susan's just asked NowNow.com... we'll post the anwer.

jt - facets ≠ easy

jb - need to understand more about motivations for tagging

q: tagging interfaces - interwoven was hard to use? Why?

- too many choices, too complicated

q: controlled vocabulary might work in a closed system, but doesn't it act as a limiting factor in terms of personal specificity: "at the shore" vs "down the shore"?

jb: problems of nuance and complications

q: controlling terminology removes some of the ability for motivation