Posts Tagged ‘tagging’

Implicit (tacit) knowledge Explicit knowledge

March 26, 2009

“[What I know of my children] is too long and deep to be exhausted in words, too twisty, entangled, and intertwingled to be made completely explicit.”
David Weinberger (2007): Everything is Miscellaneous, Times Books, New York


The Malleabaility of Words

March 10, 2009

“The malleability of words -their ambiguities, connotations, implications, and double entendres – is what makes language fun. And it’s also part of what attracts people to tagging.”

Gene Smith, Tagging: People-Powered Metadata For The Social Web, p. 84


March 10, 2009

A folksonomy is a classification system on the basis of user-generated tags.

“In a folksonomy, the relationships between tags are inferred based on their usage patterns. There are no formal relationships in a folksonomy, other than perhaps ‘degree of relatedness’. Because a folksonomy uses agorithms to look at tagging patterns, two tags that have no known semantic relationship may have a statistical relationship.”

Gene Smith, Tagging: People-Powered Metadata For The Social Web, p. 82

Metadata and Tags

March 10, 2009
    What can metadata help with?

  • Finding data;
  • Managing data;
  • Co-relating data.
    Three categories of metadata:

  • Descriptive (details about resource; often ambiguous)
  • Administrative (author, date etc.; unambiguous)
  • Structural (data about the structure/organisation of the resource; used to co-relate different sets of data; unambiguous

    Seven kinds of tags:

  • Descriptive (e.g. webdesign, drama)
  • Resource (e.g. video, blog)
  • Ownership/Source (e.g. nytimes)
  • Opinion (e.g. cool)
  • Self-reference (e.g. mystuff)
  • Task Organizing (e.g. toread, todo)
  • Play and Performance (e.g. squaredcircle, seenlive)

Gene Smith, Tagging: People-Powered Metadata For The Social Web, p.63/7

Tagging: design decicions

March 10, 2009

“Before you start designing your interface or planning social navigation features, you’ll want to think about the relationships and rules between users, resources, abnd tags in your taging system. In other words, you’ll want to think about the architecture of your tagging system. In this context, architecture simply means the abstract design decisions involved in creating a tagging system.”

Define rules and relationships between users, resources, and tags:


  • Who are they (personas)
  • How are they associated with the site? (Membership)
  • Turnover: What happens with joiners and leavers?
  • Activity? How enthusiastic are they?
  • Community: How do they engage with other users? (Followers, e.g. – Contacts, e.g. facebook – Groups, e.g. Ma.gnolia)
    Resources (system or user-contributed resources)

  • What is being taged? Original or pointer? “Systems that use pointers [like] are collaborative – many people can tag the same resource with their own unique tags. Systems that use originals are not collaborative at the resource level. althoughh you can still aggregate tags acroiss users.”
  • Privacy: Everyuthing is public – Configurable bur public by default – Configurable but private by default – Everything is private
  • Restrictions (e.g. file type, object, genre/category, origin)
  • Dynamism (number of resources and the rate at which they change impacts on tagging behavious)

  • Permissions: Who can tag what? “Most taggings systems with user-contributed resources have a simple and straightforward way of determining permissions. The people who contribute resources get to tag them and can add, edit, or remove the tags later if they want [e.g. wordpress blog]. In some cases, these permissionscan be extended to other users in the system. In Flickr, for example, you can give your contacts and friends permission to tag your photos (they can also delete the tags the add).” p.50
  • Truth: Where are the tags? (actual file or reference in database)
  • Control: Should you censor tags? Social policing might be a feasible and effective way to ‘control’ tags.
  • Patterns: Understanding the Power Law (the distribution characterized by a few elements occurring with a high frequency and most with a low frequency.” (e.g. Pareto’s law that 80% of wealth is held by 20% of people or Zipf’s law …)

Gene Smith, Tagging: People-Powered Metadata For The Social Web, p. 39-53

The Value of Tagging

March 9, 2009
    Value for users: Return on Experience

  • Ease of Use: Tags are simple, flexible, extensible, can be aggregated
  • Managing Personal Information: Tags and Folders
  • Collaborating and Sharing: Seeding Communities, social proof
  • Having Fun
  • Expressing yourself
    Business Benefits: Return on Investment

  • Facilitating Collaboration
  • Obtaining Descriptive Metadata
  • Enhancing Findability
  • Increasing Participation
  • Identifying Patterns
  • Augmenting Existing Classification Efforts
  • Sparking Innovation

Gene Smith, Tagging: People-Powered Metadata for the Social Web, p. 23-37


March 9, 2009

“Folders may represent, if only crudely, a person’s emerging, often hard-won, understanding of the information items contained within, their relationships to each other, their important properties.”

Gene Smith, Tagging: People-Powered Metadata For The Social Web, p.25

Tagging vs. semantic Web

March 9, 2009


Tagged content does not equal structured data.


  • are labels with words or a combination of words that are attributed/attached to pieces of content – most often retrospectively by users;
  • describe content implicitly and hence are ambiguous (like any term in human language)
  • ;

  • tagging is a light version of the semantic Web.
    The semantic web

  • is like a logical layer that sits on top of the Web (of documents);
  • the semantic Web is a Web of structured data and a data-field is the smallest semantic unit in the Web;
  • using (standard) identifiers and expressions of relationship between data fields, description of content is made explicit and non-ambiguous;
  • the semantic Web is based on interoperable metadata standards such as identifiers and expressions of relationships.

The way from the Web of Documents to the Semantic Web requires a shift in thinking: from “thinking repository scale to thinking Web scale” (Silver Oliver)

Side note: Interestingly, according to semiotic thinking, human language terms (as utilised in tags) become meaningful in differentiation to competing semantic units, i.e. by implicit logical exclusion – ‘is not this, is not that’ – whereas in the semantic Web meaning is defered by explicit attribution of identifier and a limited number of expressions of relationships.