The Semantics of Tagging - Social Semantics: The Search for Meaning on the Web - page 110

a term ('music') or neologism ('toread') in natural language. The third space is the
resource space , the set of all resources , where usually each resource is a website
denoted by a unique URI. 2 A tagging instance can be seen as the two edges that
link a user to a tag and that tag to a given website or resource. Note that a tagging
instance can associate a date with its tuple of user, tag(s), and resource.
From Fig. 5.1 , we observe that tags provide the link between the users of the
system and the resources or concepts they search for. This analysis reveals a number
of dimensions of tagging that are often under-emphasized. In particular, tagging is
often a methodology for information retrieval , much like traditional search engines,
but with a number of key differences. To simplify drastically, with a traditional
search engine a user enters a number of tags and then an automatic algorithm labels
the resources with some measure of relevance to the tags pre-discovery , displaying
relevant resources to the user. In contrast, with collaborative tagging a user finds a
resource and then adds one or more tags to the resource manually, with the system
storing the resource and the tags post-discovery . When faced with a case of retrieval,
an automatic algorithm does not have to assign tags to the resource automatically,
but can follow the tags used by the user. The difference between this and traditional
searching algorithms is twofold: collaborative tagging relies on human knowledge,
as opposed to an algorithm, to directly connect terms to documents before a search
begins, and thus relies on the collective intelligence of its human users to pre-filter
the search results for relevance. When a search is complete and a resource of interest
is found, collaborative tagging often requires the user to tag the resource in order
to store the result in his or her personal collection. This causes a feedback cycle .
These characteristics motivate many systems like del.icio.us and it is well-known
that feedback cycles are one ingredient of complex systems (Bar-Yam 2003), giving
further indication that a power-law in the tagging distribution might emerge.

5.2 Detecting Power Laws in Tags

This section uses data from del.icio.us to empirically examine whether intuitions
regarding tagging systems as complex systems exhibiting power law distributions
hold.

5.2.1 Power Law Distributions: Definition

A power-law is a relationship between two scalar quantities x and y of the form:

cx α + y = b (5.1)

2 Most systems such as del.icio.us store only the URI. The resource space, in this definition,
represents whatever is being tagged, which may or may not be websites per se but resources
themselves.