Abstract
This chapter discusses folksonomies as a novel way of indexing documents and locating information based on user generated keywords. Folksonomies are considered from the point of view of knowledge organization and representation in the context of user collaboration within the Web 2.0 environments. Folksonomies provide multiple benefits which make them a useful indexing method in various contexts; however, they also have a number of shortcomings that may hamper precise or exhaustive document retrieval. The position maintained is that folksonomies are a valuable addition to the traditional spectrum of knowledge organization methods since they facilitate user input, stimulate active language use and timeliness, create opportunities for processing large data sets, and allow new ways of social navigation within document collections. Applications of folksonomies as well as recommendations for effective information indexing and retrieval are discussed.
TopIntroduction
A key problem facing today’s information society is how to find and retrieve information precisely and effectively. Substantial research efforts concentrate on the challenges of information structuring and storing, particularly within different sub-disciplines of computer science and information science. In this context, information retrieval studies focus on methods and algorithms to enable precise and comprehensive searching of document collections (Frakes & Baeza-Yates, 1992; Stock, 2007a). In addition, techniques of knowledge representation have been established (Cleveland & Cleveland, 2001; Lancaster, 2003; Stock & Stock, 2008). Most prominent are approaches of document indexing: i.e., assigning content-descriptive keywords to documents. This enhances retrieval techniques and aids users in deciding on a document’s relevance. Different knowledge organization systems (KOS) are developed to support sophisticated document indexing. Common examples of KOS include classification systems (taxonomies), thesauri, and controlled keywords (nomenclatures).
Recently, a well-known problem of indexing documents with content-descriptive metadata has been addressed from a new, user centered perspective. Within the so-called “Web 2.0” (O’Reilly, 2005), web users have begun publishing their own content on a large scale and started using social software to store and share documents, such as photos, videos or bookmarks (Gordon-Murnane, 2006; Hammond, Hannay, Lund, & Scott, 2005). And they have also begun to index these documents with their own keywords to make them retrievable. In this context, the assigned keywords are called tags. The indexing process is called (social) tagging, the totality of tags used within one platform is called folksonomy. A tag cloud is a popular method for displaying most frequently applied tags of a folksonomy visually (Figure 1).
Figure 1. An exemplary tag cloud. Tag clouds display the most popular tags within a folksonomy based system. The bigger the font size, the more documents have been indexed with a tag.
Thus, a folksonomy is an indexing method open for users to apply freely chosen index terms. Peter Merholz (2004) entitles this method “metadata for the masses”; the writer James Surowiecki (2004) refers to it as one example of “the wisdom of crowds.” The term “folksonomy”, as a combination of “folk” and “taxonomy”, was introduced in 2004 by Thomas Vander Wal and cited in a blog post by Gene Smith (2004). Smith uses the term “classification” for paraphrasing folksonomies. This term arouses a misleading and faulty connotation. The same holds for the term “taxonomy.” Folksonomies are not classifications or taxonomies, since they work neither with notations nor with semantic relations. They are, however, a new type of knowledge organization system, with its own advantages and disadvantages.
Key Terms in this Chapter
Tag: Within a given context, a tag is a keyword assigned to a document to describe it. Tags can be used for document retrieval. Folksonomy tags can be freely chosen by the users of a folksonomy-based system.
Folksonomy: An indexing method open for users to apply freely chosen index terms. The term “folksonomy” was introduced in 2004 by Thomas Vander Wal as a combination of “folk” and “taxonomy.”
Knowledge Organization Systems (KOS): Knowledge Organization Systems are (structured) representations of a knowledge domain, used for document classification and indexing. Common classical knowledge organization systems include classifications (taxonomies), thesauri, and nomenclatures. Folksonomies and ontologies are new forms of KOS.
Broad and Narrow Folksonomies: Broad and narrow folksonomies differ in whether multiple assignments of identical tags are possible or not. Systems with broad folksonomies allow to assign the same tag to one document several times (thus the tag frequency can be counted), whereas narrow folksonomies record every tag only once.
Knowledge Representation and Indexing: In the context of information storage and retrieval techniques, knowledge representation is concerned with providing methods for organizing and representing knowledge domains and sorting documents accordingly. A traditional way to do this is by document indexing: i.e., by assigning keywords or notations (usually taken from a controlled vocabulary or classification scheme) to a document to describe its content.
Tag Cloud: A tag cloud displays the popularity of tags, either for tags assigned to one single document or for all tags within a complete folksonomy-based platform. The bigger and broader a tag is displayed in a tag cloud, the more often has it been used.
Tag Distribution: The frequency of tags assigned to one document (or within a platform) can be counted and visualized as a tag distribution graph. Some specific forms of tag distributions are dominant within folksonomies: for example, the emergence of a “long tail”, which reacts to the rules of the power law. A “long trunk” may appear as well; the curve then follows an inverse-logistic distribution.