XTech 2006 news

Newsletter sign-up


RSS and Atom feed icon News feeds

Semantics Through the Tag

David Beckett (Yahoo! Inc)
Open data St. John 2

Tagging is a lightweight way to describe or label information and it is successfully used mainly for describing bookmarked URLs, images and weblog items. A tag is a word which has a meaning to the tagger without need for brittle pre-arranged category structure and multiple tags can be used in a tagset. The deliberate low barrier for using tags has the consequence that the meaning of a tag has to be implicit as there is no easy way to find out some tag means by a directory or catalog. The words are mostly written in some human language but that is not usually recorded.

Tags can have URIs when they are associated with well known services such as technorati, del.icio.us and flickr. These centralised services do indexing of tags either by web crawling, receiving pings from weblogs or having total knowledge when the tag database is owned by the service. These sites then allow you to search or browse from a tag (URI) to the known items described by the tags. There is usually a web service available for these operations as well as feeds of new items about some tag, providing some context.

A tag or tagset thus may be able to give you more information but not all tags are available this way, and there is still no way to find out what the meaning of a tagset (the words) is.

This paper will discuss how tagging can be separated from the services that provide them, allowing use of tags as separate entities. This separation will use a mapping of tags to and from RDF, showing how the context of tags (who, where, when made them) can be represented and preserved. There are multiple choices for modelling tags in this way which will be discussed along with their benefits and some of the unusual uses of tags will be considered as well as the different way that existing tagging services use tagging differently.

The paper will also discuss methods of how to find the meaning of a tag – going from a tag to something a human can understand as a description, which may be evolving – in a decentralized, web friendly way. Without the use of ontologies!

Chair: Ian Davis