The Tagcloud

With the emergence of Internet technologies summarized under the buzzword Web 2.0 user participation was pushed to a new level. Websites as del.icio.us, Flickr.com or CiteULike.org make use of meta information provided by the users themselves. The phenomenon of thousands of web users classifying, indexing or describing websites, links, photos, songs etc. was appreciated enthusiastically and labelled as social bookmarking or folksonomy, to describe the collective intelligence at work. In fact the actual act of information retrieval on the Internet got a new tool, the tag, a digital post it, added to a link, a website, a photo or a song. Although the W3 recommendations suggest adding meta information to websites and their content such as pictures, often these information are missing or not supporting the actual request. The promise of the collective act of providing meta data is to attach semantic information to the content published on the Internet. The act of tagging includes besides the description of content, a validation. It adds a second layer to the stored data, a layer that promises to revolutionize the way users search and deal with information.

Instead of providing a paper to PUMS, I set up this work in progress which will address the phenomenon of meta information in Web 2.0. This includes an explorative research on websites and services such as last.fm, an online radio station, CiteULike.org, a social bookmarking service for academic papers, Flickr.com, the most popular website for posting and tagging fotos, and del.icio.us, a service for storing and tagging boomarks. The presentation at PUMS will summarize the literature on meta information in Web 2.0 and highlight the problems and expectations related to it.

In my presentation I will describe tagging as a quest for objectifying the subjective, like personal taste and special interests. In conclusion I will argue that meta information for useful and efficent information retrieval will need communities that share the same interests and expect the same standards of the data they are tagging. In that way one could think about a man-machine interaction to define and to retrieve precisely the kind of pictures, texts or the tunes one was looking for. Social bookmarking would become a way to separate the wheat from the chaff.