Using keywords for the classification of data is a centuries-old concept. Traditionally part of pre-defined taxonomies, keywords were also integrated with early computing software. It was about five years ago, when tags as we know them today – free-text keywords with no hierarchical order – were introduced to the web. They quickly revolutionized the organization of information. With the arrival of user-created tagging on sites such as delicious.com and Flickr, the year 2004 marked the birth of folksonomies. Today, the use of tags is one of the backbones for many “Web 2.0” services.

It is therefore no surprise that this year’s ACM conference on Hypertext and Hypermedia, Hypertext 2009, had plenty of presentations on folksonomies. The research presented covered many topics on how to utilize the implicit information contained in tags. Tagging by users is an uncoordinated distributed process, that generates a lot of meaning beyond attaching a keyword to some content. Mining this data allows for all new ways of personalization (such as social search, that takes into account the user’s tag cloud), user experience (e.g. tag recommendation when adding new files) and exploitation of the “wisdom of crowds” (discovering the wisdom hidden in tag rankings and occurrences).

Over the last years, Valve has been working on a variety of projects where tags are integrated as an integral part of information architecture. We are excited to explore the latest trends in tag-based computing and are always looking forward to make use of the latest technological possibilities to serve our clients and their users with the best solutions.

Hypertext 2009, the annual conference of ACM’s Special Interest Group on Hypertext, Hypermedia and the Web, was held in the end of June at Turin in Italy. Besides folksonomies, the conference covered the evolution of hypertext-based information structures. Participants discussed the role of social contacts as hyperlinks, the use of hypertext for narrative tasks, the tracking of user behavior in hypermedia systems and last but not least Ted Nelson’s alternative to HTML, xanalogical data structures.