Social Bookmarking

What is Social Bookmarking?

The holy grail of search engines is relevance. One recent development towards finding relevant material is social bookmarking. Exampels of social bookmarkign systems include del.icio.us, CiteULike and Connotea.

Social bookmarking systems use the behaviour of many people (the 'swarm') to identify resources of interest. It is based on the idea that if many people bookmark a resource, it is important. This is similar to the way that Google, Teoma and other search engines rate sites according to the number and rating of other sites which point to them. It also allows you to follow the tags applied to a resource to find other resources which share the same tag, or to follow other tags used by the people who tagged the resource.

The problem with most social bookmarking software is that it makes no distinction between 'important' people and the rest. In fact, simple weighting means that the most significant people are those who create the most bookmarks. Most social bookmarking systems are also anonymous, so you can't identify and follow the tags of people you know and respect.

The swarm is also far too democratic - it is unlikely to have a focused interest on early 15th century French literature! The most popular tags on del.icio.us are mostly IT-related - blog, design, web, programming, software, web2.0, ajax, css, music, tools, reference, linux, webdesign, news, javascript, video, art, java, blogs, and shopping and on CiteULike (which is more oriented towards academics) cancer, collaboration, design, evolution, information, network, review, social, structure, theory.

HEURIST's take on social bookmarking

HEURIST attempts to overcome the democracy of the swarm by not accepting anonymous users, and by allowing the identification of groups of colleagues who share specific interests - we call these User Clouds, although our programmers think they should be called Friends (but who is to say we actually like the people who occupy the same academic space?).

HEURIST then orders search results according to the average of the ratings provided by you or by identifiable groups of colleagues. This means that referernces rated highly by a group of people with identifiable shared interests will appear near the top of search results. Not everyone is limited to a single research area, so each user can define several User Clouds covering different areas of interest, allowing focussed discovery of new resources. This is a much more powerful methodology for discovering resources of specific interest than a generic 'popularity' index. We are still working on the weighting algorithms as at March 2006.

The HEURIST approach to social bookmarking will also allow the identification of colleagues with shared interests through identification of users who have a similar pattern of bookmarking. We expect to work on this in third quarter 2006 once all major functionality is complete and the user base starts to grow - it's not very useful until then.