We are each, in our own ways, experts on something. The great among us might have expertise across several domains, but every one of us has a passion, a hobby, a predilection, into which we pour tens of thousands of hours of (generally) unpaid work. These passions define us to our peers, far more than any simple explanation of what we do to earn a living. Furthermore, these passions help us select our peers: we are more likely to spend time with individuals who share our obsessions. Nor are these passions singular, as they invariably present themselves in a multiplicity: family and home repair and computers and cinema. As our passions wax and wane, this multiplicity constantly rearranges itself, weighted by the interests of the moment.
We can declare our passion by attending meetings, visiting websites, purchasing material goods, and so forth. These are the outward signs of an inward state. Just by observation, it is possible for one person to learn of another’s passions. Where your heart is, there your feet will follow. Where are you spending your time? What are you doing with that time? Time, as the ultimate currency of the 21st century, represents the atomic unit of attention. Our attention is absorbed by the objects which fascinate us – our passions expressed.
It is possible – and relatively easy – to monitor an individual’s attention. In the context of the online world, it’s actually harder to ignore attention than to monitor it. Every web server, every chat room, and every blog keeps detailed records of who was there, when, and for how long. These records are maintained for the benefit of the provider of the service, but, since this information is an essential part of the data shadow of the individual, generated by their activities, it would also benefit the individual if this information could be fed back to them.
Although the individual generates a data shadow, this information is normally inaccessible to the individual; there are no mechanisms in place for the individual to store this information, nor any well-developed techniques for the individual to analyze this information for their own benefit. Yet, in the few situations where this information is put to use – Amazon.com’s recommendations being an outstanding example (adapted from Pattie Maes’ “Firefly” project at the MIT Media Lab) – an individual can improve the online experience enormously. A small investment in data analysis can produce enormous returns.
Analysis of an individual’s data shadow can be treated as a type of metadata; that is, information about information. Although Tim Berners-Lee, the inventor of the World Wide Web, has been preaching about the importance of metadata since its earliest days, only in the last 24 months has metadata has emerged as a major developmental focus in online services. Metadata has become the defining feature of “Web2.0”, most often implemented through “tagging.” Tagging allows an individual – or, in rarer cases, an algorithm – to assign semantic keywords to any particular piece of information – a clip of video, a URL, a blog entry, etc. These tags do not reside in the data, but rather, sit in a “cloud” around the data, creating an envelope of context for data which, in isolation, has no specific meaning.
The two most familiar examples of tagging in use today are the social link sharing site del.icio.us, and the photo sharing site Flickr. Each allows the individual to provide an arbitrary set of keywords – “tags” – which are thereafter permanently associated with a link or image. Thus, the individual’s continued interaction with these services gradually generates a second-order set of metadata, the total set of tags created by an individual, which has come to be known as a “folksonomy”, that is, a user-generated taxonomy.
Tags have become instrumental in the individual’s quest to find specific information inside the riotous wealth of information presented by the Internet. Given that most individuals will create tags that are, more often than not, similar to those created by other individuals, folksonomies tend toward isomorphism: it is likely that we will use the same tags to describe the same things. del.icio.us actually performs an analysis of folksonomies, presenting a list of tags from the collective folksonomy when a link is being recorded in the system – you can use the tags selected by others to tag a link. Furthermore, del.icio.us presents your entire “tag cloud” – the set of all tags you have ever generated – when you tag a link, making it easy to keep to a regular folksonomy. A regular user of del.icio.us quickly develops a rich folksonomy (my own runs to several hundred keywords, from “ABC” to “Zen”), and this metadata about metadata has value in itself, which is, at present, completely unutilized.
What would happen if we took a community of users – who have self-selected themselves around a common interest, such as travel, or sport, or the existence of extraterrestrial intelligences – and analyzed their tag clouds? In the case of each individual within the community, it would reveal their specific passions: this one loves Bali, that one the West Coast Eagles, and another, the Zeta Reticulans. This is valuable information, and can be used to greatly improve the experience of each of these users. Taking it to the next logical step, what would happen if you analyzed the tag clouds collectively, searching for common threads across members of a community? Here you’d find the communities of interest within a community: some people are obsessed with Asia, others with Rugby Union, still others with abduction experiences. There’s a high degree of likelihood that this analysis would closely mirror the communication affinities within these communities: that is, individuals with a high degree of isomorphism in their respective folksonimies would be more likely to be in communication with one another. When this tag cloud analysis is compared to the actual record of communication within the community, it should show not only who is talking to whom – reinforcing the comparative analysis of the tag clouds – but it should also clearly indicate who should be talking to whom.
Comparative analysis of folksonomies within any community will reveal the total knowledge encompassed by that community; this is the first step to establishing systems which allow individuals within that community to quickly and easily connect with others who possess some specific knowledge, or who have some specific need for knowledge. When an individual can immediately find the “go to” people with expertise in any specific area of interest, that individual’s effectiveness is multiplied. This straightforward approach will create communities of collective hyperintelligence, where the collective value of the community’s knowledge is greater than the sum of its individual parts.
It remains only to put this thesis to the test. The pieces are in place to make it all work: we have sufficient capability to record and store metadata, we have the ability – if not the will – to expose this metadata to the user, and we’ve begun to develop techniques to analyze this metadata. We need to move from analysis of the individual to an analysis of the collective, laying the exposed individual’s data shadow side-by-side with other co-communicants, to reveal the hidden structures of knowledge and expertise within the community. Once exposed, these structures can be reified with existing communications technologies, amplifying the expertise of the entire community. If this thesis is correct, communities which leverage their metadata to improve their effectiveness should quickly leap ahead of their peer communities in almost every endeavor. Passion, meet acceleration.