17 Feb 2004 ib   » (Master)

A few more words on communal and social designators.

The difference, I believe, lies not so much in the presentation as it does in the emphasis in either design. Communities are defined as a group of individuals, thus emphasizing individual attributes, such as eye color, political orientation, or number of friends within a given structure. Orkut, in this definition would be such a community site, creating new goals (finding a date, a business opportunity, or gaining more "friends") in itself. So is, for that matter, FOAF, which seems to be little more than vCard data in a new format (RDF). Why the designers of FOAF chose to use RDF rather than an existing and widely adopted format remains a mystery to me, maybe there are advantages, maybe it's just the old NIH. Both formats are easily extended, allow for external descriptors to be declared, and have provisions for inter-personal relationships to be documented and managed. In addition, vCard has a long list of supporting tools, from most mail applications to PIMs and address books.

Social structures stress the links (relationships) between nodes (individuals) rather than the individual itself. A good example of a technical design which focuses on relationships, rather than nodes, is XFN , the XHTML Friends Network.

XFN has a few advantages over communal systems such as Friendster, Orkut, or Tribe, in that it correctly recognizes and displays asynchronous relationships, such as Anna calling Bob a friend, but Bob considering Anna merely an acquaintance. Relationships in 'the field' are rarely as clean cut and simple as Orkut's mutual friends paradigm would suggest. Some are even one-sided, that is, they are maintained unidirectionally, even though this should be considered an exception, rather than a rule.

If social networking technologies are to raise above the current trend of dating facilitators and battles for the most "friends", emphasis needs to be shifted from the individual to its relationships with others. XFN, assisted by tools such as RubHub could be the answer.

The lowercase semantic web offers two interesting means to visualize and ultimately analyze such constructs on the web through its implementation of XFN, the XHTML Friends Network, and VoteLinks, a set of constructive semantics to blueprint approval or disapproval with propositions and similar data.

VoteLinks allow publishers to denote approval or disproval by adding a similar <code>rel</code>attribute.

One of the more interesting application, in this case, would consist of a few simple steps:

  • propose or discuss a proposition in writing on the web.
  • (optionally, see below) Define "seed" individuals
  • Call for vote
  • Readers respond on their webspace
  • Either tally all VoteLinks or
  • Define inclusions or exclusions based on XFN data
Such inclusions could involve simple queries, such as "only tally votes by individuals that have been tagged as friend or acquaintance by me", or "only tally votes that have been linked as 'met' by my seed individuals". The possibilities are endless, covering exclusionary approaches ("do not count votes from individuals with Little Green Footballs in their blogroll") as well as weight assignments ("three votes to my friends, two to their friends, one to everyone else").

Here XFN becomes a quite usable, and readily available, means to build working (if rudimentary) webs of trust, based on familiarity not with the voting party, but on acquaintanceship between a defined seed, who enjoys the proponents trust, to the voter.

The advantages of such a system are quite clear - by embedding a vote into a publicized information fragment, discussions of the voting behavior are easy and fundamentally remain associated to the vote. Neither XFN nor VoteLinks require additional software or technical changes to the proponents or voters setup. Infrastructures, such as Technorati and RubHub are already available, and inband designators should speed up adoption considerably, compared to out of band solutions.

One last thing, before I hang up:

Previously, I mentioned five attributes, I believed need to be assessed and considered in social network analysis: proximity, density, source, type, and direction. After speaking to and discussing the issue with Tantek Çelik and others, it has become clear, that I have indeed missed another important relationship attribute.

Frequency, the distance in time and repetitiveness of relationships.

The importance of frequency as a social designator becomes apparent, when seen in the context of relationship proximity and source. A casual encounter might result in a short term state of close proximity, yes may yield a low-density-high-distance link after some time.

That's it.

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!