However, until recently, face-to-face meetings were part of this assessment process. And close physical proximity tended to mediate relationships for most of us.

But now we face the challenge of determining who people are and if they are our friends. And today this process is often mediated only by online channels.

The primary problem is identity - are they who they say they are? The secondary problem is authenticity - are they being real or assuming a different persona, and can I rely upon that persona? The tertiary problem is how to know and possibly to reference the varying degrees of relationship.

These problems have been subject of much discussion between myself and friends or colleagues over the past few years - both in real life and online. Our answers to the first two problems tend to come down to the fact that over time it is hard to maintain an alternate persona or identity and to keep it consistent. Thus sock puppets are usually uncovered due to inconsistency or they lose interest and fade away. Over time as you interact with someone they tend to reveal themselves in various ways and to demonstrate consistent patterns of thought, conversation and behaviour. There are also a number of identity management initiatives under way, such as OpenID.

The the third question is one to which there seems to be no easy answer. One possible solution is a representational approach as outlined in the XFN™ or XHTML Friends Network. This is described as "a simple way to represent human relationships using hyperlinks". But I'm not certain that merely representing the relationships actually helps us to know people any better. Nor am I certain that representing these relationships in a hierarchical classification will actually improve those relationships.

I have no answer to this last question and am interested to hear other opinions.

Some other people who've considered this topic recently :

Is a Social Media Friend Really a Friend?
Be a Real Friend to Your Social Networking "Friends"

By Carruthers via Aide-mémoire