Cesar Brea's Weblog

Our KM Cluster
panel on technology for social networks at IBM Research three weeks ago
was well-received.  Bill Ives asked me to recap some of the points.  Here goes:

The question on the table was, “what makes good technology for social
networking applications?”  To answer this, the logical thing is to
ask first, “What makes a good social networking application?”

In summary, you have to have something valuable to exchange, and not
lose it in a thicket of other junk.  Second, you have to group
users into tight “affinity groups” within which they are likely to
share.  Finally, third, you have to make both contributing and
consuming information really easy.  To illustrate these points, at
great professional risk, I re-told the (true, I swear) Tale Of The Binge-o-Matic.

So then you ask yourself the following questions about technology:

1. at what cost can I modify it to focus the feature set only on the one or two things that are most valuable to exchange?

2. does its scheme for defining and managing groups and
permissions  support the affinity group structure I think will
maximize sharing?

3. at what cost can I modify it to support the simplest possible structured contribution and consumption of what’s shared?

Generally speaking, the first and third of these are easy if you’re
custom-building a web app. Any good package should also make them
easy. 

The second is hard.  The right way to do it is with abstractions
that support inheritance of group properties and permissions.  But
abstractions can be seriously slow if not done well.  (Getting
this right is part of what makes OpenACS/.LRN really special.)

Thanks again to Bill, and to Kate Ehrlich for hosting us.