Next-Generation Distributed Social Software Networks:
Designs and Applications

http://ctin511.randomfoo.net/

Leonard Lin

http://randomfoo.net/

lhl@usc.edu

Social Software

Social software is an enabling technology, facilitating mediated communications, collaboration, and other human interactions

Stewart Butterfield outlines five devices:

The social networking system I will be discussing will encompass all of these devices

Current Competition

There's a new bubblet going on. What's out there in the Social Networking space?

  Friendster Tribe.net PeopleAggregator
Status Current front runner; dead end? Initial slow growth, now gaining momentum Pre-alpha; FOAF based
Funding $13 million in a round of funding led by Benchmark Capital and Kleiner Perkins Caufield & Byers $6.3 million by Knight Ridder, The Washington Post Co. and Mayfield none
Notes Antipathy towards user-base; lack of tool development; however, largest network Actively developing new features; currently specializing in listings Notable for using FOAF; front-end for larger digital lifestyle aggregation framework; see dev notes

Why Bother?

Lynchpin of entire next-generation of collaborative and information technologies.

Distributed Social Software

What's the difference between a distributed social software system versus a centralized system?

Defining an end-to-end architecture:

Distributed Centralized
Platform Destination
Open Closed
Extensible Fixed
User-centric Provider-centric

Why DSS?

Metcalfe's Law:

The value of a communications system grows as the square of the number of users of the system (N2)

Goals:

Features

Technology Framework

DSS Architecture

Scenarios

  • Collaboratively Filtered Everything
  • Food Sniffer, reviews
  • Trust-based lending, reputation system
  • Semi-permeable blogging, data exchange

Challenges

  • User/Network migration
  • Fragmented ID/Trust Segmentation
  • Privacy Management
    • Dealing w/ programmatic access, control
  • Trust Network robustness, attack resistance
  • Extension architecture: plugins, autodiscovery
  • Scalability - caching mechanisms, performance, enabling arbitrary queries

Timeline

This field is moving quite rapidly, both in the academic and commercial realm, quite time sensitive.

By leveraging pre-established open standards and open source technologies, time-to-market can be cut by a significant (magnitude) factor

Components Est. Man-months
Core Architecture 12-18
Web Interface 6
P2P Client 6
OpenGIS extensions 4-6

Thank You

Questions?