Thursday keynote notes:

Hello everybody, welcome to my cult ;-)

Social Software: Software that supports group interaction -- many2many.

-Ridiculously easy group forming is really new. Still learning how to make these things.
-Definition doesnt specify a specific class of applications. Email, Weblogs not necessarily social, but can support social patterns.
-Groups are a runtime effect.
-Unforseen social patterns.

^^^Why a group is its own worst enemy

WR Bion, a psychologist studied how neurotics were conspiring to quash therapy. Should he view it as individual actions or coordinated action? People are fundamentially individual and social. Group effect is deeper and happens quicker than we expect.

Social stickyness: when you dont leave a party you dont like. But mass exit still prevails when there is a triggering effect...the paradox of groups. Complicated moment of a group coming together, then you get group effects.

Bion saw the group defending itself and trying to defeat getting treatment. Three common patterns for all groups:
#sex talk - away from the sophisticated dialogue
#identification and vilification of external enemies
#religious veneration of an icon ( clay, clay, clay )
His conclusion: Group structure is necessary to defend a group from itself.

Communatree, an early BBS, founded on the principles of open access and free dialup. Throw off structure and new beautiful patterns arise. This does happen when you institute a new mode of communication to a group. Teenage boys subsumed the BBS (overrun) with sex talk. Site shut down.

Q: A technical or a social problem?

A: Doesnt matter -- can't seperate the two.

Pattern keeps repeating itself, as people dont learn from history.

People who make social software are closer to economists and political scientists than programmers.

First crisis forces people not just to make rules, but to have rules about making rules. Constitutions are essential.

^^^Why now?

Reason #1 -- we didnt know what we were doing

-The number of people writing tools to support and enhance group interaction is exploding.
-Size and scale doesnt support the dense interlinked pattern necessary for groups. Less is more.
-New tools: weblogs, wikis, RSS...
-Has nothing to do with technology, took a long time to realize that people talking to each other is what its about

Reason #2 -- its web native

• lightweight, loosely coupled, web-based

Reason #3 -- we can now have a small pieces loosely joined
Happening - multimodal, loosely joined
-Emergent Democracy

Reason #4 -- Ubiquity
-In many situations (some groups), all people have access to the network
-Lets you take it for granted
-Now all offline groups can have an online component
-Assume that people can be in-room and using an online mode
-Library of Congress Meeting Wiki: shared repository for group memory

^^^Core concepts

-What is it that makes a large online group successful over time? It depends.
*Natural grace and supernatural grace
*Normal experience of groups is they are failures (skinny tail of a power-law in yahoo groups)

Accept

1. You cannot completely seperate technical and social issues.
2. You cannot program social issues. The group will assert its rights. Put into the hands of the group the responsibility for defining itself. Members are different than users. Allow the core group to express itself or it will find its own way to do so.
3. Core group has rights the trump individual rights in some situations. Leverage the core group to maintain sophisticated goals and progress.

Design

1. All groups have a constitution, formal (in code) and informal (how we do it around here).
2. Handles the user can invest in. Anonymity and weak pseudonymity doesnt work in conversational spaces. Reputations are not linearizable or portable. Ebay works because its not a social situation, its pure linear transactions. Has to be a penalty for changing handles. Changing your identity is really weird
3. You have to design a way to recognize members in groups standing. Reputation enhancement by linking handles.
4. Need barriers to participation. To give the core group ways to defend itself. The user is the group and the ease of use should be for the group.
5. Spare the group from scale. Have to let users hang on to the less is more proposition. Soft-forking, loose clustering. Protect your users from scale, or it will collapse, dissipate or turn into broadcasting

---

Q&A

• Online communities are centralized. Boards and other communities are the source of literature about what might be common principles, but new software forms are emerging -- more decentralized, on/offline.
-Why not richer media? Immersive environments are a strange new effects, but less effective for getting conversations done. Email as the serial killer app. Games may not transcend to the business world. -Need more goal oriented software.

---

Links to Other Summaries
Journal of Hyperlinked Organization write-up
Jon Lebkowsky's notes
-Trevor Smith's notes

-"Shirky's Social Software and the Politics of Groups (March 9, 2003) article" <http://shirky.com/writings/group_politics.html>