Stewart Butterfield's talk:

The Game Context as a Testing Ground for Social Software

O'Reilly Emerging Technology Conference 2003

Cory Doctorow doctorow@craphound.com

--

* Raph Koster: The future of gaming isn't flashy graphics and blowing up elves,
it's average people connecting with each other.

* We built a social software oriented game and run an alpha that we eventually
had to shut down -- it was so significant to peoples' lives that when we shut it
down, people all over the world logged in at the same time to cry together

* In the Game Neverending prototype, we made a social index -- you could count
people as friends, acquaintances and enemies. We had a leaderboard whose
unintended result was that all newbies were flooded with requests to be
acquaintances/friends, in order to build their scores.

* The social network maps that came out of this were so complex we couldn't load
them; at a couple of removes, the relationships are too complex to capture
anything interesting from them

* We wanted to have a model for the groups and group actions in our game. We
thought we could provide a polling tool -- the leader of the groups could poll
the members and see what they wanted to do. A contrasting proposal was to allow
people to put forth proposals that succeeded or failed based on whether people
were inspired to action.

* We thought about reputation systems -- Ultima Online had a problem with
player-killers; newbies would get repeatedly killed by mature players who'd
accumulate the XP and little bits of gold. So Ultima created a system where
killing people changed the color of your name; the response was that
player-killers teamed up to take turns killing each other (killing a player
killer was good for your reputation).

* Ultima tried Notoriety, where people could finger each other. Everyone painted
himself as a victim.

* Finally, they separated the system to allow servers where you could get killed
and servers where you couldn't.

* MUDS are really hard. CLIs suck. Without windowing, MUDDing was really, really
hard, but windowing, which allowed for descriptions and chats to be separated
out, made it easier.

* ConFab scopes the system by limiting those who you could speak to -- to people
who are in the same room as you. Changing rooms was too quick, though, and so we
had to add a delay to create the transition.

* ConFab lets you annotate your contacts

* The design objective is to enable communications groups -- you can drag
present people into your chats and they'll pop in.

* You can limit the scope of conversations to your groups

* The history of civilization is the history of creating names for ourselves --
i.e., John Miller