I wrote a report together with Constance Steinkuhler from the MUD
Dev conf and from the GDC for another mailinglist, but i thought it
might be of interest here as well!
Its a bit messy, but here you go! (warning, its long)
<EdNote: Minor non-content changing edits made>
************************************************************
Conference notes from:
Game Developers Conference 2004
San Jose, California, March 24 - 26
MUD Dev Conference
San Jose, California, March 27 - 28
Constance and I from the SSOWS list (Society for Study of Online
Worlds) went to the conferences, and she has posted her report at
http://www.joystick101.org/ <EdNote: Appears to be down>
I complete my notes with Constances notes in those cases we went to
different sessions with relevance for MMOGs or research. Constance
Steinkuehler is a MMORPG cognitive researcher at the University of
Wisconsin:
http://www.sit.wisc.edu/~steinkuehler/
In advance I want to apologize for the not-so-perfect English. Its
not my first language, so please excuse the errors. This report is
also muddled by having not-so-intergrated parts: I paste in
Constances notes and post conference notes here and there.
The report that follows is a chronologic recapture of events:
It is Tuesday evening, I have been travelling for 24 hours and I
get out of the flight terminal in San Jose. I get into a taxi, and
ask to be driven to the Best Western Hotel. The driver asks
whether I want to go to the one close to the airport or to the one
downtown. I show him the address on the hotel voucher, he says
'downtown' and off we go. I look out the window and I see palm
trees and low buildings.
I get out of the cab, into the hotel and present my voucher to the
portiere at the reception. He looks at it and then says I don't
have a booking. 'Look here' He points in the text 'There must be
a typo or something, because it says California.
'Typo? This is California, right?' I smile.
He looks worried. 'No. This is Costa Rica.'
I say 'Costa Rica? Where is Costa Rica? I am in America, right?'
He answers slowly. 'Yes. You are in central America.' I look at
the clock on the wall. The conference is starting in twelve hours.
Lets get this straight. Costa Rica is a little country just below
Nicaragua, and just above Panama.
I take a room for the night, and plan to go to the airport as soon
as it opens in the morning. In my tiredness I mistake Costa Rica
for Puerto Rico, so when I try to reach people for help there is
some confusion about where I am, but after a couple of hours of
sleeping between phone calls (different office hours in different
countries) I have a booking on the first morning flight out of
Costa Rica.
When I sit on the plane on my way to Phoenix I am finally able to
see some absurd humour in the situation. And I have bought a hemp
bag with the text Costa Rica printed on it, and that is at least
something.
When I finally arrive to the right San Jose I take another taxi,
which takes me to the Best Western hotel. The downtown one. I throw
in my bags at the room while the taxi waits for me, and I'm able to
get to the session:
AI and Design: How AI Enables Designers
Speaker(s): Brian Reynolds
Basically what he said was that the game gets better if the ai
programming is done in cooperation with the game design. It was
really interesting to see how they evolved the AI from the first
version of Civilization (1991) to Colonization (1994) to
Civilization 2 (1996) to Alpha Centauri 1998 and to Rise of
Nations. He also provided some advice for AI development
After that I try to get to the roundtable about Massively
Multiplayer Engineering. It's held by the guy who was the lead
programmer for Asheron's Call 2, Jeff Johnson, and thus a crime not
to go to. Unfortunately I'm not the only one thinking so, so when I
get there its already full. Instead I make a run for:
AI: Gameplay & Design: A Marriage of Heaven or Hell?
Speaker(s): Jonty Barnes, Peter Molyneux
What I took home from this somewhat visionary lecture was mostly
feeling comfortable that Peter Molineux thought that what was
most important for future game ai development was
1. Language understanding,
2. Player personality AI and
3. Complex AI characters.
He also showed some neat screenshots from 'Fable' where NPC:s in
a village share a 'group mind' and are driven by a set of needs:
food, wood, rest, sex and protection.
Suddenly while listening to the lecture I see Michael Mateas and
Andrew Stern sneak out in an attempted discrete way. It becomes less
discrete when I capture them and we make dinner plans. They have to
hurry to the Game Awards ceremony since they are among the nominees.
Michael made Façade together with Andrew Stern. Façade is the
project that is nominated. It's one of the most interesting thing
that happened in the area of interactive story telling for the
recent few years, read about it here:
http://www.quvu.net/interactivestory.net/
Also check out Michaels papers, don't miss the one about the
office plant:
http://www-2.cs.cmu.edu/~michaelm/publications.html
And while we are at it: He is setting up an experimental game lab
at Georgia Tech, read about it here:
http://egl.gatech.edu/
At the awards ceremony flounder around looking for Anna, a collegue
who worries about me still being in Costa Rica, but I somehow must
be a bit confused cause I end up in the VIP area. When I realize
that I try to hop over to the non-VIP area but are stopped by a
person who tries to find drinks for me. He has seen the Costa
Rica-bag, gotten a brief summary of the strangely booked flight
tickets, and is determined to make things better. Turns out that he
is the organizer of the Austin Conference, (which is solely
concentrated on massively multiplayer games September 9-10, 2004,
check it out here:
http://www.gameconference.com/ ) . The awards
ceremony starts and proves itself to be a very lengthy thing: many
prizes to hand out. I sat through all of this:
Game of the Year
STAR WARS: KNIGHTS OF THE OLD REPUBLIC (BioWare Corp.)
Original Game Character of the Year
HK-47 from STAR WARS: KNIGHTS OF THE OLD REPUBLIC
Rookie Studio of the Year
Infinity Ward for CALL OF DUTY
Excellence in Audio
Chuck Russom for sound effects in CALL OF DUTY
Excellence in Game Design
David Chateauneuf, Patrice Desilets, Jordan Mechner and team for
game design in PRINCE OF PERSIA: THE SANDS OF TIME
Excellence in Programming
Dominic Couture, Feng Quan Wang and team for graphics
programming in PRINCE OF PERSIA: THE SANDS OF TIME
Excellence in Visual Arts
Masanao Arimoto, Yoshiki Haruhana and Satoru Takizawa for art
direction in THE LEGEND OF ZELDA: THE WIND WAKER
Excellence in Writing
David Gaider, Drew Karpyshyn, Luke Kristjanson and Peter Thomas
for writing in STAR WARS: KNIGHTS OF THE OLD REPUBLIC
Game Innovation Spotlights
EYETOY: PLAY (Sony Computer Entertainment Europe)
VIEWTIFUL JOE (Capcom)
WARIOWARE INC.: MEGA MICROGAME$ (Nintendo)
Four special honors, selected by the Game Developers Choice Awards
advisory board, were also presented at the ceremony.
Lifetime Achievement Award
Mark Cerny: Production genius and master collaborator on such
successes as arcade classic Marble Madness, and the Crash
Bandicoot and Ratchet & Clank series.
First Penguin Award
Masaya Matsuura: A pioneer of beat-rhythm games, with
groundbreaking titles like Parappa the Rapper and Um Jammer
Lammy.
Maverick Award
Brian Fiete, Jason Kapalka and John Vechey: The founders of
PopCap Games for bringing casual gaming to mainstream
audiences.
IGDA Award for Community Contribution
Ray Muzyka and Greg Zeschuk: The joint CEOs of BioWare Corp.,
they contribute to both the developer community and charities
at large. For more information on the awards and recipients,
please visit
http://www.igda.org/awards.
When the ceremony is over the hysterical dinner organization
begins. Luckily I find Anna. After running back and forth between
people quite large group, mostly consisting of researchers, is
formed and we trot away to a restaurant. Among others its Jesper
Juul, (check out his work here:
http://www.jesperjuul.dk/ ) who just
finished his PHD at the Copenhagen IT university (where they do many
interesting things, check this out:
http://game.itu.dk/ ). He is
currently consulting and doing some game development. Also there was
Michael Mateas (who is in my supervisory team). We book a meeting
for the following day to talk about my research plan. There is also
Magy Seif El-Nasr who is assistant professor at Pennsylvania State
University. She does some very interesting work on believable
agents, story visualization techniques and real time lightning
design. Check her publications here:
http://ist.psu.edu/SeifEl-Nasr/publications.html . Those were the
people closest to me, but it was a party of about 20 people. Gonzalo
Frasca cracked unforgettable jokes about the ADAPTOR that he found
in Anna's handbag: a very big device she had gotten from the IT
department here at HGO to always have POWER. By that time I was
getting very tired: the longest sleep period I had since I left
Visby was 2 hours. Got a lift to the hotel by Fleming Seay who doing
his PhD work at Carnegie Mellon, Project Massive. As I've understood
it its collecting data about MMOG players and to figure out ways of
supporting communities.
DAY 2
-----
Anna and I find a Starbucks on our way so I'm loaded with espresso
when I enter the first session:
Game Design Challenge: The Love Story
Speaker(s): Raph Koster, Warren Spector, Will Wright, Eric Zimmerman
Eric Zimmerman does a really good panel moderation and the three
others present their designs. Raph Koster presents a story
generator, and Will Wright does an unforgettable design where he
puts in a sim-love-social component into a war MMOG (first person
kisser). Warren Spector problematizes the whole thing in an
interesting way instead of making a design. Will wins.
I had originally planed to go to Encouraging Innovation in Game
Development by Andy House, but I feel stressed about having missed
out on 2 meetings I had planned for the Wednesday. I get hold of
Adam Matjelan, CEO of Terraforge (
http://www.terraforge.com/ ), and
we hook up for a meeting. Together we make our way to the next
session I had marked as criminal to not go to:
Towards Relevant Research: Collaboration 101
Speaker(s): Mark DeLoura, Robin Hunicke, Raph Koster, John
Laird, Michael van Lent, Will Wright
The panel was moderated by Robin Hunicke, a PhD student at North
Western University who is working on adaptive difficulty levels in
first person shooters. Mark DeLoura is Manager of Developer
Relations, Sony Computer Entertainment. Raph Koster is Creative
Director at SOE and was creative director for Star Wars Galaxies and
Ultima Online. John E. Laird is a Professor of Electrical
Engineering and Computer Science at the University of Michigan, but
he wasn't in the panel. Michael van Lent is a research scientist at
the USC Institute for Creative Technologies and a research assistant
professor of Computer Science at the University of Southern
California.
The first topic was about technical research and development. The
panellists agreed that technical research was something that they
kept more track on than research in other areas such as social
sciences and humanities. Marc said that he kept up with the
technical research via the SIGGRAPh proceedings. Raph said that the
scope of the different tasks one has as a game developer and a
technical researcher can make it difficult to make the research
directly applicable. It was also stated that the research that is
done within companies is done within closed walls. Michael said that
it would be good to have access to the code of different games in
order to play around with ai components in it. 'Give us access to
the systems'. Will said they actually have given away code (but I
missed out on what specific game they gave out the code for). Raph
said that it would be very expensive to give out systems for
experimentation, like a whole server and request players. (At this
point Adam and I exchanged surprised glances, since we just had been
discussing how we might get components to work together.)
The next topic was about what research to make. Raph stated an
interest in soft science' i.e. not only code. Will said that they
have so tight schedules that even taking a phone call could make one
miss a milestone. But that they have a contact person at EA for
research, John Buchanan (he was involved in doing Majestic). Will
also said that interns were a good bridge to research.
This led onto the next topic: What to look for in students for
internship? Will said that it was important to match students, and
that he got students from CMU, Georgia Tech and MIT. Raph said he
only had one intern slot this year and that that would go to someone
in San Diego. The rest of the panel mostly dealed with specific
desired features of interns. I found this disappointing since the
title of the panel had made me expect a discussion of what could be
seen as 'relevant research' between industry representants and
representants from academia. I saw Katherine Isbister sitting in
the audience, so we had a chat and decided to have lunch the
following day.
Constance's notes from the same session:
This roundtable was somewhat contentious (if not slightly
condescending) toward academics, so my notes will be brief (to
avoid trouble, since *cough* I'm one of the "slow as molasses" and
"useless" academic games researchers the panel repeatedly
bashed). Rumor has it that the negative tone was in response to a
panel scheduled several days later entitled, "Academics Fragging
Developers." I have no idea, nor did I ask. Hell, I don't even
understand what the fragging title means in the first place. I
went to GDC so I could get some insight into the development side
of games, not to frag anyone. Back to topic: The discussion
started with explaining that little academic research is useful to
designers, then explaining why (in a variety of terms), then
finally reducing the issue of 'collaboration between academics and
industry' to one of 'academics training students for later work in
industry.' I then stopped taking notes. Though admitting that, at
least for MMOGs, most of the problems that designers face are
"more soft science rather than hard science" (Raph), only
technology research and industry training was discussed (and then
promptly trivialized). As a social scientist who thinks I do, in
fact, have something to offer and something to learn from
industry, I can only say that NOT ALL MMOG DESIGN PROBLEMS
CAN/SHOULD BE REDUCED TO TECHNICAL PROBLEMS. I guess that's not a
very popular stance to take at the GDC, though.
Next I went off to have a meeting with Michael Mateas about my
research plan, and after some running around after lunch boxes we
sat down.
After that I went to an AI roundtable with the topic RPGs and
adventure games, moderated by Neil Kirby. There were some
interesting discussions, but nothing that was directly relevant to
my research.
Next stop was:
Creating the Right Mix of Static Versus Dynamic Content in a
Massively Multiplayer Game
Speaker(s): Rich Vogel
He had a very specific opinion about the rate: 70% of the content in
a MMOG should be static, and 30% should be dynamic.
Constances notes from the same session:
First off, if you do not know who Rich Vogel is, stop reading this
and go google him. He is a very cool guy. Vexed a bit on the issue
of user-generated content, but a very very cool guy. Vogel began
by stating, "Hollywood tells stories. We create "living
experiences." He then did a pro and con of both static content
(designed-in experiences) and dynamic content (user generated
content), arguing that the right mix is roughly 70 / 30. He sees
the future of MMOGaming in things like: mechanisms for players to
change the world, systems that are not geared toward killing
(e.g. SWG's entertainer class), and - most importantly IMHO -
toolboxes for players to generate their own content. Woot! My
only real issue throughout the entire presentation was in how
Vogel was defining "dynamic content." At first he explicitly
defined it as "player generated," but through a large part of the
talk he also seemed to be including any content that was
time-sensitive, such as dynamic points of interest that are
designed to be consumed only once, random brief encounters (again,
designed in but triggered by an individual's action), etc. Are
these kinds of content really 'of the same sort'? I see triggering
content as very different than authoring content: I might trigger
a scold for this article, but I didn't author it. And I might
author the article, but I didn't trigger it (the speakers did). Or
something like that... Regardless, the session was well worth the
time investment and thought-provoking for those who attended.
This was the party night. We started out at 'the green
thing'. (People in green tee shirts had been handing out green
invitations). The occasion was to celebrate that EA had decided to
fund USC, School of Cinema and Television, with 8 million dollars.
Eddo Stern was there with his friends from C-level. They do
interesting work in the intersection of art and games, their site
its well worth having a look at:
http://www.c-level.cc . Also Celia
Pearce (
http://www.cpandfriends.com/) was there, and she presented
me to Bill Tomlinson, (Assistant Professor, University of
California, Irvine
http://www.ics.uci.edu/~wmt/) who is interested
in similar topics as I, but with his background as a biologist he
attacks them from a different angle. Both Celia and Bill are working
with the Grid. The Grid is a Massively Multi-User Online Grid-Based
Game Server in support of education and research. Read more about it
here:
http://visservices.npaci.edu/gaming/. At the same place I ran
into Steffen Waltz and his boss Gerhard, who is setting up a game
education in Switzerland. They still haven't found the right person
for running it. After a while we headed for a party hosted by
Carnegie Mellon University, where we found Jesper Juul and Aki
Järvinen. (Who already posted his conference report, read it here:
http://www.gameswithoutfrontiers.net/gdc2004inreview.html it has
pictures!) Had a nice chat with some students from Carnegie Mellon,
they do collaborative game projects with many institutions
involved. The evening ended at the Fairmont, and now it becomes
pointless to count up names, because everybody was there :).
DAY 3
-----
I made it just in time for the Embodied Agents session:
Embodied Autonomous Agents
Speaker(s): Bryan Stout
It was a little bit of a dilemma, cause this was a roundtable I had
wanted to go to on the Wednesday (when I instead was on an aircraft
coming from Costa Rica), And I had originally planned to go to the
ArtModJam which Celia was hosting and where Eddo were going to show
their new game. But the day before I had run into the 3 guys hosting
the general AI roundtables, one of them being Eric Dybsand, whom I
had already met at an AI conference in Canada the summer before this
one. All three of them said that the embodied agent session was
really good, so I didn't want to miss it. The roundtable turned out
to be focussed on character animation, so we dwelled on methods for
contextual character animation.
Next I made my way to:
Triangulation: A Schizophrenic Approach to Game Design
Speaker(s): Will Wright
Which was a wonderful wonderful speech. Will Wright is a genius and
a hero! He described the different methods he uses when he gets a
'designers block'. And about Russian space crafts.
I go out for a smoke, and on the balcony I find Eddo and his
friends. Luckily I get a chance to have a look at what they
presented at the ArtModJam.
Then I hurry away to Michaels and Andrew's presentation of Façade.
Beyond Finite State Machines: Managing Complex, Intermixing
Behavior Hierarchies
Speaker(s): Michael Mateas, Andrew Stern
Anyone with the slightest interest in interactive story telling,
believable agents and conversational systems will find the project
well worth studying (url is previously put into this text).
Anna and I meet Katherine Isbister for lunch. Ken Perlin (NY CAT) is
supposed to come with us too, but comes running and says I needs to
fix a bug in a demo he is going to show to someone later the same
day. The three of us trots away to deli, and Katherine's friend
Robin joins us (she was the one who moderated the relevant research
panel). (K made a speech at the conference that I missed (10 tricks
from psychology for making better characters.))
The next session I go to is:
Storytelling in Earth & Beyond
Speaker(s): Chris Klug
Chris had been through quite an ordeal while being lead content
designer on Earth and Beyond, Westwood Studios' massive multiplayer
online RPG. The outset was to make an epic story and to sell a brand
new sf universe to a savvy audience.
Chris stated that in dramatic form, change in story is due to
conflict. Therefore, gameplay had to change as the players took
part in the story. In theory all people in the production team and
management agreed to this. The head writer left early in the
process, but before he produced a massive story line that they
decided to keep.
Chris and his co-workers broke the story down to three acts that
each consisted of 15 'steps'. A lot of effort needed to go into
each step in scripting and dialog writing for all the npc:s and the
causal dependencies within the story. The dialogues were written as
tree structures. To talk to the NPCs was the way for the player to
move forward in the story. The story would to the player be as a
puzzle, fragment by fragment players could get an understanding of
the whole picture. The story had shades of grey instead of black and
white, no obvious 'good' or 'evil' characters. For newplayers
getting a grip of the story they relayed on the common player memory
that 'older' players would explain to the new ones. There was an
idea to let the internal (dramatic) driving forces inherent in each
of nine different player factions have a role in the story. Each
faction was defined by a certain view of reality. Sadly the story
never got the chance to develop to that point.
This session made me sad - the team didn't have the right tools in
order to keep up the production rate (not even a dialog-debug tool!
*sob*). Strange decisions from the management forced them to rewrite
the first part of the story five times, and the live content
updating was shut down before the story was ended. And the whole
game closes down the 22nd of September.
To me it seem like they wanted to author the whole story statically,
but make it appear as emergent. The amount work in order to use
similar methods for storytelling as used in single player game but
do it in a MMOG would be insane; this would be the illustration for
it. It's clear that other methods are needed. What also is clear
from this production though is that players really appreciate a good
story, the handcraft of storytelling, deep/rounded characters and
shades of grey.
But an interesting note is that when they had an NPC lying, the
players wouldn't perceive that as a lie, but assumed that the NPC
was bugged. I put it down to personal player histories of gameplay
experiences where false casual relationships in stories have
resulted in dead ends.
Post conference note:
E-mail correspondence with Chris: he thought that this description
well encapsulated the situation.
The last session I went to was:
Practical Game Theories: Academics Fragging Developers
Speaker(s): Gonzalo Frasca
Some of the topics discussed were what type of research people from
the industry would want: tools for more efficient game production,
and again, internships in game companies for students.
More from Constance's notes:
Designing a Massively Multiplayer Game
Raph Koster, Rich Vogel, Gordon Walton
This full-day tutorial, led by the infamous UO/SWG trio Koster,
Vogel, Walton, targeted developers who are at the very beginning
stages of starting an MMOG or even earlier - at the point of
merely considering it. A few take-homes were clear. First, it is
UNREAL the amount of resources it takes to create an MMOG! Quite
honestly, my respect for MMOG developers has now shot through the
roof. Second, the differences between single player and
multiplayer games are so profound and extensive that one wonders
if they can even be classified under the same rubric "games." In
a single player, launch date means job done. In a multiplayer,
launch date means shift in type of work but not in quantity. In
single player, the gamer is basically a consumer. In multiplayer,
the gamer plays several roles that shift as the product move from
beta to published. In Koster's own words: "Why is your game an
mmog? Because if there is any reason that it doesn't need to be,
then it probably shouldn't be. If you don't have a passion to do
online games in particular, don't do it. If you're not in love
with the potential of online, then don't do it. because you'll
fuck it up." Other interesting that were touched on: Possible
market transition to a micropayments model rather than
subscriptions (e.g. exchanging real money for items rather than
time). Mocking up the game system in detail via design documents
is tedious, next to impossible, and crucial - this is where SWG
forgot to include 'ctrl+R' to reply to last tell message and now
EVERYONE bytches about it. Developers have to negotiate between
money people, other developers, and insatiable gamers. Not
easy. When asked why he still does it, Koster responded: "The
things that players accomplish, sometimes despite us - those are
the things that keep you excited about it. It's such a wonderful
feeling to know immediately that you are making contact with
people & effecting their lives... the stories are endless.
Absolutely, its worth it."
Notes: oh yeah. mucho notes:
http://www.sit.wisc.edu/~steinkuehler/J101/designMMOG.rtf
Presentation:
http://www.legendmud.org/raph/gaming (coming soon)
MMO Communities: Fans and Flames
Kevin O'Hara
This roundtable, despite its more general title, focused primarily
on MMOG bulletin boards. The discussion began with a description
of how the community function and facilitation has to change when
the game shifts from prelaunch to beta to live. From there on,
how to create a constructive community and deal with gripers was
the central topic. Despite O'Hara's well-received closing
comments - "Some say that players don't know what they want. I
contest that. I think gamers know exactly what they want. They
just don't understand the tools and constraints we work with. So I
would urge you to take your player base seriously." - it seemed as
if the conversation dwelled far more on the community stinkers
rather than the community builders, with the developers present
sharing horror story after horror story about their official or
unofficial boards. Some of the stories were sufficiently
traumatizing for me to sympathize with those running the games,
yet the fact that these communities are oftentimes
self-monitoring, free advertising, and pro-developer seemed
entirely left out of the equation. As the SWG bulletin board
community manager, O'Hara is both experienced and articulate, but
after a while the roundtable began to feel a bit like an MMDA
meeting (Massively Multiplayer Developers Anonymous) - all the
designers bonding with each other through troubles-telling and
player-bashing with little fair representation of the MMO
communities themselves left in the equation at all.
Notes:
http://www.sit.wisc.edu/~steinkuehler/J101/MMOGboards.rtf
Presentation: n/a
MUD Dev Dinner:
---------------
First thing I needed to do was to find Constance. As said earlier we
had planned to write a review together of the MMOG related stuff
happening at both conferences, and had originally planned to meet on
wednesday morning but yes - you know - Costa Rica.
This proved to be a challenging exercise, since it was in the last
desperate hour of the GDC and everyone running around in a final
social race. After an hour of herding a constantly dispersing group
of academics we finally ended up in the bigger group - the
Muddevians. Our destination was a restaurant that served sauerkraut
and big pieces of meat and I ended up (by accident I swear) flanked
by three persons programming on WorldForge, an open source MMOG
engine (
http://www.worldforge.org/), I heard some beautiful
descriptions of Atlas, the middlewaresystem/client/server-protocol
but also a full-fledged ontological description of an abstract
world.
After we had been fed the evening dispersed into floundering about
talking to people, among them Dave Rickey and J C Lawrence. Dave
writes regularly for Skotos.net
(
http://www.skotos.net/about/pr/06032003b.html). JC (owner of the
MUD Dev list), gave by a frequent use of the word zipf curve a brief
history of the list. One of his concerns for the MUD Dev list was
that some threads tended to end very abruptly, and he saw a clear
correlation of those abrupt endings to issues of intellectual
property, since many of the list members work in big
corporations. In another corner Raph Koster (creative director of
SOE, made StarWars Galaxies and Ultima Online) was holding court,
giving a session of the whole history of MUDs to a group that grew
bigger over the evening. After staying as long as we could at this
Germanic outpost in California we ended up at the Fairmont again,
where I spent most of the evening talking to Eric Zimmerman
(Co-author 'Rules of Play', 2003), who was already planning for the
next years GDC game design contest.
MUD-Dev Conference DAY ONE
--------------------------
The day started out by Andy Tepper (Teppy - creator of and pharaoh
in A Tale in the Desert, one of the most interesting MMOGs
http://egenesis.centralserver.net/ ) and myself getting caught by
the police at the shuttle to the convention centre. I had rushed
onto the train with the idea to pay on board, and Andy had not
succeeded to feed the machine with twenty-dollar bills. I felt like
I was in an American TV series, the police said things like 'I could
do anything to you but to shoot you!' And he was wearing one of
those uniforms frequently exposed by media. I wanted to take photo,
but I didn't dare to, also I was busy apologizing so that we
wouldn't have to go to jail :).
The first session was by one of the lead programmers of WorldForge:
Designs for Continent Scale Persistent Online Worlds
Speaker: Alistair Riddoch
Alistair presented the terrain generation system he built within the
WorldForge project. He also gave a brief history of the WorldForge
Project, which started six years ago, but, fell apart due to people
having different goals, and then reformed. One of the goals behind
the terrain generation system was to not have big chunks of data to
download. They chose height fields early on for ease of editing and
for ease of downloading and rendering. They wanted to generate the
terrain data on the client. He described pros and cons of different
algorithms, one with diamond square interpolation (with in inherent
problem of change: change of one control point would affect 16
tiles), and square interpolation. A neat thing is that the code is
identical on both the client and the server (phew).
Post conference notes:
Alistair says via emial: Our Mercator terrain system has been
written as a lib, and is available immediatly from our cvs
server. Details of how to get anonymous access to the cvs server
can be found somewhere on
http://www.worldforge.org/
The second speaker was:
Hierarchical World Generation
Speaker: John Arras
John described how he goes about to generate content for text
MUDs. His motivation for building this is the unpredictability of
the content, the fact that people can change things - he meant that
its easier for authors to let go of content control if the content
is generated rather than authored:
http://www.sf.net/projects/genmud/
In his system there are whole societies of NPC:s that not only live
in certain areas, but also alter the areas (this was presented at
last years MUD Dev Conf). John showed code examples of the little
code seeds that in the system recursively grows to become cities
with buildings of different kinds and sizes which are populated by
items and NPC:s. (He mentioned a book that had inspired him: 'The
algorithmic beauty of plants'). John also showed the result of a
simulation of generating building, which made me associate to Karl
Sims classic simulation. John wanted see more elements of RTS in
MUDs generally.
At lunch I ended up at the same table as Justin Lloyd and Sharon
Volin, both developers but there in the role of journalists,
covering the event for GameDeveloper Magazine. Justin kept video
taping the event, and said that he might edit a little conference
video of the first day. Randy Farmer was the fourth person at the
table, and I asked him about what community system in existing MMOGs
he found best. He put his vote in for Asheron's Call. There was a
trolley with desserts on it.
Full of food we sat down again to listen to
Developing independent games panel
Members: Brian Green, Andy Tepper and Daniel James
Brian Green - Meridian 59
Andy Tepper - A Tale in the Desert
Daniel James - Puzzle Pirates (winner of many awards at the GDC ceremony)
The panelists started out by remembering GDC four years ago: 'If you
don't have $20M dollars, don't do it!'
Then a quick recap of how they did it anyway.
- Andy had other business that he used to fund the development of
ATITD. He spent 1.1 M during 4 years and had good help from
students at the local university. (But only for the art, not for
the code 'like volunteering to be an apprentice diamond cutter')
- For Meridian 59 3DO paid for most of the development cost. After
2 years of work it was shut down. Brian Green (and I'm not sure
who else was involved) bought it off 3DO. During the production
they got help from fans, for example one player provided the
hosting service. They used what they had in art assets, and BG
lives a frugal life: only made 12k last year.
- The team making Pirates had a similar approach as ATITD. They
put in their own money, in total around 750 k. They didn't use
interns in order to protect the communication in the small team of
six people.
This was followed by discussions about use of interns, health
insurance for employees and revenue models.
Constance's notes:
Developing Independent Games
Brian Green, Andrew Tepper, Daniel James
For those of you who don't know these developers: Brian is the man
behind the resurrection of Meridian59, Andy developed the MMOG "A
Tale in the Desert," and Daniel wears a pirate hat and is the
genius behind "Puzzle Pirates." If you haven't played these
titles, go try them out - there could not be a more diverse
collection of games, though all three are well worth playing. The
three panelists discussed their personal sacrifices to raise money
for the games, how to make a game work on a shoestring budget,
what the production pipeline is like for small development teams,
and their takes on how to relate to gamefans. For me, the
highlights of this panel discussion (other than hearing Brian
speak for the first time, which I had long been looking forward
to) was when Andy admitted that he actually put his REAL mobile
phone number on his game's website.
Andy: I've only had 6 inappropriate calls to my number the
entire year.
Constance: OMG I only run a pledge, not a game, and I've had
more than 6 inappropriate calls in the last year and my home
telephone number was never posted anyplace. o.0
Notes: yes but sparse
http://www.sit.wisc.edu/~steinkuehler/J101/muddev_indegames.rtf
Presentation: n/a
Next was:
Providing Resources for MMOG Guild
Speaker: Constance Steinkuehler Leaders
Constance described the practice of work for being a guild/pledge
leader in Lineage; the amount of work is massive. She pointed out
couple of things that developers could offer or sell in order to
support guild leaders. The benefits for the developers are that they
first of all benefit from the work of the guild leaders, since they
work as game masters, keeping players in game, solving conflicts and
adding content by leading and motivating group activities. As an
example she described the local mythos developed around a game item,
an orb, which no longer had a function in the game, but that was
given a new function. Instead of 'gaming on top of the story' they
created their own story around it. Constance requested server space
and website templates and pointed out that players would be prepared
to pay for this. It would also give the developer a chance of
monitoring/controlling the content. She also requested that in game
data about guild members could be available to guild leaders.
(Recommended Jenkins book 'Transmedia Storytelling and Fandom')
Immediately after the audience pointed out examples, as the web
tools at guilds.com and the available user data in DAoC. The centre
of the discussion was around player created content. On one hand the
fear for the lack of quality, and on the other hand examples of the
opposite, like Second Life where players can model own 3D
objects. Also the problem of legal issues was pointed out, for
example having player created music to be played in SWG turned out
to be a legal minefield and thus too much of a hassle to go through
with.
Constance's notes:
Providing Resources for MMOG Guild Leaders
Constance Steinkuehler (that's me)
The main point of my talk was that guild leaders constitute a new
market & a new resource, and that developers can take advantage of
this by building & selling them tools to support their work -
hence, getting greater profits off of them while, at the same
time, having more control over what they actually produce since
what they produce would be shaped/constrained by the tools they
provide. I think that point got sort of lost, though. Instead, the
post-powerpoint discussion focused only on my assumption: that
'user-generated content' is possible & good. So, my hangover and I
ran with that, arguing that the real IP issue is not merely a
'legal' or 'economic' one but a profound cultural one of who is
going to 'own' our cultural products - media conglomerates (e.g.,
Sony) or the people for whom that content is meaningful (e.g.,
fandom)? Thank you, Henry Jenkins, <link to HJenkins site> who
made that argument in a talk a few years ago <link to Textual
Poachers_ much more persuasively than I could.
Developer in Audience: Your talk seems very fan-centric...
Constance: Why yes. It sure is. I'll stop being fan-centric when
leaders in the field like Richard Garriot stop publicly staying
things like:
Giving away you IP is a VERY BAD THING. A VERY BAD BAD
THING. DON'T EVER DO THAT *spoken dramatically with hands
waving at last year's E3*
Notes: no
Presentation:
http://www.sit.wisc.edu/~steinkuehler/papers/SteinkuehlerMUDDEV.pdf
The next session was:
The future of Cyberspace economies
Speaker: Edward Castronova
Constances notes on Castranova:
Castranova's talk struck a remarkable balance between focus on
virtual world economies and focus on people's well being,
regardless of which world they're in. He began with an
interesting discussion of the human drive for play, then moved on
to a discussion of virtual worlds as a form of cultural criticism,
asking "For how many people do virtual worlds compare favourably
with life on Earth?" Later, Castranova discussed the broader
implications of growth on virtual culture/worlds, such as the
migration of labour supply and consumption to online spaces, and
the idea of the avatar as an extension of the body that, in fact,
has the potential to raise people's well being. I'm not sure what
I expected from an economist, but Castranova's work seems to me a
really innovative and well-received balance between focus on the
economics of virtual worlds and the economics of play / well-being
/ human value. If you think his academic papers are impressive,
try his presentations.
Notes:
http://www.sit.wisc.edu/~steinkuehler/J101/muddev_castranova.rtf
Presentation: (ask Ted)
And closing the day:
Habitat Redux
Speakers: Randy Farmer & Chip Morningstar
Randy is now strategic analyst at Yahoo! And Chip is architect at
Avistar Communications.
As I perceived it, this session was originally planned to be divided
into three sections, 1) Lessons of Habitat, from 1990, 2) New
lessons from Habitat, from nineties and beginning of this century,
and 3) What now. Due to time constraints, this was the last session
of the day; we never really got to section 3. But here is a recap of
what was said. 1990 Randy and Chip presented the paper 'The Lessons
of Lucasfilm's Habitat' at The First Annual International Conference
on Cyberspace. It was published in Cyberspace: First Steps, Michael
Benedikt (ed.), 1990, MIT Press, Cambridge, Mass. The paper is
available at:
http://www.scara.com/~ole/literatur/LessonsOfHabitat.html
Habitat was done 1985 and ran on networked commodore 64
computers. This was ten years before the Palace and ActiveWorlds,
and of course, pre-Internet. The lessons are in short
- A multi-user environment is central to the idea of cyberspace.
- Communications bandwidth is a scarce resource.
- An object-oriented data representation is essential.
- The implementation platform is relatively unimportant.
- Data communications standards are vital.
- Detailed central planning is impossible; don't even try.
- You can't trust anyone.
- Work within the system.
- Get real.
All this is fleshed out in the paper. New lessons are:
- You can't tell people anything. (When an idea is new and not yet
implemented, people will make an interpretation, and understand
what they would like the idea to be.)
- Beware the platform (its easy to become obsessed by the platform
and its tools)
- Its business (People that invest are not interested in awesome
stuff, they are interested in money)
- User generated content is not free (It costs to filter)
- First movers get slaughtered
- Program small things, one at a time. (Plans will change anyway)
- Smart people can rationalize anything (and give solutions to
problems that not yet exist)
- Compartmentalize (Its lethal to apply boxed single player game
architecture to MMOGs. Use the best distributed object model you
can afford. Don't think client/server, think 'division of
labour'. 'Threads are evil.' 'State machines are for ppl who
doesn't understand programming')
- You can't trust anyone so you may just as well give away the
whole source code.
This is as far it got before the time ran out. The final
exclamations I noted was 'Stuff changed!' and 'It's time to build!'
I oh so much wanted to hear the continuation of this. (argh)
Post conferences notes:
I posted a question on MUD Dev and this is the answer from Chip:
> at the MUD Dev Conference about a week ago Randy Farmer and
> Chip Morningstar in the end of the first day held a speech
> with the title Habitat Redux.
> Due to time constraints (we had to leave the conference room
> and > hop into cars to go for dinner) the speech was kind of
> cut off.
Unfortunately we got squeezed between some
too-fascinating-to-cut-off audience discussions on one end and
an unexpectedly-eager-kick-us-all-out hotel staff on the other.
*sigh*
> I'm so curious about what the end of the speech would have
> been, > since Randy and Chip also proclaimed that this was the
> last time > they would speak about Habitat.
I don't think we said this was the last time we'd talk about
Habitat, just that it was time to look forward.
> What was it that you would have said there, in the end?
Y'all will just have to wait for the written version, which
we'll put up on the web just as soon as we're done writing it
(Real Soon Now :-)
Until then, you can make do with a copy of our PowerPoint slides
at:
http://www.fudco.com/habitat/
> All this is fleshed out in the paper. New lessons are:
...Mirjam's partial summary elided...
The quick summary is:
-- You can't tell people anything
-- Beware the Platform
-- It's a business
-- Resilience is better than planning
-- Smart people can rationalize anything
-- Security is a design discipline
-- Capabilities trump User Identity
-- Compartmentalize
-- Use the best distributed object model you can afford
-- Don't think Client/Server, think Division of Labor
-- Threads are EVIL
-- You can do it without Shards
-- Incent users to crash your world, early & often
-- Don't punish people for following the rules.
The full exposition will be coming, we promise (really!) but I
do want to correct one point that got a bit garbled in
translation there:
> - You cant trust anyone so you may just as well give away
> the > whole source code.
Before anybody jumps on us for this, we were definitely *not*
advocating this. Not that there aren't sometimes virtues to
giving away source code, but that's not what we were arguing
for. What we were advocating was that you publish your
*protocols*. The kiddies will reverse engineer them anyway;
publication provides a kind of partial immunization against the
sort of marketing solipsism which fantasizes that you can attain
security by keeping a secret of the way things work.
> This is as far it got before the time ran out. The final
> exclamations I noted was "Stuff changed!" and "It's time to
> build!"
Time to build, indeed. When we started talking about this stuff
15 years ago, there weren't many other folks doing this sort of
thing so what were saying kind of stood out. Now there's a
whole community of community builders out there, many of whom
have as much or more experience at this as we do, so we're
really getting more interested in building than in pontificating
(not that we're likely to stop doing the latter -- we may not
always be right but at least we're opinionated!).
> I oh so much wanted to hear the continuation of this.
Hey, us too! I didn't get a chance to use some of my best
jokes. :-)
Chip
Andy Tepper and I got a ride to Dave & Busters by an anthropologist
from Stanford, Richard. After dinner and a couple of drinks (with
loads of fruit in them and described as little parties in glasses by
the menu) we went down to the arcade part of the place, which was
huge.
MUD-Dev Conference DAY 2
------------------------
This morning I came rushing again as the train came clattering, but
luckily Andy saw me running and bought me a train ticket from the
picky machine, phew. After a stop at starbucks, filled with vanilla
latte and espresso, we sat down to listen to
Designing a persistent-world server for flexible rule development
Speaker: James Turner
James is the other :) lead programmer of WorldForge, and his speech
concentrated on scripting languages in general and on scripting in
the WorldForge Project in particular. He went through the pros and
cons of some scripting languages and motivated his choice of java
script. He went into issues of top-level design, protocol isolation,
standard messages, implicit updates, perceptions, storage
virtualization and putting it together. James gave the advice to
decide in advance on what to make scriptable, and to chose one
language and stick with it. The discussion afterwards turned into a
mapping of what scripting languages that was used by people in the
audience, and stances to scripting in general. Raph said that they
used python when they made UO, java for SWG, but that the movement
in SOE was to not use scripting languages at all. In Second Life,
for security reasons only 5% of the CPU is used for scripts, the
scripts being independent of rest of the system. Skotos - tried
three different scripting languages. Meridian 59 - single threaded
custom scripting language, which proved to be efficient for
production. Tess said that if a custom scripting language is to be
built - avoid to be too clever, a new idea can often prove to be a
bad idea. Andy said they used 2.5 years making the scripting
language, and after that 2 years building ATITD, and it made it
possible to create an enormous amount of deep content. Raph pointed
out dangers of scripting language for large production teams - for
SWG it was a bad thing that java was brought in since people went
outside the script design, important to segment the access to the
power. And that python, as being totally object oriented nested
objects turned into nightmares. Skotos - good thing to have all
objects of a single class, easier to update and add, flexible.
Due to a change in schedule Nicole Lazzare hopped in and gave the
same speech as she gave at GDC:
Why We Play Games: The Four Keys to Player Experience.
Speaker: Nicole Lazzaro
Description from the GDC site:
Level completion, high scores, and cut scenes do not make a game
great. Players want more than something to think, see, or
do. What players value above all else are their emotions: their
hopes, their fears, and their dreams. Entertainment that taps into
who they are offers a more compelling and enchanting player
experience.
Why do we play games?
1. Emotion
2. Challenge & Strategy
3. Immersion
4. Social Interaction
This class unearths recent treasures on how to create exceptional
entertainment through doing. It covers the Four Keys to creating
player experiences from XEODesign's independent research of over
30 core, casual, and non-gamers; and our experience in designing
casual entertainment such as Cosmopolitan Virtual Makeover® 2,
Totally Mad (the Mad Magazine archive), and Shockwave games such
as for ChevronCars.com. Additional insight on human computer
interaction and the psychology of fun and emotion comes from
research of Piaget, Norman, Ekman, and Csikszentmihalyi. This
class provides tools developers can use to understand why we play
games and how to pump up the emotions in them without relying on
pre-scripted sequences. If you have ever wondered why people play
games, and what makes players angry, fearful and joyous, the
difference between a social smile and real enjoyment, or how to
get more emotions into your game without relying on story attend
this talk.
The 46-page paper that was the basis for the presentation is
available in the GDC proceedings.
JC announced a moment of reflection on the conference:
Things discussed included:
- How to market the conference? (If to market?)
- Difference from the Austin Conference. (Which is focussing on
MMOGs. Raph said that the Austin one was more teaching oriented,
and therefore bigger, as opposed to muddev as being more
research oriented and naturally smaller)
- This years high presence of academics at the conference.
- What happened to the people that came three years ago?
- The exclusiveness of the group, and the hassle to update all
new people on 20 years of history.
- Is it froufrou?
Round table:
At each table: five people + paper + colour pencils + topics
The fact that people moved table after half time, after quite deep
discussions, created a good discussion climate.
We moved onto the barbecue at Jon Leonard's house. Thanks for that,
it was a great evening!
... and thanks for the whole conference!
//Mirjam Eladhari