[This is a draft I wrote for my column, P’s&Q’s in ACMs interactions magazine. For it, I interviewed Bob Moore who was at the time with now defunct Multiverse. It came out in Volume 15 Issue 3, May + June 2008. The final version can be found here.]
Chatting with virtual world researchers Jeffrey Bardzell and Shaowen Bardzell, I found out that a seriously desired artifact in Second Life, not unlike in First Life, is hair. Swishy, shiny hair. But there is one problem with this real-dreamy hair: Such hair is computationally costly compared to render in comparison to your average avatar body. And so, sadly, your avatar body arrives before your hair. For a matter of moments, no matter how fashionable the outfits, everyone is bald as coots.
And it’s worse than that. Often from other players’ perspectives, your hair has failed to “rez,” but not from your perspective. So you think you look hot-REALLY hot–until, that is, some newbie says, “Why are you bald?”
There are now apparently a lot of people routinely spending time hanging out in virtual worlds. MMO Crunch reported 36 million regularly active MMORPG players in August 2007. And they are getting more popular.
“Virtual worlds are fundamentally a medium for social interaction. One that takes the face-to-face conversation as its metaphor. As such it leverages users’ common sense knowledge: I see a humanoid avatar and I know that if I want to talk to that player, I should approach his or her avatar with my own” says Bob Moore interaction researcher and designer at Multiverse. As with all metaphors, however, there are fractures in the metaphor. Which matter and which don’t? When do we happily immerse and feel like we are really there and when do we get amusingly or annoyingly jolted by something that doesn’t work? What are key things designers are thinking about right now? Bob and I chatted for a while, reflecting on the work that has been previously done on the research areas of graphical worlds, collaborative virtual environments (CVEs) and online gaming environments. Given my own background in researching text-based virtual worlds and animated interface characters, I was intrigued to her Bob’s perspective. Here are some of the main points we covered in our conversation:
(1) For virtual worlds, user interface design is also social interaction design
Bob argues that one thing UI designers for virtual worlds have not yet fully grasped is that when users are interacting with the system, they are often at the same time interacting with other users. When standing avatar-to-avatar, if you ask me, “Did you get the sword?” and I promptly open my inventory, that UI action is a relevant part of the interactional context. Or if I say, “Let’s go” and you promptly open your map, that’s a relevant next action that I should know about. In almost all current virtual worlds, opening your inventory or your map triggers no public cues, only private cues for the individual user. So you not only need to give the individual user feedback about what the system is doing, also need to give the other players feedback about what the individual user is doing. In other words, users’ interactions with the system should be made public.
(2) Where avatar bodies are not like physical bodies
Unlike real life, most people tend to play virtual worlds with their camera view zoomed back so they can see their avatars rather than in true first-person view where you can only occasionally see your hands or legs. There are a couple of good reasons for this. Computer screens don’t allow for peripheral vision. But pulling back the camera can help mitigate this limitation by widening your field of view. Similarly, as Bob points out, avatars don’t allow for proprioception, or our awareness of the positions of our body. Zooming back the camera also helps players deal with this fact. I may think my avatar should be waving like the Queen of England or winking flirtatiously at someone else because I typed /wave or /wink, but it’s hard to be sure if I can’t see my avatar. Maybe I mistyped the commands or maybe the animation associated with the command actually looks more like a New Yorker hailing a taxi than Her Royal Highness.
(3) When the conversation lags behind the avatars
Despite a lot of the interesting audio experiments that were being conducted over 15 years ago–see Benford, S. D. and Fahlén, L. E., Awareness, Focus, Nimbus and Aura – A Spatial Model of Interaction in Virtual Worlds, Proc. HCI International ’93, Orlando, Florida, 1993 as a good example–most virtual world conversations take place through chat. The avatars may be wandering about gesturing and wiggling their hips, but chat does not come out in audio from the mouth of the avatar. One problem this causes is the discontinuity, the lack of congruence, between action and uttered words. Bob points out that typing a chat message is another kind of action that other players should know about, and he recounts a case in which a team of players is about to attack a group of “mobs,” or computer-controlled opponents. While one player is composing a question about how the team might change its tactics, a fellow player initiates combat. The tactical question then publicly appears too late. The pseudo-synchronous chat lags behind the synchronous avatars. Bob, who was trained in Conversation Analysis, explains that a key feature of real-life conversation is that you can hear a turn unfolding in real-time. This enables you to do things like determine who should speak next, anticipate precisely when the turn will end so you can start your next turn with minimal gap and overlap, and even preempt the completion of the current speaker’s turn if you don’t like the direction its going. In other words, the ability to monitor other people’s turns-in-progress is a requirement for tight coordination in conversation. Most virtual worlds (with the exception of There) use IRC- or IM-style chat systems, and therefore, do not allow players to achieve this tight coordination among their turns-at-chat and avatar actions. The result is an interactional experience that feels very unnatural (at first) and which motivates players to invent workarounds to the system.
(4) Perennials of place
One of the amazing things about virtual worlds is how quickly we get a sense of being co-present in a place with other people, even though it may be an image on a screen, a world into which we are kind of peering. And, as in the real world, ambience is created by building and room size and scale in relation to crowd size. In my research on MUDs/MOOs with Jeni Tennison and then later work with Sara Bly, we found that even very simple text exchanges in textually described “rooms” can make dyads and groups feel co-present, imersively in the virtual world together. Humans tend to get engaged with each other as long as there is some consistent chain of action-reaction. Indeed some analysts would argue that turn-taking in conversation is the fundamental unit of human communication and connectedness.
As we explored these concepts, Bob described a comparative ethnographic study he’d done of bars and dance clubs in multiple virtual worlds. He hung out in these social public spaces and analyzed features of the design that impacted the success of the space as a social environment. One key feature of club design is size. Bob discovered that, while construction in these worlds is cheap compared to real life, it is more difficult to fill these spaces with people than it is to fill the real-life urban centers that researcher William H. Whyte examined. As a result, the dance club in City of Heroes and the majority of player-built clubs in Second Life are simply too large. They feel like an airport terminal or concert hall rather than a corner pub.
So in order to achieve the kind of social density necessary for a vibrant social space, or “third place” as academic Ray Oldenburg would call it, designers should make virtual bars and clubs much smaller than they currently do. The most successful virtual third place that Bob discovered was a Second Life bar that was intentionally tiny. In order to get into the place, you had to “rub elbows” with other patrons. The place felt “busy” with only five players and “hoppin’” with twenty. And everyone was within everyone else’s chat radius, which facilitated the public conversation. In other words, lessons from real-life urban design appear to apply in several ways to the design of virtual public places.
(5) A range of skill sets and a modifiable world.
There are challenges in designing a world where “newbies” or newcomers can learn the ropes. If you want to learn about interacting in one, you simply have to get off (or rather for most people onto) the sofa and get in there. Go in-world. It is much easier to learn in-game than to learn out of game, just like learning to play golf requires you take up a golf club and try it. You just can’t learn by watching someone else. Bob says, “I’d recommend getting into a pick-up group and go and doing some adventuring”. There are plenty of folks in-world who are willing to help out, to show off their knowledge. Bob agrees that not everybody likes to help, and admits that whether you are more of less likely to be helped depends on some other factors……you guessed it-having an attractive female avatar means you are more likely to get help.
There are other things to learn aside from interactions and activities. Many virtual worlds allow people to buy, build and exchange things. Second Life is perhaps the primo example, it is a sandbox with a constructive geometry, which enables them to stream your data and create your objects on the fly. But that also means people arrive and build stuff and leave it behind. There is a certain “I just learned to build today” look that Bob identifies as one of the main scars on the aesthetics of the virtual world. Despite the parallels that some people make between the real live music/art/dance festival known as the Burning Man Project and playful interactions and explorations in Second Life, there is no motto inviting us to “leave no trace” in Second Life. Of course the ecological consequences are somewhat different but the visual aesthetic of clutter and detritis is experienced as the same for Second Life afficionados. So, there is a tension between giving people freedom to build anything they want and making sure the world doesn’t end up looking like the aftermath of an afternoon in a Montessori School for Gremlins. Beyond clean-up though, there is another point. At this juncture the building tools are much better than they were but they still leave a lot to be desired. I say: If we want Gaudi not Brutalism, we need to provide better tools to scaffold the building endeavours of the folks in-world.
To get an understanding of issues like these above and push on understanding how people really experience these virtual/online places, Bob advocates a close and detailed analysis of what is actually going on as it unfolds in real time, looking at patterns of action and interaction, how those patterns develop and are understood, learned and evolved, and identifying patterns that are persistent and prevalent. Too many people have theories about what is going on that are based on something completely external to the situation. “You need to get close to the phenomenon/experience” says Bob. And we need to “be concerned both with the in-world simulation of face-to-face interaction AND the usability of the interface for puppeteering the avatars and interacting with the system.”
By looking at the challenges in interaction that people routinely encounter and work around, it is possible to ask how important – or disruptive to interaction in-world – those challenges are, and propose ways to address them through interaction, interface and system (re)design. I could not agree with him more.
And on that note, it’s time for me to get my hair on and go build a shack.
Reference: Bardzell, S., & Bardzell, J. (2007). Docile avatars: Aesthetics, experience, and sexual interaction in Second Life. Proceedings of British HCI 2007. Lancaster, UK