« Meta Grinding and Nemo | Main | Sticky NeoPets »

Jun 18, 2005

Comments

1.

A quick question, and not much feedback: Did you take this name from Rene Margrite's painting? Delicious, if so.

A stimulating discussion must surely follow. Here's my two cents:

I've never played a MUD or text-based multi-user online game. The only things I've been exposed to are three-dimensional games, and I feel that if I were to switch completely to text-based games, I am positive that the experience would not be as entertaining or as engrossing as a "fully realized" world. I also find interesting parallels to Dungeons and Dragons--a very social and, in some ways, immersive game, but at the same time, very flat and two-dimensional. I saw Johnson on the Daily Show the other night and what he said seemed to tie a lot into this sort of discussion: it's the INTERACTION with something tangible that has caused everything "bad" (or mainstream/popular culture) to become "good" (rising IQ test scores, for example) for you. I'm left wondering, then, is D&D (or tabletop games) on the same level as the MUD, or is it a step higher? (Or any other variation)

2.

It would be wrong to talk about this topic without referencing Magritte in some way :-)

3.

I've long been a fan of Magritte's work as well.

"Ceci n'est pas un pipe," indeed. :-)

As to the assertion that "Physically simulated 3D worlds are fundamentally different from text worlds", while I might quibble with the word "fundamentally," I'm not sure many people are going to disagree in any substantial way.

> kill dwarf with knife
The dwarf disappears in a puff of greasy black smoke.

Compare that experience from an early text-mode game to the experience of "firing" a BFG 9000. You're no longer just being told what happens -- you get to see pretty weapon fire effects, simulated blood spraying and body parts flying, and positional audio coming out of your 5.1 surround sound system that shakes the room.

The former experience is evocative. The latter is visceral.

Writers are often told to "show, don't tell." A text game tells. A graphical game shows.

But I still fire up Crowther and Woods's "Original Adventure" from time to time.

--Flatfingers

4.

Analogies:
- text game || 3D game
- newspaper || TV news
- fantasy novel || anime

Paradox:
I believe there's a study that says people who read news get more than those who watch news. Do text-gamers get more (because they have to process more - a more subjective experience perhaps)?

Good point:
However, you (Cory) are right about the shared (piano) bit - people may intepret texts differently but their interpretations of images should be more or less the same.

5.

I agree with the proposition (from the argument concerning collaboration spaces, mostly) , but I think a couple of arguments need challenging in detail.

Thus far Cory: Sure, text games could have a beautiful description of a sunset that could only be reached after reading a description of racing several competitors to the beach and another description of the long and dangerous climb to the top of a cliff, but claiming it is the same experience as driving a simulated car against other people and then fighting them up the cliff face is just silly.

With all respect, you're setting up strawman arguments here. A 3D world can just as easily have static objects that represent your "competitors" as scenery, in the manner you describe text games as doing, and text games can (and do) allow real competitors to engage in real competition using their (textual) avatars. You can't race a shopkeeping NPC any more than you can race a room description, but other players have the same abilities and inclinations as you do.

If MOOs (and the LP codebase rather than Diku) proved one thing, it's that nothing -has- to be scenery. Flat-panel backgrounds are a shortcut, not a law of nature. Any sufficiently complex assemblage of, well, Stuff can in theory support physics-and-engineering systems - look at what you can do with real-world Lego. The experience of doing it is different, especially from a real-world usability point of view, but the processes involved for an immersed player are similar.

Thus far Flatfingers: Writers are often told to "show, don't tell." A text game tells. A graphical game shows.

Bad text games tell, bad graphical games tell. Bad writers make bad games. Good text games show, in the literary sense - "don't impose your opinions on the reader, don't rub their noses in facts, give them some hints and let them work it out and integrate everything for themselves".

But it's also possible to see "show" as "here's an object, this is what it's like, take it or leave it" whereas "tell" would be saying to them "This is what the object is like, imagine it for yourself", and it seems to me that this is how you're trying to use it. However, we're looking at different levels here. You aren't showing players an object as distinct from telling them about it, you're showing players a 3D picture of an object as distinct from letting them read what a good writer wants them to know so they can build a mental picture of that object. Immersion lets people treat that manipulable, interactable 3D picture as though it were a real thing, and that's good, but that level of immersion can be, and is, achieved in textual worlds too, with the same result. You "see" the "object" only in the same way that your world supports "seeing".

6.

Well, you've got some things right, and some things wrong I think. Are 3d worlds fundamentally different? Generally no, though they do allow for movement that seems much more analog than text does (although it is not actually analog movement, it is close enough for our purposes). Are they different? Of course. Anything that is not that exact thing is different.

One of your main assumptions has problems from the start. You assume that text is a subset of 3d. It makes as much sense to claim that 3d is a subset of text, using the same logic you use. A "3d game" can have text just as easily as a "text game" can have 3d. Purely a matter of semantics. They are both methods of communicating information. They may be combined, but neither is a subset of the other unless you simply assume that one is a subset of the other.


Sure, text games could have a beautiful description of a sunset that could only be reached after reading a description of racing several competitors to the beach and another description of the long and dangerous climb to the top of a cliff, but claiming it is the same experience as driving a simulated car against other people and then fighting them up the cliff face is just silly.

This makes no sense. You are as much 'doing' something when you issue commands in a text world by hitting keys as you are when you issue commands in a graphical world by hitting keys or mouse buttons. In neither case are you experiencing climbing a cliff or fighting them up the cliff face. In both cases your brain is building images for you constructed out of data you receive. In neither case are you actually climbing a cliff or fighting up a cliff, at all, and the image your brain actually experiences is going to depend on the person involved. For instance, I may find that a text description of the result of my actions results in a much richer image in my head, whereas you may find that a graphical description of the result of your actions results in the richer image. Different strokes, that's all.

Of the four things you list, I think only #4 has any validity, but only in theory. I personally find that all of the graphical MUDs I've ever played feel extremely restrictive in terms of my range of expression compared to text. I'll grant that graphical MUDs have the potential to simulate face to face conversation, with the richness of body language and tonal variation and whatnot, far better than text MUDs, but I also think that currently none of them even come close to surpassing the range of expression available in some text MUDs precisely because of the need/desire of designers to show everything graphically.

Incidentally, a piano in 3d isn't a piano if you're playing it with a qwerty keyboard. In that case, all it is is something that can synthesize the tones of a piano given the correct input. That input could be graphical. That input could be textual. You could hook up a controller to it that plays just like an 88 key piano keyboard, and it still wouldn't make a difference what the screen displayed, because what matters with a piano is the input, which can be the same regardless of graphical or text output, and the audio output, which can also be the same.

--matt

7.

Flatfingers wrote:

Compare that experience from an early text-mode game to the experience of "firing" a BFG 9000. You're no longer just being told what happens -- you get to see pretty weapon fire effects, simulated blood spraying and body parts flying, and positional audio coming out of your 5.1 surround sound system that shakes the room.

Well first of all, the positional audio has nothing to do with 3d graphics. Positional audio could be done in text as well. 2nd, you're not seeing a gun fire when you fire a BFG in a graphical game. You're getting a representation of a gun firing. How close to the reality of a BFG firing your brain is willing to believe that representation is is going to differ from person to person. It is similar with text.

--matt

8.

Oh boy, yet another iteration of the medium/method debate.

Cory: Physically simulated 3D worlds are fundamentally different from text worlds.

This strikes me as a really bad way of framing this issue. Are 3D worlds different from text-based worlds? Of course! A better question to ask though is how are they different rather than to propose some claim that they are fundamentally different and we can effectively throw out our understandings based on what came before.

The entire point of looking at the continuity of history is that while technology changes (not as fast as we would like to think) human behavior tends to be fairly conserved. My favorite virtual gender-bender in history happens to be Marie-Sophie Germain, who in the early 19th century obtained university lecture notes and collaborated with Gauss by spoofing her identity.

Cory: #1 Place matters

Of course place matters! Barker and Schoggen figured this out in the 1950s with the concept of "behavior settings." Of course, we in the "social sciences" feel the need to reinvent the wheel every generation, so the notion that you get different behavior depending on if you stick a piano or a hot tub in a room sounds really radical now.

Cory in the linked article: Rather than the generally asynchronous nature of blogs or wikis, groups working together in digital worlds are able to engage each other in real time, leading to more learning, greater cooperation, and a much richer experience.

Well, I have an objection to this on two counts. First of all, there are some pretty huge differences between synchronous and asynchronous settings. And yes, I do believe that we should talk about asynchronous sites as behavior settings with different structures, behavior, and communities. Wikipedia is a bad contrast to address the issue of text vs. 3d graphics.

Secondly, the "more learning" claim sounds a lot like selling snake-oil. There are advantages and disadvantages to both synchronous and asynchronous modes. With synchronous, you have a person right there willing to help you through a task, or practice a skill. With asynchronous, you can reduce the barriers of shared time and space, and have access to an archive of previous community knowledge.

In practice, actually showing that you get strong learning gains from either has been problematic. Synchronicity or asynchronicity on their own do very little to promote learning. It's what you DO in those environments that promotes learning.

Cory: #4 Simulation matters

To which, I would say, yes. Simulation matters. But to what degree?

I'll throw down the gauntlet and make an equally (but in my opinion, better-supported claim.) The paradigms by which human beings interact with other human beings has not changed in the last 100 years. Much less in the 20 years between MUDs and virtual worlds. The fundamental lesson of 20 years of CMC research is "same shit, different mode."

9.

Kyle K: I think the big advantage of tabletop roleplaying games especially if you have a good gamemaster, is that you can ad lib not only details of your character, but details of the setting as well as you go along. As an actual example, my dwarven siege engineer is the rear-guard in a fight down a stairwell to rescue a kidnap victim. Since he has nothing to do, he starts looking at how humans build their houses.

"So, what's that wall made of?" I ask.

"Wood planking."

"What's preventing me from going through that wall with my axe?"

The DM thinks about it, "oh, about three rounds."

"Well, heck, I got nothing better to do."

It is impossible to think of every possible way that a player might try to solve a problem in advance. D&D gives the GM the option of ad libbing through unanticipated situations or strategies posed by the player.

I agree with Matt in that text MUDs offered an amazing range of expression through emote commands. It does not matter if, everyone in the room got a different image in their minds in reading: "Jora waves goodbye and leaves, her body fading away piece by peice, until her waving hand vanishes with an audible *pop.*" I suspect there would be quite a cognitive dissonance between the text description and what the person actually sees on screen.

10.

My typical stance has always been "Text is better; graphics must die." But for some reason, when I read Cory's provocation, I realized that I agreed with it.

So.. this is why:
First, his list ("sales of virtual items, weddings, stalking, gender bending, and governance") of similarities are all sociological phenomena. (Economical and political stuff is arguably sociological, for the purposes of this thrust.) It's the virtual aspect that makes them possible uniquely. The medium is utterly irrelevant except in specific applicative scenarios.

Second, the very preference I have for text over graphics makes it clear that somewhere in my subconscious, I have them fundamentally split: otherwise why would it matter?

However, I'm not sure I agree with Cory's four points. Instead, I'd put forward that text is fundamentally different from graphics simply because they're built for different personality types. Text appears to favor introspection; graphics seem to favor extroversion. That is to say that the imagination invoked by text happens from the text to the mind, whereas graphics-invoked imagination happens from the mind to the graphics.

11.

Oh wow, a text versus graphics argument! I haven't had one of those for ages!

I prefer text: the pictures are better.

Cory>Physically simulated 3D worlds are fundamentally different from text worlds.

What do you mean, "fundamentally"? They're both worlds, they both try to reproduce themselves in the imaginations of the players, but they use different methods: text talks to the mind through the language of the mind, ie. words; 3D graphics talk to the mind through the language of the senses, ie. pictures.

3D graphics and text certainly have a "fundamental" difference in interface, by definition, but you seem to be saying that the worlds are fundamentally different because of the interface. Certainly the range of what can be expressed in the worlds is different, but 3D comes off worse in that regard (text is far more expressive than graphics, which is why we're posting here using text and not pictures).

OK, so you explain why you feel that there's a fundamental difference between worlds that are 3D graphics and worlds that are text. I'll take your points one by one:

>#1 Place matters

Yes it does, but place isn't the sole province of 3D. Textual worlds have place; indeed, it would be hard to call them "worlds" if they didn't.

>#2 Reading is not doing

Correct, but neither is looking. Typing is doing, as is clicking a mouse.

>Reading a description of a place plugs into very different parts of your brain than seeing it, hearing it, and interacting with it.

I agree. Reading allows the author to connect with the imagination of the reader; images allow the author to connect with the senses of the reader, which must then be interpreted before they can reach the imagination. The imagination is where all the action is. Because the human mind is wired up to process visual patterns very easily, it can be less work to understand a scene from a picture than from a description; however, because images must nevertheless be interpreted, they can't touch the mind as well as words can.

If you want a simple world lacking in emotional content for people who don't want their imaginations exercised (ie. a world for the majority of the population) then graphics is the superior interface. If you want a simple or complex world with emotional depth or shallowness for people who do want their imagination fired, text is the superior interface.

It's just an interface, though. It doesn't make the worlds fundamentally different, any more than being real-world blind makes the real world different; it only affects your experience of the world, not the world itself.

>but claiming it is the same experience as driving a simulated car against other people and then fighting them up the cliff face is just silly.

It is if you've never been immersed in a textual world. If you have been, your statement is also pales alongside the experience of driving an actual car, which is what you feel you're doing when you're immersed.

(I have to confess, I haven't actually driven a virtual car in a text world, but I have sailed a few boats across rough, shark-infested seas and I can tell you it's far more thrilling than watching 4 pixels in the middle of the screen, which sums up all that is 3D virtual car racing).

>#3 Avatars matter

Yes, they do. They're 3D graphical worlds' way of representing characters, which matter even more.

>Avatars are perceived as real and plug into the many parts of our brains that have evolved to handle interpersonal communication.

Yes, but then again, characters are perceived as real, too.

>the range of behaviors and communication possible with avatars far exceeds text.

What?! Tell me anything you can make an avatar do in SL that it would be impossible ever to make a character do in a textual world. Go on, anything.

As an example of the reverse, my MUD2 arch-wiz arrives with a puff of thunder and disappears in a crash of smoke (as opposed to arriving with a crash of thunder and disappearing in a puff of smoke, which is the default). SL can't do that, because images are less expressive than text, not more.

I'm prepared to accept that SL avatars can produce behaviours that are recognised quicker than textual ones, but you weren't claiming that.

>As an aside, remember that text is a subset of 3D. There was an assertion that 3D couldn't impart smell to an object. Not true, since you can easily have text within a 3D world to cue the many concepts, senses, and factors that don't yet lend themselves to 3D visualization.

OK, two points here.

Firstly, if your only defence that text is more expressive than graphics is that you can put text into graphics, you're basically conceding the argument.

Secondly, if you feel that text is indeed a mere subset of graphics, how can you sustain your assertion that there's a "fundamental" difference between the two?

>Anything that can be done in a text world can be done within a 3D world. The converse is certainly not true.

In a textual world, I can stand in my own mouth, seeing my surroundings get light and dark as I open and close it. I can be part of a painting I am carrying under my arm. I can appear as a frog to one person and a beautiful princess to another. I can have internal organs. I can photograph an opinion. I can share control of my body with another player. I can drink from a Klein bottle. I can be of no gender. I can unerupt a volcano, store the world in a box, hold a soul in the palm of my hand, dance with the colour cyan.

Do that in your 3D world.

>#4 Simulation matters

Yes it does, but simulating what? You're simulating the real world, but the world of the imagination is so much greater. 3D? What happened to 4D? What's so great about tying us to physicality?

Simulation matters, because the human brain is wired up to process the real world quickly. If you present a simulation of the real world, that means the brain can accept it as fact more easily, thus aiding immersion (at least in the early stages). Simulation is therefore a good thing. Textual worlds simulate, too, but it takes more effort to get started with text. This is why textual worlds lose out against graphical ones.

However, there's a difference between being able to simulate and being only able to simulate. Text can do things 3D can't. Simulation matters, but the ability to break out of the simulation may matter more.

>Text worlds allow creators to write code and text to generate an interactive text description of skydiving.

Oh, they don't have to do that. They can stay up in the sky and have the ground rush up to meet them.

>Simulated worlds allow that

No, you can't mix media that way. Showing someone an image of the ground while scrolling text past them saying how they're hurtling towards it is giving conflicting messages. It's not the same as either showing them hurtling towards the ground without the text or describing their hurtling towards the ground without the image.

>and also allow creators to build parachutes and airplanes that enable skydiving in a 3D world. 'Nuff said.

'Nuff to expose a 3D simulation as a constraint on the human imagination, yes.

Parachutes, airplanes, chariots of fire... You can't even build a TARDIS.

There is a fundamental difference between textual and graphical virtual worlds: people prefer the latter. The interface restricts the world's functionality, but not by enough that many players care. Graphic win because of their familiarity and immediacy.

Basically, though, graphical worlds are just textal worlds with pretty pictures.

Richard

12.

Richard wrote:

Cory:>Anything that can be done in a text world can be done within a 3D world. The converse is certainly not true.

In a textual world, I can stand in my own mouth, seeing my surroundings get light and dark as I open and close it. I can be part of a painting I am carrying under my arm. I can appear as a frog to one person and a beautiful princess to another. I can have internal organs. I can photograph an opinion. I can share control of my body with another player. I can drink from a Klein bottle. I can be of no gender. I can unerupt a volcano, store the world in a box, hold a soul in the palm of my hand, dance with the colour cyan.

I think what Cory means is that because he considers text to be a subset of graphical worlds, a graphical world can just use text to do anything that graphics can't. It's a meaningless statement though, as the reverse is also true: A text world can simply use 3d graphics to do anything text can't do. I don't really know any pure graphical worlds. I know worlds that are composed of text and graphics together. I think there might be some pure text worlds, but I can't actually name any offhand (though I can name a lot that don't have 3d graphics. I wonder how Cory feels about UO? Is 2d UO fundamentally different from Everquest because it isn't 3d?)

I personally will play both text and graphical MUDs, but no graphical MUD has even come close to giving me the immersive experience I used to get when I still had the time to play text MUDs 8 hours a day. Doesn't mean they won't someday, but as Richard correctly points out, there is an infinite range of things that can be depicted with text but can never be depicted with 3d graphics or literal visual data of any kind.

--matt

13.

I agree there's not much of a clear dilineation between textual and graphical worlds. I'm one of the N3w Sch00l gamers though. :)

Richard Bartle wrote: In a textual world, I can stand in my own mouth, seeing my surroundings get light and dark as I open and close it. I can be part of a painting I am carrying under my arm. I can appear as a frog to one person and a beautiful princess to another. I can have internal organs. I can photograph an opinion. I can share control of my body with another player. I can drink from a Klein bottle. I can be of no gender. I can unerupt a volcano, store the world in a box, hold a soul in the palm of my hand, dance with the colour cyan.

Would you say this is because the mind can fill in from words more than the eyes can fill in what is lacking in what they say? I ask because I've always felt as good the graphics (and necessary sound) get, the mind can always imagine far better.

Text, to me, gets the mind started on doing that. Graphics never even give that opportunity.

14.

Matt> There is an infinite range of things that can be depicted with text but can never be depicted with 3d graphics or literal visual data of any kind.

And vice versa. The reason I commented on the skateboarding game and flocking fish was because they are both impossible to do with a text game. The skateboarding game is fundamentally a 3D game as it involves judging and timing jumps in 3D space. You could do something in a text game which described the environment in text and allowed you to type "jump", but the experience would be completely different. Similarly you could describe a shoal of fish in text and even describe them swim away as you approach, but it's not the same as having a shoal of fish reacting to you in a 3D world. In both cases 3D is superior as both examples are fundamentally 3D.

For all the more abstract examples which cannot be represented in 3D, then text is clearly the better interface and appealing to the imagination is the only option.

Taken further this argument suggests that pen and paper D&D is superior still. Not only is human language able to represent everything that a text world can, but the human generating the virtual world can describe it at any detail and respond to any desire of the players inhabiting the world. I'd love to the D&D wish spell in a computer game.

Taken to it's logical conclusion, the argument argues away all technology and ends up saying that the human imagination free of all constraint is the ultimate virtual world technology. Which may well be the case.

15.

Ricahrd> Basically, though, graphical worlds are just textal worlds with pretty pictures.

This has the benefit of being concise :-)!

Matt> You are as much 'doing' something when you issue commands in a text world by hitting keys as you are when you issue commands in a graphical world by hitting keys or mouse buttons.

Richard> Typing is doing, as is clicking a mouse.

Just to be clear, the assertion that you're making is that the experience of playing a text skateboarding game is essentially the same as playing Tony Hawk. Hopefully some of the cognitive folks can weigh in (and that I can find time to finish "Everything Bad is Good for You" so that I can pick through his sources).

On a related note, there seems to be an implicit assertion that text generates more imagination.

Richard> No, you can't mix media that way. Showing someone an image of the ground while scrolling text past them saying how they're hurtling towards it is giving conflicting messages. It's not the same as either showing them hurtling towards the ground without the text or describing their hurtling towards the ground without the image.

We play text games within a 3D world so I'm pretty sure that we integrate text more smoothly if we wanted to.

Richard> What?! Tell me anything you can make an avatar do in SL that it would be impossible ever to make a character do in a textual world. Go on, anything.

If you and I are speaking, I could sit down and make eye contact, both of which signal that I'm paying attention to you without interrupting the flow of your conversation. Sure, you could type "I sit down and I'm listening to you" but especially if twenty people did that, it would interfere with everyone else listening to you.

Another way to approach this discussion might be to think about it terms of learning. Let's say you wanted to learn about how to perform an appendectomy. You could choose 1) to learn in the real world, 2) learn from a haptic VR system, 3) learn from an interactive PC simulation, 4) learn from a video, 5) learn from interactive text, or 6) learn by reading text. Each has some advantages and in a perfect world I bet that all of us would choose a mix. After all, once I've actually done something, reviewing a description of activity helps we to review what I did. This why we visualize giving speeches, skiing moguls, etc. What I suspect, however, is that this visualization is far more effective as a rehearsal method when building on the broadest possible set of memories.

Anyway, forced to choose only one -- and assuming equal quality -- I would choose them in the order listed. Would any of you choose interactive text over any of the lower numbers?

Jim> Taken to it's logical conclusion, the argument argues away all technology and ends up saying that the human imagination free of all constraint is the ultimate virtual world technology.

Of course, but I don't think that it necessarily follows that less technology generates better imagination.

16.

Cory: On a related note, there seems to be an implicit assertion that text generates more imagination.

I don't think we have enough evidence to make that claim. I can say that I've not seen a 3d engine in which the avatar gestures provided by the software engineers offered as much of a capability to express myself as the open-ended emotes of text-based muds.

Another way to approach this discussion might be to think about it terms of learning. Let's say you wanted to learn about how to perform an appendectomy. You could choose.... Each has some advantages and in a perfect world I bet that all of us would choose a mix. After all, once I've actually done something, reviewing a description of activity helps we to review what I did. This why we visualize giving speeches, skiing moguls, etc. What I suspect, however, is that this visualization is far more effective as a rehearsal method when building on the broadest possible set of memories.

Anyway, forced to choose only one -- and assuming equal quality -- I would choose them in the order listed. Would any of you choose interactive text over any of the lower numbers?

I'm going to pull a John Stewart and say stop. Just stop. You are hurting us all by engaging in this vaporware hype-pushing.

There has been tons of research on this topic over the last 40 years and this notion that more technological complexity, more fidelity, and more synchronicity produces better learning has not been supported on the grand scale. In many cases, "high technology" does even WORSE than the "low technology" alternatives.

If you are concerned about human learning, if you are concerned about *helping* people learn, answer these three questions before you even think about media:

1: What you you want to help people learn?
2: What methods are you going to use to teach? What learning behaviors should the students practice?
3: What are the economic, political and cultural constraints?

Once you answer these three questions, the media choice will be obivous:

If you want to help people learn comparative anatomy by giving them access to a large number of exemplars, having them construct hypotheses, and constrained by animal rights concerns and cost of specimens, then VR is the way to go.

If you want to help people learn creative writing by sharing and engaging in guided constructive critique, constrained by a lack of a common meeting time, then text-based bulletin board systems are the way to go.

You seem to be suffering from the psychological problem that when you have a hammer in your hand, everything looks like a nail. By all means, 3d virtual worlds are great and getting better all the time. That doesn't mean that we should start with them as the first solution for every learning goal.

17.

I have always contended that they are all the same.

In the end, the arguments come down to simulation. But the simulation is serverside; we're really talking about the ideal representation for a given sim. Graphics does better at the skateboarding; text does better at other things. Neither is particularly good at pianos, frankly, because the music is a different sense that has nothing to do with text OR graphics.

Arguing that ease of representation of a given sim means that one delivery method is superior to the other is silly; are street signs with the text for a street name inferior to traffic signs with an icon?

Avatars, seamlessness, etc, these are all issues of representation. A textual avatar is far MORE expressive in many ways than a graphical one; and far LESS expressive in others. For example, the graphical avatar tends to be very bad at body language and emotional expression, at background (hence textual profiles)... and very good at hair, facial features, and other descriptive characteristics.

Likewise, there have been seamless textual environments and room-based graphical ones. One technique is better at one, the other technique is better at the other. And each can even embed the other within its chosen primary metaphor (isn't SL introducing node-based textual spaces via an embedded web browser?).

I would argue that it is untrue that everything that can be done within a textual world can be done within a graphical world. This is only the case because the graphical world embeds significant elements of text in order to accomplish it. Starting with the obvious--chat. But going on towards names (graphical worlds still have a ways to go before they can convey physical identity as well as a textual name, for example). Even in the real world, we don't make use of merely one medium to provide our understanding of the environment.

The urge to separate out the constituent media used to DISPLAY a virtual world and consering definitional in some manner strikes me as putting the emphasis on the client, which is the wrong place to put it. One can imagine a stunningly capable natural language parser that can provide a suitable textual interface to SL--or create an on-the-fly graphical depiction of a text world. Neither will be ideal, because the world is too complicated to display in only one way.

I also would argue that there's true artistic opportunity to be mined in choosing to represent a given world in ONLY one medium. Graphical worlds haven't actually tried to make use of their core representational strength to the degree that text was obliged to do. Strip out all the text and all sound, and then we'll see a true graphical world. Might be really interesting to see what can be done with it.

18.

I think it's also a big mistake to frame learning, creativity, and imagination as technological concerns instead of social concerns. You don't get curious, creative, and imaginative communities by dumping software on them, you get it by creating social structures to support member creativity.

With that in mind, most muds turned out to be failures even when running the exact same software as the great successful muds. Those muds that did create great structures for supporting member creativity were hugely successful.

19.

Very much enjoying this discussion, everybody. Here are a few more quick thoughts on significant differences to stir the pot:

Young children, illiterate people, and people speaking different languages can participate in graphical worlds, but not in text worlds. A graphical world like Second Life allows for intuitive and globalizing situations like the one described by James Au in "The Flat World Dancers of China" (if you follow the link you'll see a pic of Chinese text in SL :-) and csven's follow-up, "The Flat World's Shaky Virtual Ground" (which will get you thinking about low wage virtual labor as a possible force for good by teaching practical skills--a situation we'll see more of that's impossible in text worlds). Check these links out if you've got time, otherwise the elevator pitch is: graphical digital worlds are becoming a world-flattening force, integrating people from all nations in a common social and business environment that amplifies our collective imagination about what's possible. This will not happen in text. Ask yourself why that is. (I understand that real world connections sever the true virtualness that many of us here want, so if that perks your ears, I'm sorry. Just adding another perspective.)

•I'm looking forward to reading Steven Johnson's "Everything Bad is Good for You" for more of the youth, learning, and preparing for the future perspective. I remember seeing Steven on the "Roots of the Matrix" DVD (break out your Matrix boxsets! Second Life, There, and The Sims are also featured among lots more) and he made a comment to the effect that the most interesting thing happening on the planet right now is that young people all over the world are innocuously learning how to deal with, create, and manage the graphical virtual agents and environments that will increasingly come to populate the Web and behave more intelligently and autonomously. On that educational side of graphical simulation and interaction, e-learning professional Clark Aldrich recently gave a talk and read from his new book "Learning by Doing" which you can hear and see slides of right here (15 minute talk with the conclusion from the book). Slide #2 in particular might raise your eyebrow: it's a graph of "infatuation with books" with the stages of pre-literate, literate, and post-literate filled by fast food workers, academics, and enterprise leaders, respectively. Paraphrasing Clark with a wink, 'Yea, academics love that one' :-D

3D worlds can make you more skilled and imaginative because you can actually build your ideas* (complete with scripted behavior if you'd like). Think of the gap between building an intricate castle and describing how lovely it is in a few lines of text. I love books and good writing, and this isn't meant to say that an amazing description isn't valuable and immersive, but maybe the kind of deep, structural understanding that atomistic graphical worlds can express make better authors. 3D is new and people are still freaked out by it. I'm 23 and totally surprised by how hands-off friends my age can be about building in 3D, like it's not something that's really for them. Makes me feel old already! So we have to think of 3D as a valuable skill set to aquire that text can't provide us. *(I'm guessing they can also make you more "socially skilled" because they free up more of your brain to see through to the relationships--you don't need to clog up your cortex supporting a world that isn't there :-)

Again, look at this picture of Bedazzle's Abraxas Winter Castle next to a MOO text description of a comparable place (pulled from Google):

"You are in a small Northumbrian Village dominated by the huge, red edifice of Bamburgh Castle. Opposite the shops on the village green, is a Victorian church, standing on the windswept headland. The church and the castle are separated by a row of houses built of the same red stone as the castle. Between the houses and the cricket pitch is a narrow opening, the Wynding, which leads to the beach, from where there is a very good view of the Farne Islands. If you look very carefully, you can see a dark alley beside the tearooms. If you need more help, type "

This doesn't demonstrate that one is Better than the other, but different, yes. If the party-line is really that there is no difference between text and graphics, then this is the biggest academic/philosophical sausage party on Earth! (ha, I kid ;-)

•I love Will Wright's definition of simulations as "prosthetics for the imagination". Having a visible simulation mezzanine to build on frees up your cortex to understand the system and ask new questions about what's possible. And maybe there's something in that extension over immersion metaphor that's causing this rift here. We're all agreeing on the differences, we just have different virtual values. :-)

Eventually we'll be moving 3D data out of our graphical worlds and manufacturing virtually prototyped things for real (art, houses, clothes, vehicles, etc.) (the idea is making new rounds). There's no data like that we can take from text. You can say that just because you build it in 3D doesn't make you more creative or imaginative, but mass-customization via designs prototyped in 3D will certainly make us collectively more creative and imaginative. Look at what open creative markets like Threadless are already beginning to do with simple t-shirts. These aren't people aren't describing the shirts they'd like to have. They're going out and designing them, entering them into what's essentially an imagination market, and, if the idea is strong enough, perhaps having them built for real. So markets like this and the atomistic, user-created ones in Second Life most certainly excercise the imagination to the extreme.

In a closing vote/request, I would also like to hear someone chime in with differences happening in the brain when people are doing things in graphical versus reading and imagining in text worlds. That's where a lot (most?) of the honey is.

20.

> Basically, though, graphical worlds are just textal
> worlds with pretty pictures.

I dunno, Richard. On that logic, the worlds depicted in film and TV are just textal worlds with pretty pictures, and worlds depicted in radio narratives are just textal worlds with pretty audio effects, and worlds depicted in comics are just textal worlds with pretty pictures. Or devolving technology *Ubik*-style all the way down to the birth of communication, before the written word, worlds described by the village storyteller around the bonfire are just textal worlds with pretty vocal delivery. At which point, it becomes totally unclear to me what is so meaningful about the original formulation.

21.

•3D worlds can make you more skilled and imaginative because you can actually build your ideas* (complete with scripted behavior if you'd like). Think of the gap between building an intricate castle and describing how lovely it is in a few lines of text.

Well, actually most spaces in textual worlds consisted of several pages of hypertext (hypetext before html, imagine that!) with scripted interaction. So mud building is not just about a set of nodes with pretty descriptions, but anticipating every object that "lives" in that space on a regular basis, and every action that a player might take in regards to that space.

Now that is not to say that 3D modeling and construction is not a creative skill, or a valuable skill, or a skill that will be needed in the future. But I think that text muds were quite a bit more complex than you are presenting here.

In a closing vote/request, I would also like to hear someone chime in with differences happening in the brain when people are doing things in graphical versus reading and imagining in text worlds. That's where a lot (most?) of the honey is.

I think where a lot of the honey is in whether the ways in which people interact with each other radically change as a result. At one time, it was "nobody knows that you are a dog." Now we know better, and find that much of the issues of power and control that exist offline, manage to be repeated online.

Who wins? Who looses? What kinds of harassment exist? Who gets high status and how? Who is put in low status and how? What does that mean? How do participants signal their identities? How deep is the collaboration? What does the participant structure really look like? (Too many of these articles fail to deal with the Pareto's law of online participation, or fail to acknowledge how they are sampling from an elite.) Given that people are people, what warrant do we have for assuming that we need to start all over again in learning how to answer these questions for 3D virtual worlds?

Or to put it another way, are these differences revolutionary, or evolutionary? By asserting that the shift to 3d environments makes them fundamentally different, you end up reinventing the wheel all over again. Perhaps I'm merely a pre-post-literate academic, but I think that a lot of what Barker and Shoggen said about the old fashioned drug stores and banks, applies quite well to the new-fangled virtual worlds.

22.

Kirk Job-Sluder > ...mud building is not just about a set of nodes with pretty descriptions, but anticipating every object that "lives" in that space on a regular basis, and every action that a player might take in regards to that space... I think that text muds were quite a bit more complex than you are presenting here.

Yes. Thanks for pointing that out, Kirk. Totally agreed, and your points here are well taken.

23.

While there are certainly experiments that show that we can personify IRC chat, the range of behaviors and communication possible with avatars far exceeds text.

As a long time IRC addict, I'd be curious for a cite.. :)

=darwin

24.

Cory Ondrejka>Just to be clear, the assertion that you're making is that the experience of playing a text skateboarding game is essentially the same as playing Tony Hawk.

No, the assertion I'm making is that for someone immersed in a textual world the experience of skateboarding can be closer to what the player understands to be skateboarding than Tony Hawk. However, given that most of them probably got their notions of what skateboarding involves from playing Tony Hawk anyway, there probably isn't all that much of a difference...

>On a related note, there seems to be an implicit assertion that text generates more imagination.

It doesn't generate it, it uses it.

>We play text games within a 3D world so I'm pretty sure that we integrate text more smoothly if we wanted to.

If you wanted, you could implement MUD1 in SL merely by displaying the text on a static background. But then as Matt Says, hey, guess what? I could implement SL in MUD1 merely by showing the pictures over static text.

SL has text because you can't do everything with graphics. However, there are some things you can't do with both. In a textual world, I can put my character inside a box that my character is holding. You can't implement that in a 3D graphical world, because the world model isn't up to it. Thus, any physical consequences of it (eg. what happens if I shake the box) can't be read off the physics. Would an observer standing next to me as I shook the box also shake? What if I were inside the box but they weren't? The nodal world model you get from a textual world can handle this, but the 3D graphical world model can't. You'd have to layer the nodal model atop the graphical one to access its physics - but then you couldn't use the graphical model to render the image.

>If you and I are speaking, I could sit down and make eye contact, both of which signal that I'm paying attention to you without interrupting the flow of your conversation.

Can I sit down and not make eye contact, or does that happen automatically?

>Sure, you could type "I sit down and I'm listening to you" but especially if twenty people did that, it would interfere with everyone else listening to you.

What, and it's beyond the wit of a programmer to convert 20 "[name] sits down and listens" into "The audience sits down and listens"?

>Anyway, forced to choose only one -- and assuming equal quality -- I would choose them in the order listed. Would any of you choose interactive text over any of the lower numbers?

So you're saying that 3D graphics are better than text because training a surgeon using only 3D graphics is only slightly less a bad idea than training them using text? That's hardly a ringing endorsement...

Besides, it depends on the surgeon and the quality of the simulation. Current 3D systems are lousy at modelling interiors, and if the surgeon had actually seen the inside of a patient before then an interactive text might be more suited to their training than 3D graphics. As you say, a mixture is going to be best.

>Of course, but I don't think that it necessarily follows that less technology generates better imagination.

Nor does it follow that more technology generates better imagination. The difference is that I'm not claiming the former, but the pro-3Ders seem to be claiming the latter.

Richard

25.

Jerry P>Young children, illiterate people, and people speaking different languages can participate in graphical worlds, but not in text worlds.

Well tra-la-la. Blind people can play textual worlds but not graphical worlds. This is just an interface issue.

>A graphical world like Second Life allows for intuitive and globalizing situations like the one described by James Au in "The Flat World Dancers of China"

I should have kept transcripts from the Norwegian incarnation of MUD1, where some people would speak to each other in English and others would use (gosh!) Norwegian.

>graphical digital worlds are becoming a world-flattening force, integrating people from all nations in a common social and business environment that amplifies our collective imagination about what's possible.

That's true, they are (or, given that it's only SL doing this, I suppose we should say "it is"). You're confusing causalities, though. It's nothing specific to 3D graphics that's doing this - after all, WoW has 50 times the number of players that SL has but it's not doing it); rather, it's SL's deliberate lack of a magic circle and its developers' conscious attempt to connect the real and the virtual in a newbie-friendly manner.

>This will not happen in text. Ask yourself why that is.

Text isn't newbie-friendly.

>•3D worlds can make you more skilled and imaginative because you can actually build your ideas*

More skilled, yes; more imaginative, no. They teach craft, not imagination.

If you want to create 3D objects then a 3D world is clearly a better place to learn the necessary skills than is a textual world, because the interface better matches your medium. Then again, if you want to write stories you'd be better off in a StoryMOO, because the interface is made of the very stuff of your trade.

I can't build my ideas in 3D because 3D isn't expressive enough for me. Now any medium constrains that which can be created in it, and working within those constraints can lead to creative solutions and art. It can fire imagination, but it can also limit it. For some people, 3D is their medium, and SL will be perfect for them. For other people, manipulating objects in space is not what they want to do, and SL would be useless to them. That's because the interface is more appropriate for them (note that's interface, not world).

>Think of the gap between building an intricate castle and describing how lovely it is in a few lines of text.

No no no, when you build a castle in a textual world you really do build it, you don't merely describe it. You have to put together components in a structured fashion in order to realise (virtualise?) the object. The castle is reflected in data structures, not just text. If you think building a textual castle isn't building, then neither is gluing triangles together to make a graphical representation of a castle.

>maybe the kind of deep, structural understanding that atomistic graphical worlds can express make better authors.

And maybe they make worse authors, or prevent people from becoming authors altogether. It's just a medium; you're making claims for it way beyond its reach.

>*(I'm guessing they can also make you more "socially skilled" because they free up more of your brain to see through to the relationships--you don't need to clog up your cortex supporting a world that isn't there :-)

I'm staggered by this remark. Would I be right in suggesting that you have never read any papers on Cognitive Maps?

>This doesn't demonstrate that one is Better than the other, but different, yes.

It demonstrates that the interface is different, but not that the world is different. The assertion Cory was making is that physically simulated 3D worlds are fundamentally different to textual worlds. All you're saying, however, is that they're different because one uses pictures and one uses words, which is kinda obvious.

>Having a visible simulation mezzanine to build on frees up your cortex to understand the system and ask new questions about what's possible.

No, the imagination is the crucible here, where the new questions must go to be asked. You can only test the system predictively when you've internalised it, which means you have to get it into your imagination - you can't rely on having it as a "real", testable physicality that will satisfy all your speculative needs.

>Eventually we'll be moving 3D data out of our graphical worlds and manufacturing virtually prototyped things for real.

Again, why does that make the world fundamentally different? Different data can be taken from textual worlds and used in the real world, but that doesn't make the worlds different, just how they're depicted on the screen.

>You can say that just because you build it in 3D doesn't make you more creative or imaginative, but mass-customization via designs prototyped in 3D will certainly make us collectively more creative and imaginative.

I disagree. It will give us more opportunity for our imaginations to blossom, and it may allow collaborations that wouldn't happen in real space, but it won't make us more creative or imaginative than we were anyway. It enables the expression of imagination in ways which might previously have not been open to so many people, but I'm not going to let you claim that the glories of SL will make the unimaginative imaginative, or the imaginative even more imaginative. You're overstating your case.

>In a closing vote/request, I would also like to hear someone chime in with differences happening in the brain when people are doing things in graphical versus reading and imagining in text worlds.

Be careful what you wish for: any evidence that 3D virtual images affect the brain chemically will be picked up by the ban-the-games crowd and used against you..!

Richard

26.

Darwin> As a long time IRC addict, I'd be curious for a cite

https://www.citeulike.org/user/klouie/article/216057

As part of the experiment, the subjects were sometimes trading with other people via text and sometimes with a computer via text. The impact of oxytocin on increased trust only appeared when people were told that they were trading with other people. They used it to test if oxytocin was driving overall risk tolerance or trust.

27.

Think how contentious this thread would be if half of Terra Nova wasn't getting over DiGRA hangovers :-)?

Richard> Can I sit down and not make eye contact, or does that happen automatically?

No, you can choose where to look.

Richard> What, and it's beyond the wit of a programmer to convert 20 "[name] sits down and listens" into "The audience sits down and listens"?

Actually, I was thinking about the case of people entering an area to listen to a speaker over time. And, yes, I'm sure that you say "well, the programmer would just . . ." but we're now drifting into the combinatorial explosion of stage magic versus simulation, which will be future (equaly contentious) post.

Kirk> I'm going to pull a John Stewart and say stop. Just stop.

Wow, I've always wanted to be "John Stewart"-ed :-) So, you'd rather learn surgery straight from a book? I only ask, because, having a) actually flown planes, b) having flown full motion simulators, c) having played flight simulators, d) having seen videos about flying planes, and e) having read books about flying, I'm still pretty comfortable with the order I listed them.

Raph> [several good points about the hybrid nature of current graphical worlds]

Yes, although nowhere was I arguing that we would want (or should) abandon text, audio, etc.

Raph> are street signs with the text for a street name inferior to traffic signs with an icon?

Well, usually the opposite is the case, right? On the other hand, do you want your map to be a test description of the streets "As you walk down Oak Grove you will cross 2nd Street" or would you prefer a map? I realize that we probably wouldn't get a universal preference here, but the fact that different people would choose different options is likely a peek into some of the different cognitive functions at work.

Also, to be clear, I'm generally *not* making normative claims about whether text or graphics are better. The assertion is that they are fundamentally different in how they are perceived and experienced and those differences lead to differences in subsystems likes communications, collaboration, or creation. I am not arguing -- nor do I believe -- that text games suck.

I also find it interesting to now read the "text is less user-friendly than 3D" posts. During SL development, when amateur-to-amateur creation was still controversial (you know, way back in 2003) the big arguments against something like Second Life were rooted in designer arrogance. You needed professionals to make this stuff. In response, people would turn to MUDs, MOOs, and the web to point out how prolific and pervasive user creation was. The nay sayers would then nod and say, "sure, but that's text, everyone knows how to type" but for 3D creation you still need professionals. Now that we see that people really enjoy creation in 3D spaces -- and, given the availability of tools and a community to support and teach them how to use them -- that they can produce some exquisitely crafted content, the argument now flips that 3D is actually easier.

I agree with Raph and Richard -- people are creative independent of medium or media. But I do feel that it is important to look at both the positive and negative differences between them, as well as the different opportunities they offer. And, yes, to think about whether they are transformative or not.

28.

Could the fundamental difference be the amount and type of information communicated over the net? Could the fundamental difference be the way we process the information received?

Cory’s four factors all points to physicality as the fundamental difference. This is an argument that show is different from tell, different from write.

Philosophically, I recall a couple of old sayings and phrases:

“A picture is worth a thousand words”
“A single word can kill”
“Jumping off the pages”

My 2 cents,

Frank

29.

Matt > For instance, I may find that a text description of the result of my actions results in a much richer image in my head, whereas you may find that a graphical description of the result of your actions results in the richer image. Different strokes, that's all.

This is an empirical proposition, and someday we'll have cognitive psychologists writing papers about it. (I imagine there's a somewhat related literature out there, but I would question whether it really pertains. It's probably about cyberspace in general, and has not develop metrics for this immersio' thing that's so important in fantasy world-building.)

Here's a theory rests on two key variables. The first has been emphasized here, and we might call it 'imaginitive scope.' Tolkien wrote that Secondary Worlds should only appear in text, because drama (and film) fix the image and thus prevent the mind from developing the subjectively most accurate image. Thus you might say that text has more imaginative scope than graphics.

My colleague Annie Lang has done some work (can't give you a URL yet - it's in press) that points to a second variable, which we might call 'cognitive processing load.' She shows that the brain has to make some meta-level decisions about what to pay attention to, and this involves a resource-constrained choice problem. In milliseconds, the brain has to choose how to respond to a stimulus, sending orders to muscle groups ('look at that flashing, curvy thing'), calling up matching patterns from storage, invoking reasoning and judgment cycles and so on. Something that really stimulates the system gets a big fat load of resources, and so you can conjure up in your head a Mordor that's more terrifying than anything depicted on screen. But it takes time and effort to conjure that image. A 3D image gets you to the guts of the image much more quickly, with less brainwork. You can get your mind into the state of the world almost without thinking about it, and that makes it feel real.

When we shift from text to 3D, we lower imaginitive scope, but we also lower the processing load. The first decreases immersion, the second increases it. There's a sweet spot in there, but I suspect, as Matt says, that it's different for different people.

I think that moving from 3D to a multiplayer 3D may raise or lower processing load. When you're by yourself, you have to do a bunch of suspension-of-disbelief activity, which is a pain. Put yourself in a group of role-players, though, and your 'society-knows-truth' functions will help you with your disbelief suspension. OTOH, put yourself in a group of non-RPers, and the the 'society-knows-truth' function works the other way. Or, it encourages you to immerse yourself in the game mechanics, not so much the fantasy atmosphere.


30.

Cory:

I'm still wondering if you feel there's also a fundamental difference between 2d UO and 3d Everquest, as your assertion is that it's 3d, not just graphics, that makes a fundamental difference. I suspect you don't want to go there, because you know that what you call text MUDs rarely lack graphics. They just lack 3d graphics.

--matt

31.

Cory: Wow, I've always wanted to be "John Stewart"-ed :-) So, you'd rather learn surgery straight from a book? I only ask, because, having a) actually flown planes, b) having flown full motion simulators, c) having played flight simulators, d) having seen videos about flying planes, and e) having read books about flying, I'm still pretty comfortable with the order I listed them.

This is completely avoiding the points I made in my post. I learned to ride a motorcycle by taking a class that put a bunch of novices on top of 50cc training bikes on a big vacant parking lot, so I agree with you for that particular domain of cognitive-motor tasks. But you are making a huge leap of faith in saying that what is good for learning to fly a plane, is good for learning a programming language, learning to read Spanish, understanding WWII as a temporal sequence of events, or identifying the motifs in Japanese film.

One of the keys here is dealing with abstraction. You can create a piano, but how do you express the structure of a gavotte, or a concerto using 3d iconography with a low degree of ambiguity? How do you communicate the role of a narrative "hook" in drawing people into a story? How do you support the creation and peristance of extended narratives in a communications medium that lacks long-term history and pushes the user towards short utterances? How do you deal with learning tasks such as abstraction and generalization?

Humans are visual animals, but they are also storytelling animals. In many cases, we learn something by struggling how to put concepts into words.

Cory: Well, usually the opposite is the case, right? On the other hand, do you want your map to be a test description of the streets "As you walk down Oak Grove you will cross 2nd Street" or would you prefer a map? I realize that we probably wouldn't get a universal preference here, but the fact that different people would choose different options is likely a peek into some of the different cognitive functions at work.

Certainly, there are different cognitive functions involved in following a narrative map, vs. following an 2D iconic map. The fact that there are differences would suggest that we should tailor the medium to fit the cognitive function we want for learners to practice, does it not?

Cory: Also, to be clear, I'm generally *not* making normative claims about whether text or graphics are better. The assertion is that they are fundamentally different in how they are perceived and experienced and those differences lead to differences in subsystems likes communications, collaboration, or creation. I am not arguing -- nor do I believe -- that text games suck.

*WOW*. I think this is really the core of this debate. Not whether 3D communication is different from 2D communication, but what is the system, and what is the subsystem. To me, communication, collaboration, and creation are the core of what should be happening in building online worlds, and the choices of what kinds of text or VR presence should be built around that.

I'm also thinking that your medium->perception->communications model has very little validity, and that the kinds of social structures you build in a space will have an extremely strong influence on how signs (whether video, audio or text) are preceived.

Third, I'm actually investigating synchronous discussion in a 3D virtual world, and while there are differences, just dropping kids into a 3D space does not magically change their discourse into something radically new, creative and innovative.

Edward: When we shift from text to 3D, we lower imaginitive scope, but we also lower the processing load. The first decreases immersion, the second increases it. There's a sweet spot in there, but I suspect, as Matt says, that it's different for different people.

Well, I think there are other loads to think about as well. I'm certainly in agreement that 3D lowers the processing load from being able to comprehend one's environment. However, I suspect the management load of just dealing with synchronous messaging is going to be a bigger factor in communication and society.

Most 3D evironments I've seen require mode-switching when you want to "say" something to another user. We know that mode-switching is bad. Heck, recent studies suggest that we've replaced the martini lunch with Microsoft Outlook. I suspect (based on experience) that once you've switched into a synchronous text-messaging mode that they are all very similar.

32.


Text is just one form of pretty graphics.

I tend to find a broader array of graphical representation to be more immersive. And I also generally find simulation to be more immersive.

Most textual worlds (not all) are defined at a very high granularity. Nodes are limited to those explicitly defined, with a predetermined number of connections. In a simulated world you have connections / actions which are implicit as well as the ones explicity defined. For example, jumping from a cliff or tree into a courtyard instead of walking through the gate.

I find it far more constraining to be limited to options which an author thought to fill in. Simulations generally ofter a much wider range of choice than a narrative world.

One of the biggest appeals of PnP gaming is being able to step outside the narrative options. The GM is your simulator, for extending the world beyond what was explicitly defined.


In a textual world, I can stand in my own mouth, seeing my surroundings get light and dark as I open and close it. I can be part of a painting I am carrying under my arm. I can appear as a frog to one person and a beautiful princess to another. I can have internal organs. I can photograph an opinion. I can share control of my body with another player. I can drink from a Klein bottle. I can be of no gender. I can unerupt a volcano, store the world in a box, hold a soul in the palm of my hand, dance with the colour cyan.

The Klien bottle may be a difficult one, but otherwise sure they can be done. It seems to be that Richard's gripe is more about the ability to creatively and easily author than it is truely about the ability to experience, and immersiveness.

33.

Flatfingers> I'm not sure many people are going to disagree in any substantial way.

/wince

Should have known better, shouldn't I?

Matt> positional audio has nothing to do with 3d graphics.

Look at these words in the original proposition: "Physically simulated 3D worlds".

People have assumed this to mean "graphics," but where is that a requirement?

Audio is an additional channel for physically transmitting information energy. And as noted, so is smell. So are tactile feedback controllers. I'd suggest the word "multisensory" to describe entertainment experiences that use multiple sensory channels to communicate with the consumer, rather than using the single channel of text.

So to recast the question: would Game X presented wholly as text be experienced by most people in a fundamentally different way from Game X presented over multisensory channels?

In other words, does the medium define the message?

1. As Raph pointed out, despite some exceptions, most text games have "seams," while most multisensory games are seamless.

I interpret this as meaning that the granularity of place is usually finer in multisensory games than in text games. Text games usually implement place as "rooms" -- moving one's sense of place occurs like a scene change or a jump cut in a movie. Multisensory games, with much finer granularity of place, give the appearance of smooth motion over many tiny steps taken very rapidly, as a master shot in a film gives the perception of continuous motion (although each frame of the film is a discrete image).

It's reasonable to think that perceiving motion as continuous is a different (I would say "more realistic") kind of experience than perceiving motion as stepwise.

2. Text is imaginatively active; multisensory is imaginatively passive.

This is what I meant when I said that text is evocative. It evokes a whole image by describing the key aspects of that image (the "telling details"), and then expects the reader to participate in the creation of the image by actively filling in everything else that matters.

Multisensory experiences do more of the imaginative work for the consumer. You don't have to imagine the moonlight gleaming on the castle's towers, or the pennons snapping in the wind, or the palpable thunder of approaching hoofbeats -- those aspects of reality are created for you.

This distinction between active and passive imagination requirements makes different communication forms more or less appealing to different people. Some people prefer movies to books because the former, through engaging more of the senses, require less imaginative effort. Some people prefer books to movies because the required exercise of imagination allows a more personalized experience.

If a text game is more like a book than a movie, and a multisensory game is more like a movie than a book, then doesn't the preference of some people for books and other people for movies suggest that the form of a computer game also matters?

Or is there no "fundamental difference" between a book and a movie?

...

Finally, like Cory I'm not really coming at any of this from a "better/worse" perspective. Better/worse depends on effectiveness at producing a desired effect; so the first question ought to be whether there's any difference of effectiveness at all.

--Flatfingers

34.

Just to throw another petrol bomb:

Richard> In a textual world, I can put my character inside a box that my character is holding. You can't implement that in a 3D graphical world, because the world model isn't up to it.

The MASSIVE-3 and SPLINE 3D virtual environment systems can both implement this and the TARDIS example you mentioned earlier by linking together multiple locales. 3D doesn't imply a single euclidean coordinate system.

I still think that text worlds are better for some things, 3D worlds are better for others and that creativity is independent of medium though.

35.

Kirk> . . . good for learning a programming language

You have read about MUPPETS, right? They talk at length about the pedagogical benefits of teaching Java within 3D.

Krik> I'm actually investigating synchronous discussion in a 3D virtual world

Great, I look forward to seeing your results!

Matt> I'm still wondering if you feel there's also a fundamental difference between 2d UO and 3d Everquest

I think that this will have to wait for the "simulation versus stage magic" post . . .

36.

I think the distinction that is relevant here is a (gasp) philosophical one. Since Smartbomb (my book coming out in November 2005, shameless plug ;) deals with this very issue I'll give my 5 cents, which are drawn from the Introduction:

The difference between text (and even movies) and virtual worlds (by virtue of being videogames) is like the difference between getting directions and using a map.

Put more bluntly, text (and arguably even streams of images) are descriptive, that is they function largely by using propositional content to describe states of affairs: eg, "It is raining," "You are bleeding," etc.

Videogames, on the other hand, are fundamentally models--dynamic computer models to be precise. Models do not rely on descriptions to communicate. They are not propositional in nature. Rather, models function by directly instantiating the very properties they seek to represent.

Consider the following:

If someone wanted to learn about the orbits of planets in the solar system, they could consult books, watch educational movies, or even attend lectures. Each of these sources would teach about planetary motion by describing them, using words and/or a stream of sound and image.

Another way to learn about the solar system, however, would be to use an orrery, a collection of spheres mounted on a system of armatures and gears. The orrery represents the solar system not by describing it, but by serving as a model of it.

Each videogame from Pong to single-player Tetris to the most sophisticated virtual world, where thousands of players interact in a fully 3D world at the same time, is ultimately a model. Videogames allow players not simply to tour a space but to influence what happens in that space. And while one can learn about the solar system both from books or films, learning from a model like an orrery or a videogame, or even simply playing with one, requires and encourages very different cognitive skills and imparts a different kind of knowledge (which philosophers have cunningly dubbed, 'model knowledge').

In a society in which entertainment has nearly become a secondary education system – studies consistently show that people today spend more time consuming entertainment than they do with either their families or in school - this shift to model-based entertainment will have an undeniable impact on the way people think and communicate. The invasion of the videogame, then, gives new life to the anthropologic saw: “Show me the games of your children, and I’ll show you the next hundred years.”

So what does this do for our discussion?

To the extent that a virtual world employs text, it is abandoning (I mean that neutrally) it's function as a model and adverting to the power of description. When text is used, it is used as a descriptor. Further, text used in such a fashion means that some part of the model in which it is contained is either incapable of, or not desired to, serve the designer adequately.

Interestingly text-as-object is capable of serving both functions: as text, its content is descriptive, as an object (that is, if the text is treated as an object in the model) it can become part of the model itself.

Of final note, there is nothing essential about models that requires 3D. A map is a model, even if it is only rendered in two dimensions.

What I hope to accomplish here is to provide a bit more solid framework and common ground for discussing the ways in which text and objects are used in virtual worlds.

Cheers


37.

I think there are a large number of fundamental issues that separate the applicability of each realm. I don't accept that either realm (text worlds versus 3D worlds) is a subset of the other... each has areas where the other will not work as effectively.

That doesn't, however, mean that I think they are equally effective at creating "worlds", since the term has different boundaries in each case.

One of the fundamental issues is that of information bandwidth. Basic information theory as elucidated by Shannon tells us that bandwidth limitations will constrain what we can communicate per unit time. In general, visual and audio channels can carry a vastly greater bandwidth of largely non-subjective information than can be achieved with serially reading an attempted textual description of the same circumstance.

Textual descriptions can carry a "shorthand" description, and that's fine when you're talking about a situation where you just don't care whether multiple participants substitute their own personal interpretations.

There are, however, a large number of situations where the artist is attempting to convey something where the temporal relationship between the components is an integral part of the experience that they are attempting to convey. Converting that experience to text and then having users read it at their own personally-constrained pace can greatly decouple (if not outright destroy) that temporal relationship.

The impact is therefore not the same. Visual and audio media can provide a concrete temporal relationship that is often much less subjective and non-consensual... the participant often doesn't have the OPTION of choosing to read (or re-read) the description at their leisure.

For me, this breaks down into a split regarding the degree of "shared experience" that the artist is attempting to convey. Kirk wrote:

I agree with Matt in that text MUDs offered an amazing range of expression through emote commands. It does not matter if, everyone in the room got a different image in their minds in reading: "Jora waves goodbye and leaves, her body fading away piece by peice, until her waving hand vanishes with an audible *pop.*" I suspect there would be quite a cognitive dissonance between the text description and what the person actually sees on screen.

It might not matter to you that everyone in the room got a different image from that description, but it might matter immensely to an artist that wants to make a PARTICULAR image associated with an event have a specific meaning. If the set of messages that are being conveyed don't carry the same impact without specificity, then allowing users to substitute their own imagined images leads to experiences that deviate substantially from the expression that the artist wanted to achieve.

The more shorthand that is required in a textual description, the greater the difference in experiences that are being conveyed to the readers. In a real-time application, there's simply not the available bandwidth in text to convey a SPECIFIC and detailed vision to a large number of parties and know that they are all operating from the same basic experience.

That's why there are experiences that can be conveyed in film, where any textual description of that experience will simply fail to convey anything even close to the same message. If you don't care whether the message is the same, that's fine. Some people do care, immensely.

I have a player experience that's being built into my current project that relies very, very heavily on the 3-D nature of visual perception in humans. The meaning of the experience leverages very strongly off of the "aha" gestalt experience, and I don't see any method where the same experience could be conveyed textually. We don't process text information in the same way.

Players substituting their own imagined versions of the images I'm creating would utterly fail to get the point... the point would have to be given to them rather than discovered organically. That would wreck a major aspect of experiential nature of the story arc, IMO.

Richard wrote:

Reading allows the author to connect with the imagination of the reader; images allow the author to connect with the senses of the reader, which must then be interpreted before they can reach the imagination. The imagination is where all the action is. Because the human mind is wired up to process visual patterns very easily, it can be less work to understand a scene from a picture than from a description; however, because images must nevertheless be interpreted, they can't touch the mind as well as words can.

Err... wait a tick. Are you alleging here that the words somehow don't have to be interpreted, and therefore touch the mind directly somehow? If so, I think you're grossly off-base. In general, I would suspect that there is vastly more ambiguity and interpretation in conveying concepts with words. That's why we tend to suck at it. When you abstract away the visual and audio clues that come from interpersonal information conveyance via speech, you increase the ambiguity further still... as a natural function of stripping away communications bandwidth.

As to the "imagination versus senses" dichotomy, I would state it along the lines that text is far more intrinsically intellectual in nature, whereas visual/audio conveyance is more intrinsically experiential in nature.

Each have different functions and applicable domains. I would suspect that the choice would come down to whether you're trying to convey a largely open-ended free-form set of concepts, where every individual's take-away is likely to be quite different, or if you're looking to convey something very specific as a shared experience.

When it comes to which is more world-like, the choice seems natural to me (but is largely so because of my philosophical disposition). For the most part, my interactions with others in the real world are a shared experience with (I suspect) rather little deviation from the fundamental reality. Since the world (reality) is my fundamental yardstick for measuring the concept of "world-like", it seems like the shared experience is vastly more world-like than everyone making up their own versions with billions of (sometimes major) inconsistencies between them.

(And yes, I'm deliberately ignoring the individual and minor differences in personal information processing that each of our bodies performs, and the selective interpretations and meanings we each apply to those experiences... because the same thing happens when two people read the same text passage, only more so.)

Richard>


It's just an interface, though. It doesn't make the worlds fundamentally different, any more than being real-world blind makes the real world different; it only affects your experience of the world, not the world itself.

As politically incorrect as it might be to say, I contend that there are experiences that the blind simply cannot share effectively with the sighted, because they cannot share the same inputs... and any textual or verbal description of that input will be far enough removed to have it be a different experience entirely.

When you limit the interface to a given reality, you constrain what you can accurately share with others (per unit time). A disembodied brain with only enough nerve attachments to receive a sequential Braille feed of a description of what's going on around it will probably not be able to share a common world experience with you in a meaningfully aligned way.

Richard>

It is if you've never been immersed in a textual world. If you have been, your statement is also pales alongside the experience of driving an actual car, which is what you feel you're doing when you're immersed.

Again, I think this is the fundamental divide between imaginary/intellectual versus experiential. I contend that we are, for the most part, wired for experiential interaction with the world. The closer the interface comes to our signal processing mechanisms, the more accurately we can map to the set of experiences we might experience if the virtual were actually real. That's why I see virtual world software moving towards integration with our sensory appartus, rather than towards our intellect.

I suspect that the most world-like virtual interface that is every achieved by mankind will most definitely NOT be anything like a text-world interface. I do believe it will contain a strong 3-D visual component, because that's how we're wired to perceive the world... we're not wired to perceive the world as a serial string of written words, IMO.

Richard>

What?! Tell me anything you can make an avatar do in SL that it would be impossible ever to make a character do in a textual world. Go on, anything.

I'll give it a shot, but I'll address what I see as the fundamentally more important question... what I as a player experience, rather than what my avatar "does".

The first time I played Half-Life and reached the point where my character stepped out of the cramped tunnel where he had been crawling and onto a sheer cliff face, in a span of ten seconds, my avatar took a step too far and plunged to a shattering death.

I didn't actually see the first moment of body chunks flying, because the visual illusion of acceleration caused by the 3D environment representation of such a fall caused a vertigo attack, and I had my head in the luckily-nearby trash can as I vomited.

You're welcome to try to construct a textual description of such an event that tricks my body well enough to cause such a reaction... I contend you can't. That first fall was experiental, not intellectual. That interface was crude... I suspect we are moving more and more towards interfaces that come closer and closer to providing our bodies with different experiential realities. I predict they will not be text interfaces.

Novels provide me with different imaginary and intellectual realms, but I rarely garner something from them that is as close to experiential. It's an entirely different realm.

Richard>

Yes it does, but simulating what? You're simulating the real world, but the world of the imagination is so much greater. 3D? What happened to 4D? What's so great about tying us to physicality?

I would think the fact that we perceive reality through our physicality is a pretty big deal. You can intellectually discuss, imagine and contemplate other realities where physicality doesn't apply, but I contend that you largely won't be able to make them experiential, because we don't have information-inputs in our bodies that map you whatever you imagine.

The more severely you disconnect your "world" from the reality of how we gather information, the less world-like you make it IMO. It becomes an abstraction that will be less and less a shared experience, since people will not map to it in the same ways.

Richard>


Simulation matters, because the human brain is wired up to process the real world quickly. If you present a simulation of the real world, that means the brain can accept it as fact more easily, thus aiding immersion (at least in the early stages). Simulation is therefore a good thing. Textual worlds simulate, too, but it takes more effort to get started with text. This is why textual worlds lose out against graphical ones.

It's not just a function of effort. Because of bandwidth limitations in processing text, it's simply not possible to gain the same degree of overlap of experience with text as you can with visual analogs, and even more so with the combination of audio and visual input.

That allows a vastly greater level of information per unit time to be conveyed, and will eventually approach the levels of what we perceive today in everyday life. The closer we can get to actually experiencing an event along with others, the closer we will come to having something that can meaningfully carry the term "world" without diluting it to the point of being nearly meaningless.

Richard>


However, there's a difference between being able to simulate and being only able to simulate. Text can do things 3D can't. Simulation matters, but the ability to break out of the simulation may matter more.

There are applications where text is certainly useful. I would, however, challenge your contention that it is text that carries that power. I contend that it is language itself that carries the power you are attributing to text.

Is there anything that you can do with text that can't be done with a combination of visual and audio input (the spoken word, for example)?

I would contend that there are many spoken experiences that simply can't be conveyed properly in written text. Do you have examples where the converse is true (where the written form can convey something properly, but not the verbal)?

Direct integration of audio processing is a natural part of the evolution towards interfaces that map better to our bodies... that's why we see things like TeamSpeak support being integrated into 3D worlds, as well as positional audio.

Are there major efforts to integrate real-time human speech into "text worlds" that I don't know about? Who is touting the more world-like aspects of text+speech in text worlds?

38.

Flatfingers: Finally, like Cory I'm not really coming at any of this from a "better/worse" perspective. Better/worse depends on effectiveness at producing a desired effect; so the first question ought to be whether there's any difference of effectiveness at all.

I don't know. Cory in the OP, and in his linked essay did make some unqualified "better than" claims: "So, pulling this all together, why is creativity in collaborative 3D worlds better than text?" "Rather than the generally asynchronous nature of blogs or wikis, groups working together in digital worlds are able to engage each other in real time, leading to more learning, greater cooperation, and a much richer experience."

So there are a number of claims about 3D worlds:
1: Better creativity.
2: More learning.
3: Greater cooperation.
4: Richer experience.

#1 has not even been properly defined, much less defended. #2 is, at least to date, empirically false in looking over multiple studies. #3 is also unsupported in that synchronous conversation supports some aspects of collaboration, and does not support others. And to throw a bone to the "fundamental difference" camp, I do think #4 is true, but the road from #4 to #1-3 is long and hard, and strongly dependent on social engineering.

While we are talking a lot about the medium through which participants interact with and manipulate the environment. What most of you are NOT talking about is the medium through which participants interact with and manipulate each other. To the extent that the communications medium of 3D worlds shares many of the same limitations and abilities as IRC, AIM and text muds, communication, participation and collaboration are not likely to be "fundamentally different."

When communication in 3D worlds no longer involves typed short utterances transmitted one line at a time, then we can have a conversation about them being "fundamentally different."

Cory: You have read about MUPPETS, right? They talk at length about the pedagogical benefits of teaching Java within 3D.

Certainly. However reading the MUPPETS paper (at least the one I have access to) it looks like they are using a multimedia approach by using MUPPETS as an optional enrichment activity in a traditional classroom-based curriculum.

39.

Kirk Job-Sluder wrote:

When communication in 3D worlds no longer involves typed short utterances transmitted one line at a time, then we can have a conversation about them being "fundamentally different."

Are you unaware of TeamSpeak and the affect it has had on the landscape of 3D worlds and the level of "communication, participation and collaboration" this allows?

The collector's edition of Guild Wars comes with a headset and three months worth of TeamSpeak.

Please proceed now with the discussion of the fundamental differences.

40.

Barry Kearns: It might not matter to you that everyone in the room got a different image from that description, but it might matter immensely to an artist that wants to make a PARTICULAR image associated with an event have a specific meaning.

Certainly. However, 3D virtual environments (in general) might not offer the artist the ability to produce that kind of effect at all. At least not without a high degree of fairly sophisticated skill and work, and cooperation with the people who define the physics of that world. If the physics does not allow animated transparency, the artist is out of luck, and there is no way around that.

But by all means, if you are a visual artist, then text worlds are probably not for you. If you are a graphic artist, then visual worlds are not for you. I'm comfortable with that. I also don't think it is very relevant to what I see as the problems with the "fundamental difference" hypothesis in regards to the kinds of social phenomena we might see in those worlds.

I would contend that there are many spoken experiences that simply can't be conveyed properly in written text. Do you have examples where the converse is true (where the written form can convey something properly, but not the verbal)?

Calligraphy and typography. The poems of Archie and Mehitabel. Visual puns using homonyms. Tables, indexes. I would argue that emoticons don't translate well into spoken language, and have evolved to an extent that they are different from the physical gestures they represent.

In addition, I think that text offers some key affordances for the reader beyond just the artist's ability to express themselves.

41.

Barry Kearns: Are you unaware of TeamSpeak and the affect it has had on the landscape of 3D worlds and the level of "communication, participation and collaboration" this allows?

Please proceed now with the discussion of the fundamental differences.

I'm aware that speech telephony has been available as an ad-on for many systems. However, I'm also aware that most systems still use the same old IRC-like text-mode CMC.

And I have no problem with the claim that internet telephony is fundamentally different from synchronous CMC. My point has always been that CMC matters more than text vs. graphics for things like collaboration and creativity.

And still, I'm going to be contrarian and point out that there is a long history behind collaborative telephony, and talking about a shared object in a 3d space is not that fundamentally different from talking about a shared object delivered through email.

42.

^^^^^ ?? :(

Whoops that last post was by me. Cut and pasted into the wrong field.

But it did provide me with an excuse to post an example of a text-based communication that would loose something in the translation.

43.

Barry wrote: Because of bandwidth limitations in processing text, it's simply not possible to gain the same degree of overlap of experience with text as you can with visual analogs, and even more so with the combination of audio and visual input.

That allows a vastly greater level of information per unit time to be conveyed, and will eventually approach the levels of what we perceive today in everyday life. The closer we can get to actually experiencing an event along with others, the closer we will come to having something that can meaningfully carry the term "world" without diluting it to the point of being nearly meaningless.

If I get you right, you're basically saying that the brain can process more information, more easily, and (more iportant) in unique ways if the information is presented to it by means of a world-model than if it is presented to textually. I totally agree. They are fundamentally different: one relies primarily on description-based communication, the other relies primarily not on the graphics per se but on that fact that it is more wholly model-based communication. That is where the fundamental distinction lies, imo.

But it is important to note that with respect to brains, the move to 3D graphics, positional audio, real-time voice chat, etc. is successful not because it's closer to how we perceive 'reality' but because it's closer to how are brains build reality on the inside-namely, through building models.

To illustrate, there are devices that allow blind people to use other senses to acquire 'visual' information. The astonishing thing is that after awhile their brains start to map this data visually, despite the stimulus having completely obviated the optical pathway. So it's not about how we perceive the world but rather how are brains build a reperesentation of the world, and the brain largely does that by building models.


Raph said: The urge to separate out the constituent media used to DISPLAY a virtual world and consering definitional in some manner strikes me as putting the emphasis on the client, which is the wrong place to put it.

No. The client is precisely the right place to put the emphasis, because it is mode of presentation the client offers that determines how the player's brain will percevie and assess the information from the world simulator--It will do so either by processing the client's output as descriptive or model-based. This makes a huge difference because model-based and description-based media are truly completely different modes of communication. In fact, they represent about as wide a difference in mode of communication that human history has ever seen.

While I disagree with you on that point, Raph. I agree that it is silly to play the better/worse richer/poorer game. I agree for the very reason that they are such disparate modes of communication. The decision as to which one is appropriate for a given kind of communication or collaboration or experience is a question of design, which is to say it is a question of intention not inherent value.

44.

Barry: I would contend that there are many spoken experiences that simply can't be conveyed properly in written text. Do you have examples where the converse is true (where the written form can convey something properly, but not the verbal)?

Kirk: Calligraphy and typography. The poems of Archie and Mehitabel. Visual puns using homonyms. Tables, indexes. I would argue that emoticons don't translate well into spoken language, and have evolved to an extent that they are different from the physical gestures they represent.

Good voice acting in 3D worlds would be a concrete example of a spoken-word audio channel having a substantive effect on making an offering more world-like. Do you have concrete examples where any of your offered analogs have a similar impact on making text offerings more world-like (or have even played any meaningful role)?

45.

The client is precisely the right place to put the emphasis, because it is mode of presentation the client offers that determines how the player's brain will percevie and assess the information from the world simulator--It will do so either by processing the client's output as descriptive or model-based. This makes a huge difference because model-based and description-based media are truly completely different modes of communication. In fact, they represent about as wide a difference in mode of communication that human history has ever seen.

The point being that the virtual world itself is composed of data that is one, the other, or either. And other types too, frankly.

Consider a web page. Is the web page via a WAP browser or Lynx "fundamentally different" than if through Firefox? The user experience may be significantly different, but (if it's well made) you can access it either way, and the fundamental difference is reduced to a user choice. What's more, it's a user choice heavily dependent on user context--am I at a desktop with decent speakers, or on a phone?

If the argument is "a virtual world which makes use of every possible medium is richer than one that limits itself to one medium" I wouldn't argue much (other than to reiterate that there are often interesting things to be discovered by using self-imposed limitations). I'd also agree that if you have a server-side odel and you choose the wrong display medium, you may well not achieve any of your goals. But neither of these points has anything to do with the definitional question of "fundamental difference."

Perhaps I am misreading Cory's intent in saying that there's a fundamental difference (heck, I am sitting next to him at Supernova right now, so I should just ask him). I went on the record as saying that graphics were coming and would dominate over a decade ago. But I wouldn't claim that this somehow changed the fundamental nature of the medium anymore than the proscenium stage is fundamentally different theater from in-the-round. You make the choice of diplay method based on design criteria, one of which might be artistic choice.

46.

Reading the responses, I get the impression there is an embedded tendency from long-time "text" people to rather defiantly discount the importance of visceral reactions of the kind Barry describes having while playing Half-Life, and I can't help but wonder from where that defiance comes. I'd also add that I find some comments betray what seems to me to be an inability to reconcile learned, real world constraints with the mind-bending possibilities of real world-looking simulations that can violate all those constraints (Molyneaux’s “The Room” experiment for example). Right or wrong, both observations bring to mind Castenada's "Journey to Ixtlan":

The sorcerer's description of the world is perceivable. But our insistence on holding on to our standard version of reality renders us almost deaf and blind to it.

If I might I'd like to modify the original four bullet points:

1) Perception of Place Matters
2) Perception of Doing Matters
3) Perception of Actors Matters*
4) Simulation and Ease of Interaction Matter

*note: by Actors I mean both the Self, other RW Participants, and Artificial Intelligences/NPCs

As an industrial designer it's interesting to me that while the mostly academic arguments are flying here, the real world marketplace is full of corporations desperate to hire "experience designers" to impart upon their products simulated worth. What a curious juxtaposition.

47.

csven wrote:

Reading the responses, I get the impression there is an embedded tendency from long-time "text" people to rather defiantly discount the importance of visceral reactions of the kind Barry describes having while playing Half-Life, and I can't help but wonder from where that defiance comes.

It's the other way around. There's a tendency from people who want to believe they are doing something new to claim that the sort of reaction you're talking about cannot be achieved via text. I've felt bigger visceral adrenaline rushes from text games than any graphical game has ever given me. Completely incorrect assumption, for me at least. I have no problem with the idea that different mediums simply appeal to different people in different ways. To say that one can impart a visceral experience and the other can't is simply incorrect. It's also naive insofar as this action is all taking place in our heads anyway. A bunch of lines on a 2d screen isn't 3d. It's simulated 3d. Your brain interprets it and builds a 3d image for you. Similarly, your brain reads a bunch of lines on a 2d screen in a text game and builds a 3d image for you. Those images are different (I personally find images generated from text to be more beautiful but possibly less concrete) and likely different for different people.

--matt

48.

Barry: Do you have concrete examples where any of your offered analogs have a similar impact on making text offerings more world-like (or have even played any meaningful role)?

Well, that's changing the question as you go along. You didn't ask this question and I didn't answer it. The question you asked was, "Do you have examples where the converse is true (where the written form can convey something properly, but not the verbal)?"

I would add that now that my brain in chewing on this, that there are lots of things which are very difficult to express verbally, but can be expressed easily using some form of symbols printed in a linear fashion on a "page" (paper, projected, or otherwise.) Chemical formula, algebraic equations, biochemical reactions, musical scores, the periodic table of elements. The written word and the spoken word can be radically different media, use the one that best fits the job.

Now, if you are asking about "world-likeness." Well, I have to say, that's your goal not mine. My goal is to help people learn and communicate, and abstract representations, including text, can be a powerful tool to that end. Persistant knowledge in the form of wikis, mail and bulletin boards can be a powerful tool to that end. Discussions like the one we are having now with strong intertextuality that permits us to quote single paragraphs from something that was written hours before are a powerful tool to that end.

It feels like that in your drive for "world-likeness" you forget that the world we inhabit is filled with people creating text: diaries, journals, fliers, "tagging" boxcars, letters to the editor, journals, memos, stickies, forms, shopping and to do lists, calendars, and newsletters.

49.

csven: Reading the responses, I get the impression there is an embedded tendency from long-time "text" people to rather defiantly discount the importance of visceral reactions of the kind Barry describes having while playing Half-Life, and I can't help but wonder from where that defiance comes.

If the claims were limited to the kinds of visceral reactions that Barry describes, then I'd be in full agreement. Yes, more authentic simulations do a better job of provoking visceral reactions in a larger number of people.

But the claims that Cory and Barry are making go beyond that. They seem to claim the fact that you can create a more visceral reaction on its own translates into better learning, better communication, and better collaboration across the board.

Meanwhile, I'm looking at 20, 30 years of research on the relationships between the medium of instruction and learning. I'm not finding much in the way of support for this claim. People have been making these claims since the 1920s, and every time it falls flat. So what I'm seeing here is a lot of magical thinking. If we improve the physics models just a bit, the frame rate just a bit, the fidelity to reality just a bit, a miracle will happen and we will get wonderful learning environments that will work for everything a person wants to learn.

Poppcock.

Snake Oil.

Merde.

Putting people into a 3D environment doesn't mean squat where communication,collaboration and learning are concerned if they are still using the good old IRC-like interface to talk to each other. I just ran the numbers on a fresh batch of data this afternoon. And my 3D chat sample is entirely consistant with older analyses of IRC and muds. Messages sent through the internal email system are shorter and less complex than older analyses of email. If the 3D-ness is making a difference in terms of quality of discourse and collaboration, I'm not finding it.

It doesn't mean squat if you don't have norms and social structures to encourage communication, collaboration and learning. We've learned this lesson over and over again with usenet, webboards, wikis, and text-muds. Simply opening up a system and advertising the address is not going to create a community.

If you want to sell me on your ability to foster communication, collaboration and learning, don't talk at me about physics models, pixels, fidelity and granularity. Talk with me about pedagogy, message design, democracy, norms, values, social structures, economies, communication, history, language and knowledge-building. I have no doubts that people in Second Life are creating something wonderful. I just don't think it has that much to do with "world-likeness."

50.

If you want to foster communication, collaboration, learning, community, participation, or even adrenaline rushes or other forms of engagement, you must first and foremost be able to attract to and keep people in your virtual world. For whatever reasons, good or bad from an individual's POV, graphical environments have decisively demonstrated -- by multiple orders of magnitude -- their advantage in attracting and keeping more types and larger numbers of people interested.

That is, millions of people play graphical games. Perhaps thousands -- maybe tens of thousands if we're being generous -- play text-based games. Arguments from personal experience any of us may have had faovring either pale in the light of this reality.

I have to say that I'm a bit surprised to see this kind of rhetorial fistfight ("poppycock - snake oil - merde") going on here on Terra Nova, in 2005. I thought that this Sisyphean battle had grown stale even on Usenet years and years ago.

That some people are deeply invested in text-as-environment while others are just as deeply invested in graphics-as-environment is obvious. The larger question that comes to mind is: so what? Does this really make any difference in the understanding, design, or construction of virtual worlds?

51.

Raph said:

If the argument is "a virtual world which makes use of every possible medium is richer than one that limits itself to one medium" I wouldn't argue much (other than to reiterate that there are often interesting things to be discovered by using self-imposed limitations).

I'd also agree that if you have a server-side odel and you choose the wrong display medium, you may well not achieve any of your goals. But neither of these points has anything to do with the definitional question of "fundamental difference."

I don't know about others, but that isn't the argument I was making.

It's not a question of just splashing about various media and saying "more is richer". Rather, it's a question of looking at two fundamentally different approaches to communicating and updating the 'states of affairs' of the world simulation. Communicating world-state changes primarily by text encourages the player focus on taking the text state descriptions and turning them into an internal model of the game world. When you use a model to communicate world states, say by using a 3D-graphical environment with physics, discrete objects, and sophisticated camera control, you are encouraging the player to take an entirely different approach to building their internal representation of the game world.

What I take as fundamental then, and why I made the point I did, is that because descriptive presentation (whether linguistic, visual, musical) engages the brain in an entirely different manner than a model-based method, the two means are fundamentally different. This difference, I argue, grants text-based worlds a fundamentally different kind of worldliness.

As I said earlier, it's like the difference between reading about the orbits of the planets in our solar system or using an orrery. Both can give you a sense of what the solar system is like, but they impart that sense in fundamentally different ways.

In the case of virtual worlds, as designers chose which to employ or mix and match, they are making choices that affect the kind of 'worldliness' that users will experience. I think Cory's argument is that perhaps model-based communication can convey 'worldliness' more directly because models are what we use to construct reality. It's not the graphics themselves or the audio or the text; rather it's that all these are presented in space like SL in the form of a roughly coherent model. And brains love models. Especially the brains of people who like videogames.

The issue is not about the use of graphics versus text, imo, but about the use of description versus model-based communication.

Kirk said

If you want to sell me on your ability to foster communication, collaboration and learning, don't talk at me about physics models, pixels, fidelity and granularity. Talk with me about pedagogy, message design, democracy, norms, values, social structures, economies, communication, history, language and knowledge-building. I just don't think it has much to do with world-likeness.

But it does. Knowledge building via models versus knowledge building via propositions/descriptions engender very different kinds of knowledge. In turn, the kind of knowledge you get from the game world and how you get that knowledge play a very large role in determining what features of the world you find worth pursuing, worth investing in, worth cooperating for, worth proscribing, and worth fighting for.

52.

Kirk wrote:

Well, that's changing the question as you go along. You didn't ask this question and I didn't answer it. The question you asked was, "Do you have examples where the converse is true (where the written form can convey something properly, but not the verbal)?"

Yes, but you've elided away the context in which I asked the question. I invite you to re-read the entire post in question, and take it in the context of the entire series of posts.

My primary argument has been from the perspective of constructing virtual worlds. When I made the initial observation regarding the spoken word conveying meaning, it was in the context of virtual world creation. I'm sorry if that was somehow unclear.

My question regarding textual expression was therefore directed at virtual world creation as experienced through a text interface. Specifically, the sentence two above the one you quoted:

"Is there anything that you can do with text that can't be done with a combination of visual and audio input (the spoken word, for example)?"

That question was asked specifically in the context of virtual world creation, and the split between text-only virtual world representations and audiovisual representations.

I was trying to make a sincere observation regarding the distinction between the claimed properties of text, and whether the claimed properties more accurately belonged to human language rather than just the written manifestation of that language.

Since spoken language has a massive overlap with written language (with respect to the ability to evoke and communicate concepts), then if the claimed properties truly belong to the language rather than simply text, it would seem logical that we would be able to accomplish much by integrating human speech into the world-making paradigm.

Since we have physical sense organs dedicated to processing audio and 3-D visual data, but no sense organs dedicated to the absorption of textual language (and survived without them for an exceedingly longer time period than that during which we've been using text), it would make sense to me that conveying the sensation of a realistic world would be more strongly integrated with the senses we actually have.

Sight and sound aren't the entire sensorium, of course, but conveying meaning through text is abstracted pretty severely from our sensory integration with reality. We can imagine and model a set of concepts, but I contend that they will likely and largely not be perceived to be as real by us. We have hundreds of years of learning fighting a vastly larger history of sensory evolution.

That being said, realism is not always the most effective paradigm for conveying certain concepts. Sometimes, it's a downright lousy tool.

Kirk later wrote:

But the claims that Cory and Barry are making go beyond that. They seem to claim the fact that you can create a more visceral reaction on its own translates into better learning, better communication, and better collaboration across the board.

Please retract that statement. It is a misattribution of my position. Cory's claims are not my claims. I have been trying to speak to the aspect of how world-like the different approaches are to me. To the best of my knowledge, I don't speak to Cory's claims of better learning, better communication or better collaboration in this thread at all.

I would appreciate not being mis-characterized in that fashion. I have tried to be exceedingly careful to not speak to those aspects of the argument. I have instead endeavored to draw distinctions between the types of communication and interaction that are taking place in each realm.

I don't believe that I've claimed anywhere here that experience is inherently superior to intellect or imagination. I would not try to support such an assertion. Each are fundamentally different tools.

My primary contention has been that conveying tightly constrained world-like information may map more strongly to experiential/sensory conveyance than it does to intellectual/imaginary conveyance.

The choice of which mode is most appropriate for any given world-message would then seem to fall to whether it is more valuable to leverage the disparate and divergent imaginary pictures that individuals use, or to converge towards a shared experience with less deflection of individual interpretation.

I most definitely do not claim one is superior to the other... simply that there are many aspects where neither will suffice as a substitute for the other.

Text is not the end-all-be-all that trumps everything that graphics and sound can do. Graphics and sound are not the end-all-be-all that trumps everything that text can do.

IMO, there's been more than enough denigration of each camp by the other. There are special and amazing things available in 3D space + sound worlds which yield particular experiences that aren't available in textual worlds.

Accordingly, 3D worlds and their creators' efforts don't deserve the dismissal of "text worlds with pretty pictures".

Likewise, there are powerful and evocative forms of communication (and conceptual boundary-breaking and concept-play as well) that are available in the combination of human language and computer code, which we often see today in text-interface worlds. Many of these things simply cannot be conveyed in the same manner using 3D graphics.

Accordingly, text worlds and the efforts of their creators do not deserve denigration either. Each are powerful tools for their own forms of communication.

My own bias with respect to the concept of "world" is rooted purely in my own philosophical take on what the term "world" conveys to me. Stating that 3D + sound worlds can be more world-like for me is not an indictment of the text genre... it is simply a statement of my own narrow perception of what "world" means.

My intention is not to offend those who traffic in the conceptual and the linguistic by mapping the concept of "world-like" to the concept of "emulating a constrained reality". That's simply how I process the terms.

For me, textual worlds today utterly rule the realm of what I would call "virtual fantasy". 3-D plus sound (plus other inputs) will likely eventually rule the realm of what I map to the concept of "virtual reality".

Each have their own place in my perception of the future. Neither deserve to be treated shabbily by proponents of the other paradigm.

Is there room to set aside the cross-factional back-biting and embrace the strengths that each approach offers us when we want to convey something using our virtual toolset?

Please?

53.

Matt wrote: To say that one can impart a visceral experience and the other can't is simply incorrect.

I don't recall saying that. My words were: "discount the importance of visceral reactions of the kind Barry describes".

I would not claim there are no visceral reactions from using textual interfaces. However I would propose that there is a different sensibility at work while employing text which to some measure informs how and to what degree we instinctively react. I believe it's already been suggested - and I tend to agree - that reading is primarily an intellectual activity, whereas operating within a 3D simulation which mimics the real world is more an instinctual activity. As evidenced by questionable claims that some textual experiences cannot possibly be simulated in 3D, both the claim and the apparent inability to separate real world life experience from something that only simulates the real world serve to only reinforce my opinion. I'm therefore left to believe those comments are coming from the gut and not the mind; hence my comment. Furthermore, this suggests to me that there is a potential issue with spatial cognition, revealed in what I consider overly-broad claims that something is "impossible" in a virtual medium.

Matt wrote: It's also naive insofar as this action is all taking place in our heads anyway. A bunch of lines on a 2d screen isn't 3d. It's simulated 3d. Your brain interprets it and builds a 3d image for you. Similarly, your brain reads a bunch of lines on a 2d screen in a text game and builds a 3d image for you.

That's a very academic, intellectual argument; however, as stated in my original post, I believe that approach too casually dismisses our instinctive responses. When I get in my car and drive down the street with the windows up, my visual interface to the real world is really only the light refracted through the glass. Is that only "simulated 3d"? And regardless of what it is, is it more like the 3D image on my monitor or more like a text image? And to which is the average naive person more likely to instinctively respond?

While we may not react the same way now, I'd remind you it wasn't too long ago that movie patrons stampeded out of movie houses at the sight of a celluloid train coming at them on the screen. I don't recall similar occurences in history with clay tablets, scrolls or books (and I'd distinguish reactions based on individually-learned interactions with the world - big object hurtling toward the user - from reactions derived from human communication of ideas - the gods voices are captured in this object therefore...).

Kirk Job-Sluder wrote: They seem to claim the fact that you can create a more visceral reaction on its own translates into better learning, better communication, and better collaboration across the board.

I'll speak from my own recent experience.

I've been toying around with Chris Anderson's Long Tail curves which are at this point, to my knowledge, all presented as 2D graphs. While recently attempting to communicate an idea in 2D he made what I considered some non-intuitive graphs, and as someone who thinks and works in 3D, I saw a better way to represent the data and subsequently generated a quick set of 3D renderings.

Later that night I was inworld discussing this with another avatar who was unfamiliar with the Long Tail idea. Rather than try to explain via voice or text, I uploaded 2D images of the 3D renderings. Of course that didn't completely explain the idea or what I was doing, so in the course of texting additional explanations I started constructing a 3D representation as an aid. But at that point the situation changed because in the course of explaining and playing with the simulated 3D object, another unique and unexpected idea presented itself. This idea would not likely have developed on paper as either written word or sketch. It was more likely because I had a simulated 3D object which I could manipulate as in real life that allowed this progression. So in the end not only did the other person more fully and intuitively understand Mr. Anderson's theory(?), they also understood what I had done with his work and why I did it. And in addition - and most importantly - I had a new, unique solution in which they participated.

As a designer, this kind of activity is common in the real world. Industrial designers may start with pen and paper, but much of the breakthrough work occurs during the modeling stage. Furthermore, when Sales and Marketing go to pitch their new products to their customers, they don't want me to write a paper on how the product functions or why it looks as it does... they want a model they can see and touch and feel comfortable presenting.

Richard wrote: Nor does it follow that more technology generates better imagination. The difference is that I'm not claiming the former, but the pro-3Ders seem to be claiming the latter.

I wouldn't claim that at all. As a professional 3Der who is a "pro-3Der", I believe what we'll actually find is that people will initially come to realize how little imagination they really have and perhaps react negatively. On the whole they might very well be less creative. Staring at the proverbial blank piece of paper isn't any fun. In addition, I've recently remarked to friends that the game modding community seems to have significantly diminished since the advent of high-quality content; the average person just can't keep up. Having a "thick skin" is another thing many will have to deal with; possibly for the first time. Not many people like a critic.

54.

Mike Sellers: That is, millions of people play graphical games. Perhaps thousands -- maybe tens of thousands if we're being generous -- play text-based games. Arguments from personal experience any of us may have had faovring either pale in the light of this reality.

I don't think this really matters when we are looking at proof of concept. The fact of the matter is, text-based muds were being used in classroom and professional development settings before MMORPGs, and proved that medium can support meanigful collaboration, creativity and learning. Some MMORPGs proved that medium can suport meaningful collaboration, creativity and learning. A small community of 20 can just as easily prove this as a massive community of 20,000.

That some people are deeply invested in text-as-environment while others are just as deeply invested in graphics-as-environment is obvious. The larger question that comes to mind is: so what? Does this really make any difference in the understanding, design, or construction of virtual worlds?

Well, I don't think this is really about text-based worlds vs. graphics-as-environment. I think this is about medium vs. method, and I think there is a great deal at stake here. It's about what works and what doesn't work, and I think the research is pretty clear that media alone won't work.

In terms of understanding virtual worlds, I think that if we drop the assumption that the shift to 3D representation makes these worlds fundamentally different from what came before, then we have a much richer history of both interesting behavior worth talking about, and best practices worth applying to future worlds.

In terms of design, I think we need to start with a utopian vision of what type of community you want to create, and then develop the structures within that community needed to make it happen. If you want for people to sell their creative works, you need a marketplace. If you want to create a community within a community, having factions and guilds will help. If you want to support a community history, you need some form of persistent text system. You need to establish the ethos, tone, and norms of the community. You need to consider how to deal with crime and misbehavior. How do politics work in your community?

mikesan: This difference, I argue, grants text-based worlds a fundamentally different kind of worldliness.

I don't think worldliness is enough. (Or even necessary. See wikipedia.) You take 5 flesh and blood volunteers and put them in a waiting room with no decoration. Don't give them any instructions or reasons for being there, just make them wait. Pick another 5 volunteers and put them in a room with a pizza and tell them to help themselves. Give them five separate sheets of paper and have them make lists. Give them one and have them make one list. Put a TV in the room. I've just described ways to create five different behavior settings. And I think that ultimately, the kinds of behavior settings you create are more critical than the fidelity of representation.

Theatre stages provoke the same behaviors in "real life", text muds, and 3D spaces. So do hottubs, so the designer has to make some wise decisions before putting a hottub on the stage.

55.

What I take as fundamental then, and why I made the point I did, is that because descriptive presentation (whether linguistic, visual, musical) engages the brain in an entirely different manner than a model-based method, the two means are fundamentally different. This difference, I argue, grants text-based worlds a fundamentally different kind of worldliness.

As I said earlier, it's like the difference between reading about the orbits of the planets in our solar system or using an orrery. Both can give you a sense of what the solar system is like, but they impart that sense in fundamentally different ways.

Yes, of course. And my point is simply that "virtual worldness" is a simulation of the orbits of the planets that can be described either in text or in an orrery. Often both at the same time even. And the understanding that comes from THAT is even richer still.

But I would never say that the description of orbits via textually portrayed mathematics or via an orrery changes the orbits themselves. A game of baseball heard on the radio versus seen at the field versus seen on TV is still the game of baseball, despite the many many differences in the user experience.

The server is where the game of baseball happens. This is then intermediated by a method of presentation. Does this method of presentation radically change the experience? Indubitably. Is the elephant still an elephant regardless of what part the blind men touch? Without a doubt.

Neither text nor graphics feature anywhere in the core definition of virtual world, just as neither text nor pictures feature in the core definition of "book" and just as neither radio nor TV feature in the definition of "baseball."

To harken back to the Magritte reference, we seem to be arguing about whether or not it's a pipe--the point I am making is that it's still a painting. As such, it still has a LOT of lessons there to teach us even if we're working in cubism rather than surrealism--which are modes of perception.

That's why the lessons of text games are still hugely relevant, and also why graphics has a whole set of its own new lessons to teach--and yet, fundamentally, both are merely elucidating virtual worlds, which are not a client-driven medium any more than baseball is.

56.

Barry: Retracted with my apologies.

In general: I'm not championing text, I have experienced heart-wrenching failures of text-modal communities. Nor am I attacking 3D, I think that 3D is beautiful. My point is that the most incredible representation of an object or concept, no matter what the medium, is not going to be effective unless you provide a context and a motivation for engaging with that object.

I think Cory is really selling Second Life short when he attributes the innovative grinding of Second Life to it's 3D-ness. 3D CGI is a well explored field that does not seem to be producing many surprises. Building rich communities is only a handful of steps beyond neolithic beer making, (throw everything in the pot, and hope for the best.)

57.

Raph> The server is where the game of baseball happens. This is then intermediated by a method of presentation.

Exactly, I should have been more careful in describing the skateboard and fish examples as things which would have been impossible without a 3D world representation rather than as impossible with text.

You could develop a 3D virtual world with a skateboard game, access it with a text client and play the same game of timing jumps and grinds. It would be pretty hard though.

I think one reason the confusion arises is that 3D graphical clients are nearly always used to render 3D worlds and that text clients are often used to render node based worlds. Hence 3D worlds are called graphical worlds and node based worlds are called text worlds.

Raph's examples, the text rendered skateboarding example and MASSIVE-3s 3D graphical rendering of node based worlds show it doesn't have to be the case though.

58.

Richard says:

Besides, it depends on the surgeon and the quality of the simulation. Current 3D systems are lousy at modelling interiors, and if the surgeon had actually seen the inside of a patient before then an interactive text might be more suited to their training than 3D graphics.

I agree with a lot of what you've written on this subject, but I have to call a point of fact here. There's been a huge amount of work on building surgical simulators, and the ones with haptic feedback are the best, but I believe even the non-haptic 3d simulators have had significant success.

As an aside, I'm not really sure what you meant by "Current 3D systems are lousy at modelling interiors."

59.

Raph said:

The server is where the game of baseball happens.

And that is a point where we disagree. The world simulation runnning on the server is a contingent fact of current technology. There is nothing about virtual worlds that require this be so.

And I don't see how one can simply dismiss the mode of presentation to simply an 'intermediary' between the virtual world and the user. The mode of presentation is as surely as much a part of the virtual world as the simulation that drives it.

We don't interact with the laws of physics in our world except through intermediaries, yet what seems most real is our interaction with these very 'intermediaries'. We don't consider the world to solely consist in the rules and dynamics that govern it. Why should this be any different for virtual worlds?

While I have been agreeing with much of what you're saying, Raph. I do disagree with the notion that the virtual world can be located as being 'on the server'.

60.

The world simulation runnning on the server is a contingent fact of current technology. There is nothing about virtual worlds that require this be so.

Even on the day when virtual worlds are done in a purely P2P fashion, there will still be a virtual server and a virtual client. It's fundamental to the architecture of virtual worlds that there be a persistent simulation, and that's what the shorthand of "the server" really means.

And I don't see how one can simply dismiss the mode of presentation to simply an 'intermediary' between the virtual world and the user. The mode of presentation is as surely as much a part of the virtual world as the simulation that drives it.

It's simple to disprove this by simply doing the basic thought experiment of having multiple different clients for the same virtual world. If you choose to think of each of these interfaces as being intrisic to the virtual world, that's fine--we just need to realize that ALL virtual worlds have this capability, just as web pages do.

One of the biggest problems with VRML as a virtual world platform was its focus on rendering; one of the strengths of HTML is its independence from clients. At some point, we'll have a VW that can be accessed via text on a phone, via haptics and immersive VR, via 3d, via a 2d isometric view, and who knows what else. Client-independence is to my mind an inevitable development in virtual world design.

We don't interact with the laws of physics in our world except through intermediaries, yet what seems most real is our interaction with these very 'intermediaries'. We don't consider the world to solely consist in the rules and dynamics that govern it. Why should this be any different for virtual worlds?

Excellent analogy. I am saying that the laws of physics are the laws of physics regardless of whether your perceptions argue for an understanding of mass or the convenient fiction of weight. The "server" of the laws of the physics has remained the same, a discrete entity, whether or not our intermediaries for forunderstanding those laws were Archimedean, Newtonian, Einsteinian, Heisenbergian, or whatever. Each of these has given us different textures to our experience--in fact, we usually move through each of these stages during our lives even though we all "know" that the latest is "the way it is."

61.

The world simulation runnning on the server is a contingent fact of current technology. There is nothing about virtual worlds that require this be so.

Even on the day when virtual worlds are done in a purely P2P fashion, there will still be a virtual server and a virtual client. It's fundamental to the architecture of virtual worlds that there be a persistent simulation, and that's what the shorthand of "the server" really means.

And I don't see how one can simply dismiss the mode of presentation to simply an 'intermediary' between the virtual world and the user. The mode of presentation is as surely as much a part of the virtual world as the simulation that drives it.

It's simple to disprove this by simply doing the basic thought experiment of having multiple different clients for the same virtual world. If you choose to think of each of these interfaces as being intrisic to the virtual world, that's fine--we just need to realize that ALL virtual worlds have this capability, just as web pages do.

One of the biggest problems with VRML as a virtual world platform was its focus on rendering; one of the strengths of HTML is its independence from clients. At some point, we'll have a VW that can be accessed via text on a phone, via haptics and immersive VR, via 3d, via a 2d isometric view, and who knows what else. Client-independence is to my mind an inevitable development in virtual world design.

We don't interact with the laws of physics in our world except through intermediaries, yet what seems most real is our interaction with these very 'intermediaries'. We don't consider the world to solely consist in the rules and dynamics that govern it. Why should this be any different for virtual worlds?

Excellent analogy. I am saying that the laws of physics are the laws of physics regardless of whether your perceptions argue for an understanding of mass or the convenient fiction of weight. The "server" of the laws of the physics has remained the same, a discrete entity, whether or not our intermediaries for forunderstanding those laws were Archimedean, Newtonian, Einsteinian, Heisenbergian, or whatever. Each of these has given us different textures to our experience--in fact, we usually move through each of these stages during our lives even though we all "know" that the latest is "the way it is."

62.

Raph wrote: A game of baseball heard on the radio versus seen at the field versus seen on TV is still the game of baseball, despite the many many differences in the user experience.

Is it? I'm not so sure. I tend to think that there's a kind of Heisenberg Uncertainty Principle at play here, meaning that the method by which each individual experiences something necessarily changes their definition of the experience. What is a "game"? If the "game" is only the statistics then I'd probably agree. But to me that seems a rather limited definition of "game" and by extension seems to put human experience as a whole in a kind of convenient box. I can't help but notice that the way you phrase your comment points to this potentially fundamental difference in how our views might differ: you superset "game" and subset "user experience". I do just the opposite.

Raph wrote: Is the elephant still an elephant regardless of what part the blind men touch? Without a doubt.

Really?

In my way of thinking, within the context of this discussion the only relevant issue is what the individual, in this case the blind man, perceives. What you or I might call "an elephant" may in fact be part of a much larger whole which we do not perceive. For example wrt elephants, it's been only recently learned they communicate at levels well below our human sense of hearing; finally explaining the so-called "mystery" surrounding elephant ESP! So in at least one way - the audible sense - sighted humans have not entirely perceived elephants. So who is in a position to tell the blind man he doesn't fully comprehend an elephant?

63.

I think a big issue here is that we are talking about two different design axes:

abstract representation <-> concrete representation
low sociability <-> high sociability

If we just look at systems with highly concrete representations, we find that even within 3D environments there is a huge range of sociability. I found Beyond Good and Evil to be an amazing game, both in terms of plot scripting and the feeling that I was inhabiting an alien planet. I feel the same way about Morrowind in that there is an amazingly beautiful universe out there to explore. However, because these were designed with low sociability, and few structures for creating social relationships except with AI, the potential for creativity and collaboration are minimal.

If we fix the sociability axis, we find that there are text and 3D environments that have managed to create extremely high degrees of social interactions, complex networks of people working towards common goals, markets, cognitive mentorships, and communities of practice. And I suspect that the communities created in abstract/high-sociability worlds are not that much different than those created in concrete/high-sociability worlds.

64.
Neither text nor graphics feature anywhere in the core definition of virtual world, just as neither text nor pictures feature in the core definition of "book" and just as neither radio nor TV feature in the definition of "baseball."

The assumption of the interface affects the internal represation of the game on the server. So I would contentd that it usually does feature in the core definition of the game.


I think one reason the confusion arises is that 3D graphical clients are nearly always used to render 3D worlds and that text clients are often used to render node based worlds. Hence 3D worlds are called graphical worlds and node based worlds are called text worlds.

Exactly what I was thinking, Jim.


A game of baseball heard on the radio versus seen at the field versus seen on TV is still the game of baseball, despite the many many differences in the user experience.

Its not the same game if you are trying to participate in it rather than just observe aspects of it. I would hesitate to categorize virtual worlds as such a one way communication as a radio or television broadcast.

I would also hesitate to cast it as the same game if the "observed" baseball game was one I was actually interacting with, ie MLB 2005. But, don't take my word for it, diagram the two and see if they are the same.. =)


65.

So in at least one way - the audible sense - sighted humans have not entirely perceived elephants. So who is in a position to tell the blind man he doesn't fully comprehend an elephant?

I have to point out that many of us around Terra Nova actually BUILD "elephants." The elephants often behave in ways we did not expect or predict, and we don't pretend to understand everything about elephant culture, but it's nonetheless generally pretty clear to us how the elephant is built. That may be part of the gap here.

And that may be why your comment:

I tend to think that there's a kind of Heisenberg Uncertainty Principle at play here, meaning that the method by which each individual experiences something necessarily changes their definition of the experience. What is a "game"? If the "game" is only the statistics then I'd probably agree. But to me that seems a rather limited definition of "game" and by extension seems to put human experience as a whole in a kind of convenient box. I can't help but notice that the way you phrase your comment points to this potentially fundamental difference in how our views might differ: you superset "game" and subset "user experience". I do just the opposite.

just feels wrong. It's odd (and very deconstructionist) to tell a game designer that they are not perceiving what they are making. It's true in all sorts of academic ways, but from a practical point of view, it's, uh, annoying, actually. :)

I've been engaged in many many discussions about whether or not you should approach design from the user experience point of view or the systemic point of view. It's sort of a silly debate, because of course, the two are intertwined rather intimately. At the core, the design for each responds to and changes based on the constraints of the other.

Which is the superset is a semantic debate that is not worth getting into. There is no doubt, however, that we can separate the experience design from the game (or virtual world) design. You can make a chess game where the pieces are Lord of the Rings themed or blobs of snot. The identifier will remain "chess" and the fact that human language works that way is, I think, a significant clue as to how the human mind classifies experiences.

A game concept came to me--I wanted to play a game which felt and looked like a kaleidoscope. That was an experience-driven concept. Then I started working out how that would work, and ended up with a system whereby there were four overlapping planes where pieces could fall--it looked like a grid, with "walls" on the horizontal and vertical axes.

I designed the basic systems of the game on a wooden go board, using glass beads. I prototyped core mechanics that way, in turn-based fashion, even though I always saw it as an arcade game. I played it with my daughter head to head, even though it was intended as a single-player experience. I was still feeling out the shape of the elephant.

Then I implemented it on a computer. I did it with colored squares on a pixel grid first. I tried several methods of bringing it into real time. You clicked on a square to place a bead even though the actual rule was that the bead slid in from the edges. Once I found three modes I was happy with, I put pretty graphics on it--that looked exactly like a wooden go board and glass beads.

In playing with it, it became clear that the clicking interface was wrong--the interface needed to mimic the system more, because otherwise people were confused. The glass beads and wood board made it seem old-fashioned and slow, but the music was a perfect fit.

I redesigned the interface again. The beads changed to an assortment of pretty anmating jewels. The board changed into abstract blue glowing lines, and the method of plkacing beads involved using keys to slide a cursor around the edge of the board. Hinting functions went in to indicate where the bead would land.

Then I let others play it, and they immediately discovered strategies I had not understood; they found interface flaws I had not anticipated; they exploited subtle rule issues, causing me to layer on more rules and restrictions in order to preserve the challenge I wanted.

The point I need to make here is that the "game" didn't change. It got immeasurably better. But it was a process not unlike whittling off the excess wood, sanding the mechanics down so that it was smooth and polished, making it work at its best. I could show you the wood board and beads protoype, and the finished version--and for that matter, he cell phone version versus the PC version--and you would instantly know they were the same game.

So when you say that the user experience matters and is paramount, I go "well, of course. It drove essentially all development of the game past initial prototype." But when you say it DEFINES the game, I say "not even close." From where I sit as the elephant crafter, the difference is extremely clear.

The assumption of the interface affects the internal represation of the game on the server. So I would contentd that it usually does feature in the core definition of the game.

It affects the internal representation of the game because we choose that it do so. Trust me; we could maximize or minimize that effect in our designs. But it's a choice. Designing with no interface assumptions (or very few) is difficult, but possible. It may not be WISE, but it's possible.

Its not the same game if you are trying to participate in it rather than just observe aspects of it. I would hesitate to categorize virtual worlds as such a one way communication as a radio or television broadcast.

The game of baseball (actual baseball) when experienced by the Major League player on the field and the spectator at home on TV, are the same game, experienced from radically different positions. They serve different entertainment needs, but any given interactive artfact will have those two dimensions, the participatory or the voyeuristic. Different games will work well at one, the other, or both, as we have all experienced. I recently designed a puzzle game that moves in such a way that you literally cannot see what the player is doing, but which is nonetheless rather beautiful to watch (it also looks a bit like a kaleidoscope). Both are valid experiences of the system. The fact that I have a different experience of a rose than a botanist does does not render either experience invalid, nor does it mean that the rose is less a rose.

Say rather, that we're all blind reaching for the Platonic ideal of the elephant or the rose.

I would also hesitate to cast it as the same game if the "observed" baseball game was one I was actually interacting with, ie MLB 2005. But, don't take my word for it, diagram the two and see if they are the same.. =)

I don't need to; they are not the same game. Just a look at the basic atoms of interaction tells me that, because the choices made are radically different. One is an abstraction of the other.

Ironically, the voyeuristic experience of both may be almost identical, and as display technology improves, they may become indistinguishable from that perspective. Nonetheless, the games of MLB and actual baseball will be different despite their similarity from the observer's point of view.

Identity of two objects from an observer's point of view is always because the observer does not know how to look deeply enough. I suppose this is the point that the "fundamentally different" camp is making; but let us not confuse an instance for a type.

66.

And to add yet one more perspective to the mix....

For me, this has become a series of questions related to whether there's a fundamental difference between linguistic perception and spatial perception.

Isn't our processing of verbal information fundamentally different from how we process spatial information? Isn't there plenty of MRI and neurochemical research indicating that different parts of the brain process verbal and spatial information?

Does creativity depend on perception? Is linguistic creativity fundamentally different from spatial creativity?

If so, why then is it so unreasonable to guess that text experiences (which rely on linguistic perception) will tend to elicit and reward linguistic creativity, and that 3D experiences (which rely heavily on spatial perception, even of a simulated model) will tend to elicit and reward spatial creativity?

If there are more people playing computer games who are spatially-oriented than those who are linguistically-oriented, wouldn't that explain the popularity of 3D games over text games?

If there's nothing inherently wrong with preferring spatial perception to linguistic perception (or vice versa), why must there be something wrong with preferring 3D games to text games (or vice versa)?

The only thing that would be wrong would be if the spatially perceptive majority concluded that there is something wrong with linguistic perception simply because those who prefer the latter are a minority, and on the basis of that belief decided that all games should be 3D-only.

Does anyone think this is happening?

And yes, it's possible to engage spatial perception through words. Nodal text games did this all the time... but in playing such games, did you ever draw a map on a piece of paper? Why? Wasn't it because spatial information is better represented through a spatial display system?

So, back to Cory's original question: Do graphical systems allow more creative play (or better learning, etc.) than text? That now seems doubtful... unless we're talking about communicating information that is highly spatial, in which case maybe graphical systems really are better.

But the same must be true for highly verbal information, for information that requires or benefits from linguistic processing, in which case communicating that information through text will probably be more effective.

When I want an experience that continuously engages my spatial perception, I prefer graphics over language. Sim City allows you to position zoned blocks on a 2D graphical space -- could any text description of this process be as effective at conveying so much spatial information so elegantly?

When I want an experience that stimulates my language enjoyment center, I prefer language over graphics. "Now is the winter of our discontent made glorious summer by this sun of York" -- could any 3D/mime version of this speech carry all the nuanced meaning of the language that Shakespeare uses to communicate Richard's jealousy?

Does anyone still feel that one of these approaches/tools is always "better" for all goals?

--Flatfingers

67.

Raph wrote: I have to point out that many of us around Terra Nova actually BUILD "elephants."

What an unfortunate comment. Especially since it suggests that you've made an assumption about me in order to shame and silence an opinion with which you don't agree.

Perhaps I was mistaken about the quality of discourse on this site.

68.

I made the comment purely to put context to the rest of the paragraph; no intent to shame whatsoever. I'm sorry you read it that way.

I admit that I do make the assumption that most folks that I don't know who post in the comments do not work as developers of virtual worlds. It's unfair of me, despite the fact that statistically it tends to be true. I don't know you at all. I did make that assumption, and that's why I then provided the extended example of a game development process to illustrate my point.

If I were trying to silence the discussion, I certainly wouldn't have bothered with the extended example.

I don't know what to make of your final comemnt, but clearly I got you upset. I apologize for that, but still stand by the argument I made. :)

69.

Raph said:

Even on the day when virtual worlds are done in a purely P2P fashion, there will still be a virtual server and a virtual client. It's fundamental to the architecture of virtual worlds that there be a persistent simulation, and that's what the shorthand of "the server" really means.

The thing is your argument about where the 'world' is located hinges presently on being able to discretely identify a server in time and space and then claim 'that's the server'. That distinction is unlikely to remain robust in the future, and moving from a rigidly designated term to a functional metaphor (which is really what you mean by 'virtual' with respect to servers) may change the truth conditions of your claim or even make it meaningless.

Still, I'll admit that is picking nits and not central.

Raph said:

It's simple to disprove [that 'clients' are part of the virtual world] by simply doing the basic thought experiment of having multiple different clients for the same virtual world.

All this shows is that it's possible to have multiple different clients access the same world simulation. It doesn't show that they are all accessing the same virtual world. Your thought experiment begs the question because it assumes precisely what I am challenging--namely, that 'world simulation' is a synonym or identical to 'virtual world'.

In addition, I find it odd that you of all people (and I mean that with admiration not snideness), are ignorning the most interesting parts of virtual worlds in your effort to make them coextensive with 'world simulations'. What part of the world simulation contains the relationships among users, the social structures, the ethical dimensions, the very users of a virtual world itself even. Any definition of a 'virtual world' would have to include those as very real features of the world. Yet certainly these don't reside 'on the server' in any meaningful sense.

Raph said:

I am saying that the laws of physics are the laws of physics regardless of whether your perceptions argue for an understanding of mass or the convenient fiction of weight. The "server" of the laws of the physics has remained the same, a discrete entity, whether or not our intermediaries for forunderstanding those laws were Archimedean, Newtonian, Einsteinian, Heisenbergian, or whatever.

The claim that 'the laws of physics are the laws of physics regardless of [whether there are minds to perceive them]' is reasonable. But it doesn't follow that there is anything like a 'server of physical laws' in reality or even that the laws of physics have any claim to existence outside of their instantiation.

Your extension of my analogy, then, presumes a pretty hardline realism about the nature of physical laws that is not consensus among philosophers of science or even physicists (particularly empiricists of which there are still many). For example, if the laws of physics really were a 'discrete entity', as you suggest, by what interaction do they govern the evolution of the universe? How does the server tell a ball on earth to fall. Is that ball simply part of the laws of physics? I know of very few philosophers or physicists who would claim that our reality, our 'world' simply is the laws of physics.

In any event, just because the real world is mind-independent (which is also debatable, though I don't particularly endorse it), it does not follow that virtual worlds are as well.

The point of my analogy was that perceptions are not simply 'intermediaries' between ourselves and some mystical 'external world'. Rather our perceptions (and of course ourselves) are just as much features of the world as anything else in reality, including the laws of physics. And I don't see how that fact changes in the move to virtual worlds, particularly if, like myself, one views virtual worlds as mind-dependent in a way that the real world is not.

Bottom line, I don't think that 'virtual world' is equivalent to 'world simulation'. As a result, I don't find your argument that the client is not part of the virtual world compelling based on what you've said so far. I also think your identification of 'virtual world' with 'world simulation' implies something that even you don't believe--namely, that the social dimension of virtual worlds are not truly a part of the virtual worlds they exist in.

On a side note, I think people misinterpreted what I meant by 'worldliness'. I don't mean the term to carry an existence claim. When I use the term 'worldliness', I am merely referring to the 'what it's like to be in that world' quality of a given representation. It's not a value claim. I try to stay away from those :)



70.

Monkeyan>The thing is your argument about where the 'world' is located hinges presently on being able to discretely identify a server in time and space and then claim 'that's the server'.

My perception of what Raph was saying about "the server" is essentially concerning that which is collectively shared. In a P2P sense, we can all have copies of what would have been provided by a server, but the shared environment can still be considered "the server."

In the sense of the term "world", there is that which is defined and that which emerges through interaction within the defined. Since we are talking about games, this is the difference between the rules and the scope. Regardless, it is still the game. In the world, physical laws, as given in Raph's example, are part of the rules, and culture is part of the scope. I think your argument is largely semantics based on the subjective use of the term "world".

The only reason I point this out is that I just read a lot of posts that seem to be arguing semantics.

Monkeyan>Your extension of my analogy, then, presumes a pretty hardline realism about the nature of physical laws that is not consensus among philosophers of science or even physicists (particularly empiricists of which there are still many).

I understood his comments based in the context of games and simulations, but not on the real universe.

>How does the server tell a ball on earth to fall

That's a good illustration of the difference between simulation and reality.

71.

Eric wrote> In the sense of the term "world", there is that which is defined and that which emerges through interaction within the defined.

Really? That doesn't capture the actual world. I'm missing something.

Eric wrote>Since we are talking about games, this is the difference between the rules and the scope. Regardless, it is still the game.

Yes, the sum of the two and maybe some other stuff constitute the game. Do we agree there?

Eric wrote>In the world, physical laws, as given in Raph's example, are part of the rules, and culture is part of the scope.

The problem, and it's the same problem I was raising with Raph's extension of an analogy I made, is that it's far from clear that the physical laws of our universe are or operate anything like the kind of 'rules' you're talking about with respect to games. To make his analogy stick, Raph's got to adopt a what I take to be a very radical position about the nature of physical laws, one that I suspect he doesn't really hold.

Eric>I think your argument is largely semantics based on the subjective use of the term "world".

Well, here I guess I plead guilty. I don't see any more fundamental question about virtual worlds than whether they are indeed 'worlds' in some meaningful way, as the phrase 'virtual worlds' implies they are. For example, if it turns out that the 'virtual world' just is the world simulation, then I don't see what what is so worldy about them at all. All the worldlike stuff happens outside them, and they become just like any regular old model. Except that these model's designers refer to them as worlds.

Fortunately, I don't think that's the case. And I suspect that Raph doesn't either.

72.

Eric wrote> In the sense of the term "world", there is that which is defined and that which emerges through interaction within the defined.

Monkeysan>Really? That doesn't capture the actual world. I'm missing something.

I did not say the definition was limited to those two dimensions, I was merely describing the difference between those two dimensions. A difference you made that I was re-asserting in other terms. I was not attempting a general description of the actual world. I was attempting to focus on context, as again, I'm doing here.

Monkeysan>Yes, the sum of the two and maybe some other stuff constitute the game. Do we agree there?

Yes. I get the sense you perceived I was defining a game as having only two dimensions, which I wasn't. I was merely describing differences of two dimensions of a game. I could have highlighted the differences of rules, scope, players, tactics, etc., but my point was the difference between rules and scope. Thus my effort to keep focused on the argument at hand.

Monkeysan>To make his analogy stick, Raph's got to adopt a what I take to be a very radical position about the nature of physical laws, one that I suspect he doesn't really hold.

An analogy can work well for a virtual world, but not hold well for the actual world. I perceived Raph's analogy as holding well for virtual worlds. Virtual worlds have containment issues, that is, aspects of the actual world are intermixed with the virtual world because of the interaction between the two. As a simple example, a player's consciousness can exist outside of the virtual world, but not outside the actual world. I could say that an avatar, as a corporeal proxy in the virtual world, is analogous to our bodies in the actual world. This analogy holds true in both. We experience the worlds through such corporeal entities. I could further say that in a virtual world, player conciousness is external to the world, but I say that knowing I am not making any such metaphysical claims about the actual world. I perceived Raph's analogy in the context of virtual worlds.

From a virtual world perspective, whether a ball falls based on the interactions of subatomic particles, or whether a server commanded it to do so, the player observes a ball falling. In the actual world, we know it to be the interaction of sub-atomic particles, but the virtual world can have many implementations which display such behavior. The player still observed a ball falling. The player observed his environment. The environment is shared by other players, who also observed the ball falling. When someone on this board writes of "the server", I perceive it as "the shared environment." Someone could just as easily call it "the world server" and also, simply, "the world." I don't see such a statement attempting a rigorous definition of the term, though.

Monkeysan>>if it turns out that the 'virtual world' just is the world simulation, then I don't see what what is so worldy about them at all.

The term "world" is a subjective term of containment, but I understand your point to be, as an analogy, that a game is not simply just the rules, and I would agree. In that understanding, I don't see that as a disagreement with what Raph was illustrating.

73.

When I spoke of the 'server' telling the ball to fall on earth, I was questioning Raph's suggestion that the laws of physics are the actual world's 'server'. I wasn't talking about relationships in a virtual world. Sorry if that was confusing.

The term "world" as a subjective term of containment may work fine in day to day 'loose' talk. But if one wants to talk about 'virtual worlds', compare them to the actual world, infer that they are genuine 'places' and that people 'live' there, then it behooves us to figure out in what sense those words are appropriate and in what sense they are just provocative metaphors.
That starts with trying to get a handle on fundamental terms.

74.

Monkeysan -

You made two arguments I addressed.

One, If Raph asserts the laws of physics are contained in the server, how can such a server analogy apply in the real world. More precisely, you said "it doesn't follow that there is anything like a server of physical laws in reality".

To this I explained the analogy seemed to concern a discussion of virtual worlds. The server analogy can only apply to the actual world in the form of "shared environment". The shared environment and thus, the universe, caused the ball to drop.

Two, If Raph asserts the laws of physics are the world, then what of culture? More precisely, you said "What part of the world simulation contains the relationships among users, the social structures, the ethical dimensions, the very users of a virtual world itself even"

To this I explained that he seemed to be illustrating one dimension, and not asserting that the laws of physics was -the only- dimension. To do this, I needed to restate the difference between the laws of physics and culture as that which is defined and that which emerges within the defined. This allows the easy transition to the analogy of the difference between rules and scope in game theory. Although Raph was focusing on rules, he did not seem to define "game" as only a set of rules.

Monkeysan>it behooves us to figure out in what sense those words are appropriate and in what sense they are just provocative metaphors.

I agree, and have been involved and have read many discussions on it, and it is quite a complicated subject with many views, and invariably, someone brings it up. Most academic papers I have read which use it tend to define it explicitly towards their use. I've even written a post on MUD-Dev defining my understanding of what virtual world means. Welcome to the party. Write your own fully scoped definition and submit it for feedback.

It still irks me to see people talk about 3D in contrast to text worlds, but I look past it in an effort to understand their point. Both visual worlds and text worlds can be in three dimensions, and the examples of visual collaboration in the original post are, in reality, 2D. From a cognitive science perspective, I enjoyed Castronova's post best concerning the differences between imagining and viewing, and I tend to agree with the criticisms of Cory's assertions by Bartle and Koster. I would agree that imagining and viewing are different, but not for any reasons Cory gave. I would even assert something as profound - visual collaboration is best implemented in visual environments. ;)

75.

Eric>One, If Raph asserts the laws of physics are contained in the server, how can such a server analogy apply in the real world. More precisely, you said "it doesn't follow that there is anything like a server of physical laws in reality".

I completely agree. See the earlier post where he described the laws of physics in the real world as the "server". I didn't see how that analogy could stick. That's why I asked the rhetorical question about the ball.

Eric>I explained that he seemed to be illustrating one dimension, and not asserting that the laws of physics was -the only- dimension.

Ok. Then he can't maintain that the laws of physics are the world. The use of 'are' is a statement of identity.

Eric>I needed to restate the difference between the laws of physics and culture as that which is defined and that which emerges within the defined.

Ok. But then I take issue with the claim that the laws of physics of the actual world are 'defined' in any real way. I certainly agree that they exist, but there's a wide gulf between the existence claim and a claim that they are some how 'defined'. Maybe I don't understand what you mean by 'defined'.

Also, just because culture cannot be explained in terms of 'the laws of physics' doesn't mean that culture is necessarily 'emergent'. The language of biology, for example, may prove sufficient. All of this is true even if we agree that the cultural supervenes on the physical. The point I'm trying to make here is that even though it would make an elegant analogy, it may not be tenable to say that the defined/emergent is the appropriate bridge concept between rules/scope and laws/culture.

Eric>It still irks me to see people talk about 3D in contrast to text worlds.

I agree to that extent that it misses the real distinction, imo: namely the distinction between descriptive (or propositional) and model-based environments. The 'text' could just as easily be photographic images and the model could just as easily be rendered in two dimensions, and the distinction would still be preserved. The question is what that distinction means, and one thing I don't think it means is that text environments like muds are inherently less valuable or should be disqualified from being considered true 'virtual worlds'.

Thanks for taking the time to hash a bit of this out. We may disagree, but I find it pleasing to encounter others who care about this stuff.

=]

76.

I'm not going to try to respond individually to the points made in the excellent discussion above. Rather, I'll try to clarify (or even abandon) the analogy.

In general, there is some formal system, a set of abstracted rules made solid via code, that serves as a pltform for the virtual world. Hopefully we can agree that this, while not being the entirety of a full world, is certainly a fundamental portion of it. In terms of the specific sort of virtual world we discuss here (leaving out the more 'academic' sorts of stretching of definitions, such as calling Flickr a virtual world, or something), this portion of the system is essential; the virtual world would not exist without it.

To me it is definitional, as I have mentioned, that said system simulate space--a place, as Richard puts it.

(It is also possibly definitional that it is considered to "exist" whether or not there happens to be anyone connected. Ironically, with modern technology, this may not truly be the case, in that no actual simulated space might be instantiated. And yet, all the data necessary to conjure it back up does still exist. Certainly the illusion that it continue to exist seems fundamental).

The systemic rules that form this simulation are designed ones. They may not have been carefully thoguht through, and their designers may no know everything about how they work. They may exhibit emergent properties, and they will certainly be used in ways that the designers did not foresee. Without this formal simulation framework, there are none of the intangibles that also form part of the world.

Certainly users almost certainly do not know the true shape of this system. Their experience of the system is heavily mediated. Many complexities are elided, in the name of user interface, in the name of immersion, in the name of narrative, in the name of entertainment, socialization, and so on.

In addition, it is likely that the designers themselves do not know the true shape of the system. It has almost certainly been mediated for them as well, by programmers. And programmers have experienced it via code, but there is data that they likely do not see in a raw form, and so on.

So here, truly, is the elephant. It is not fully known, and yet all the formal tangible elements are nonetheless human-created and designed. Just as the designer of a jet probably doesn't know everything about, say, the precise thread count of the blankets, neither does the developer know everything about what they made. This is pretty common to all forms of human creation.

Nonetheless, this construct exists and is quantifiable ina very concrete way. What's more, the methods of interaction with it are in fact codified. They are bounded by the construct's very architecture. Developers often refer to these bounds as "the laws of physics" of the world, even though they often have little to do with physics or physics simulation. They are constraints imposed by fiat, implicit in the fabric of the simulated universe, unbreakable by actions within the space itself.

Hence the analogy. We don't get to break the laws of physics. There is no claim made that we understand the laws of physics. There is no claim that our notions of the limitations of physics are accurate. There isn't even a claim that there ARE limitations in actual physics--that's a point at which the analogy breaks down, because we DO know that there are limintations and boundaries on the simulated systems that form a virtual world.

My claim (such as it was) was that the fundamental nature of this system is not changing because of the intermediated perception of said system. Speaking a bit more formally, "the client" is the mediating channel for "the server" which is the construct. Said mediating channel may well impose additional constraints on interactions with the construct, beyond the ones that are implicit in the construct itself. (The world might go to 11, but the client might only allow going to 8). This does not mean that the data, the limitations, or the rules of interactions extan in the construct are in any way altered by the mediating channel.

No judgement is made regarding the suitability of a given mediating channel for a given simulation. Some will be better suited to certain sims than others. Different intended mediating channels may well become a design constraint on the construct, and this might even be mandatory in order to achieve a given effect, but it's not mandatory in the general case.

When examined from an architectural point of view, this set of simulatory rules may or may not reside on a remote server; it might be a consensus set of data and interactions distributed across a P2P network. Regardless of that, it's still going to be the same rough diagram:

construct <--> mediating channel <--> user

One might even argue that the more reliant on a single mediating channel the construct is, the POORER it is, in areas such as portability, longevity in the face of the march of technology, flexibility of interaction platform, and so on.

And that is why I say there is no fundamental difference in the intrinsic nature of text worlds versus 3d worlds. All of the above is fundamental to the basic model of what wre are discussing.

Can different mediating channels have strengths and weaknesses? Sure. And different levels of flexibility? Absolutely (I would imagine a mediating channel that was solely audio would be incredibly painful to use). Is an elephant still an elephant regardless of what part of it the blind men touch? To my mind, yes. And it doesn't even preclude the developers themselves from being if not blind, at least myopic regarding details of things like the dilation rate of an elephant pupil or the precise thickness of an elephant hair.

77.

Not to boil down useful explanation into too small a bite, but:

Raph> there is no fundamental difference in the intrinsic nature of text worlds versus 3d worlds.

1. Doesn't function follow form? In a practical sense, doesn't knowing that your world's interface will be primarily text (or 3D graphics, or sound) dictate to some meaningful degree the intrinsic nature of that world?

2. For form to have no effect on function, wouldn't it be necessary to extract all the "laws of physics" out of the construct and place them in the mediating channel? Is that even possible?

I don't necessarily disagree with the notion that text and non-text worlds can be intrinsically identical -- I'm just curious to see how far that concept can be pushed before it breaks.

--Flatfingers

78.

(Yes, I know it's usually "form follows function," but experience as an architect of systems suggests that the reverse is often true as well.)

79.
1. Doesn't function follow form? In a practical sense, doesn't knowing that your world's interface will be primarily text (or 3D graphics, or sound) dictate to some meaningful degree the intrinsic nature of that world?

I pretty much said the same thing earlier, in a point Raph already responded too. He contends that we can minimize or maximize that by design. I basically agree that with that. I just don't agree that it actually happens often in practice.

80.

Raph>And that is why I say there is no fundamental difference in the intrinsic nature of text worlds versus 3d worlds. All of the above is fundamental to the basic model of what wre are discussing.

I agree with you here, and I agree with your model of mediation between construct/interface/user. My only gripe was when I thought you were circling 'construct' and giving it a gold star for being identical to the virtual world.

My model of the situation

world model <-> avatar <-> user

is completely analogous. I just think the whole complex is what constitutes a 'virtual world'. But that's another story. (Please don't make too much of my use of 'avatar'. I have a reason for using that word, but it's not important here. Suffice it to say that in this context I mean it to be nearly synonymous with client/channel/mediator.)

I also agree that possessing at least the functional equivalent of a world simulation is a necessary feature of a virtual world.

I still see a fundamental distinction between descriptive and model-based mediators/avatars. But while that distinction is fundamental to how the user perceives a virtual world (which in turn influences the contours of that world itself), the distinction between descriptive and model-based 'avatars' is not fundamental to whether it is a virtual world.

I'm sorry if that muddled things.

81.

I do believe we're converging!

On the function and form issue--consider audio books. Unabridged ones, specifically. Different functions AND forms. Yet the same words. And indeed, many folks speak of "reading audio books."

Who's right? Well, it depends, as monkeysan mentions, on what level of detail you are looking at, I guess. :)

From a "define the medium" point of view, I do tend to circle and gold star the construct, because it's the irreducible portion; you can remove or change the others and still have a virtual world. Remove or change the construct, however, and you start getting other beasties, like websites, chat channels, or standalone games.

That doesn't mean that I would ever minimize the importance of the whole, though.

(FWIW, I also believe avatars are important mediators of both perception and action, but they exist between the world construct and the mediating channel).

82.

Hooray!

I do have one question though. Well I have more, but for now I'm just curious: You said something like, " you can remove or change the others and still have a virtual world."

I see how you can change the mediator and still have a virtual world, but how could you get rid of it? I'm not sure you have a virtual world anymore if all you had was a world simulation and no way for anyone to access it or even be able to observe it's behavior. Unlike some of your licensed friends, I don't speak binary. ;)

83.

It happens all the time; you run the server and do not have any clients connected. It's still demonstrably there, as you can easily verify by checking the process on the server. It's even taking independent action--NPCs are moving about, time is going by, etc.

If you tell me that's not a virtual world, I'll have to accuse you of discriminatory privileging of human participants. ;)

84.

Raph> [if] you run the server and do not have any clients connected [it's] still demonstrably there

It may still be there, but the folks over in the Developments in the MOG Market thread will call it a "ghost MOG"!

--Flatfingers

85.

Well, let me respond with this: What do you mean when you say "you can easily verify by checking the process on the server"?

The reason I press this that by "checking the process" I assume you mean that either (1) you're using tools that monitor the world simulation and can report on its states or (2) you just look and see that the program/suite that constitutes the simulation is being executed by the hardware.

In the first case, I'd argue that whatever tools you're using that let you interact with the world simulation could qualify as a mediator. So that doesn't bother me so much.

The second case is a little hazier. I suppose one might argue that even observing that the program is running is a form of mediation, but I don't find that compelling. If all you mean by "checking the process" is confirming that the software is being executed, then I'd disagree that there is a virtual world of that world model in existence at that point in time.

The NPCs, for example, are only NPCs if you have a way to map the processes that the code is directing to your metaphor 'world model of Norrath'. Without that in place, all you have is a digital computer changing states over time. The computer doesn't understand the metaphors that make the 'world model of Norrath' into 'the virtual world Norrath'. That feat is accomplished by operation of the mediator and the user. And without that operation, the metaphor fades and with it the world, leaving only the model.

There, I said it! U-S-E-R!

In some sense, I guess a stand guilty as charged of "discriminatory privileging of human participants." I'd prefer to state my crime as one of "discriminatory privileging of conscious observers." I'm not 'privileging' humans in particular.

The point is that it makes perfect sense to me to say that a world simulation cannot be a virtual world without an observer. And there can be no observation of a virtual world without an avatar (if only in a very very abstract sense, such as client/mediator, which I realize conflicts with the common usage of that term). I'm even unashamed to say I don't think there's any such thing as an 'empty virtual world'.

You can run the world simulation and even have a client running in an empty room, but until you have a conscious observer interact with the world model, there is no virtual world. Put yet another way:

The presence of a conscious observer is a necessary condition for the emergence of a virtual world from a world model.

It's kind of like Schroedinger's cat (but don't press that too far). Maybe I'll call it Koster's cat. That way, even if you don't buy it, I can still imply you endorse it. ;)

86.

Heh. I knew you were going to go there as soon as I posted that.

To my mind, that sort of ontological question is fun to play with in an academic sense, but by and large not all that practical. Yes, it's true that even typing ps at the command prompt is a form of mediation. But who really cares?

A huge portion of the trades on the sock market happen in automated fashion now. So no conscious observer. Is the market "not there" when this happens? The same thing applies to passive software display mechanisms that put up stuff like character stats on a webpage, like EQ2Players.com. These would operate whether there were a user or not. Is the world "there"? Who cares? :)

The interesting thing about Schrodinger's cat is that it exposes interesting aspects of reality to us. Is there a practical application of saying that the presence of a conscious observer is required?

87.

Raph>The interesting thing about Schrodinger's cat is that it exposes interesting aspects of reality to us. Is there a practical application of saying that the presence of a conscious observer is required?

Right. Then you're real objection is not that what I'm saying is wrong, but that even if it's true, it's not very useful. And if a theory isn't useful, who cares?

Ok. Well, as always, useful depends on what you want your theory to do.

I want a theory of virtual worlds so that I can think more lucidly about the questions virtual worlds pose for society, the future of communication, and ourselves. You know, the little things in life. ;P

I don't think a theory of virtual worlds that takes a virtual world to be identical with its world model can be an effective explanatory framework for those kinds of questions.

So, first, I turn the question back on yourself: "What practical application (other than mirroring the architecture of the technology) is there in defining the virtual world as being the same as the world model?"

But to answer your question, what is the practical application of my requirement about observers?

The short answer is about meaning and models. Consider a map. A map is a model. It is a model because it functions by instantiating the very properties it seeks to represent. This is unlike getting directions, where you are simply supplied with instructions that represent propositions about your relationship to your destination.

A map can only function as a model if there is an observer who can recognize that certain properties instantiated by the map represent properties of some region. Otherwise it's not much more than a piece of paper that is very hard to fold properly.

The observer supplies meaning to the map, allowing it to function as a model.

The same is true of a world model of Norrath. When the software that runs 'the world model' was built, it was done by programming rules for making state changes in the hardware that runs it. The computer has no concept of NPCs, for example. They are only NPCs because the dev team (a group of at least semi-conscious observers by the launch), can map those state-change rules to properties of your imagined Norrath. So, even at the level of simulation, what makes the instantiation of your software a world model is that you've created a mapping by applying meaning to the state changes of a machine.

Ok, but what about the level of the virtual world?

In the same way that meaning has to be supplied to make a world model of your software, the virtual world has to be constructed by having observers supply meaning to the machinations of the world model. By interpreting the output of the their client as a model that maps to properties the imagined world of Norrath, that is, by coordinating the behavior of that model with the behavior of some imagined Norrath. At that point, I think it can be said a virtual world exists.

Think of it like this, I'm saying that neither the world model itself nor one's imagined idea of Norrath are actually the 'place' Norrath. The place 'Norrath' emerges from the coordination of the two. Norrath is born. It is only at that point that the really interesting features of the system starts to emerge. And when thousands of us are doing this coordination roughly simultaneously from data generated by the same world model, amazing things start to happen.

Finally, about cats, practicality and explanations:

Schroedinger's cat exposes an interesting aspect of reality: It exposes that observers are required in order to make determinate certain states of affairs about our world.

Koster's cat exposes an interesting aspect of virtuality: It exposes that observers are required to make determinate the existence of a virtual world.

The practical advantage of Koster's cat and defining virtual worlds the way I do is that it can lead to a theory of virtual worlds that can explain interesting phenomena while still comporting to our common sense ways of talking about them. For example, when we ask about what goes on 'inside Norrath', we are rarely asking for a description of what goes on inside the 'world simulation of Norrath'.

I know this is too long an explanation for a post and far too short to make the theory clear, but it's worth a try here and there.

(FWIW: With respect to the stock market (or even sock markets, though they are less volatile and widely trades), I don't claim the stock market is a virtual world, so the requirement doesn't hold. However, to the extent that one wants to claim it is a virtual world, I'd say all the automated trades show is that stock market avatars do not need to act in real-time concert with the stock market world model.

=]

88.

Mike Sellers wrote:

Perhaps thousands -- maybe tens of thousands if we're being generous -- play text-based games.

My estimate, incidentally, is that somewhere between 200k and 400k people play text MUDs. Paltry compared to the size of a single game like Runescape, of course.

--matt

89.

It never ceases to amaze me how little most people know about text gaming.

> Mike Sellers wrote:
>
> That is, millions of people play graphical games.
> Perhaps thousands -- maybe tens of thousands if
> we're being generous -- play text-based games.
> Arguments from personal experience any of us may
> have had faovring either pale in the light of
> this reality.

Mike, I know of single games that have tens of thousands of users. I have personally run text games that cracked the ten thousand user mark.

Hundreds of thousands of people currently play text MUDs- perhaps millions.

90.

Monkeysan>The presence of a conscious observer is a necessary condition for the emergence of a virtual world from a world model.

One of the things Roy Trubshaw said to me when we were working on MUD1 was: "You know that question of whether, if a tree falls in the desert and nobody hears it, does it make a sound? Well in MUD, I can tell you that it definitely does. I've just written the code for it."

Put another way, the world is there because the code is there. It runs, therefore it exists. It doesn't matter whether there's anyone observing it or not - it was "preobserved" by its creation.

Of course, you can argue that the hardware might have crashed and that you can only be sure it hasn't by "observing" it. However, that's reducable to the situation where you can only be sure of anything in Reality by observing it, and even then that's only true if you trust your senses.

From the point of view of an NPC in a virtual world, it continues to exist whether there are any visitors from Reality or not.

Richard

91.

Just to get rid of that unclosed italics bracket...

92.

Richard>From the point of view of an NPC in a virtual world, it continues to exist whether there are any visitors from Reality or not.

Well, I really hope you don't mean that literally, since NPCs don't have a point of view (at least not yet). If they did, the term NPC would likely no longer be accurate. We'd need something more like NHC (nonhuman character).

The issue of observers isn't simply that without a 'visitor from reality' there's no way of knowing. Rather the issue is that without the observer, there is no virtual world at that point in time. For an explanation why this is the case, see my myriad posts earlier (if you can slug through 'em ;)

In the interest of brevity I'll just provide an intuition pump: If the world model just is the virtual world, then what on earth do we mean we ask questions like, "I wonder what's going on in Norrath right now." Clearly, gamers are not simply asking for a description of the state of the world model. They are asking about other things that lie completely outside the world simulation. And while you may want to dismiss this as 'just talk', then you also have to dismiss as 'just talk' many of the very things people at TN find so meaningful about vws. Things like economies, guilds, shared experiences among players, collaboration toward goals. None of these are subsumed in the world model.

Many gamers they take the notion that they are interacting in a 'place' very seriously, sometimes even to the point where they claim to 'live' there. These people aren't simply deluded either. There is a very real sense of place that is necessary to being a virtual world. The world model on it's own doesn't provide that, therefore it can't be the 'entirety' of what we call virtual worlds.

Of course, what I said above doesn't mean that there is no world model without an observer, because, as you've said, the mapping of meaning from imagined world to code was established at the time the software was engineered. As long as someone is around who can recognize that there is a mapping between the instantiation of the code and some imagined world, we have a functional model.

93.

apparently that wasn't the end of the italics bracket you closed earlier =]

i'll try again

94.

The presence of a conscious observer is a necessary condition for the emergence of a virtual world from a world model.

I would argue that the notion that worlds are defined by media is not very useful for some perspectives. An empty demo of a house sitting on a new sub-division is physically identical (within the limits of construction methods) to other houses in the area, may be tastefully decorated to create an illusion of being lived in, but lacks the standing patterns of behavior that make it "inhabited."

I'm reminded of one of my favorite themes from science fiction and horror cinema, the notion of finding yourself in the city where you have lived, and finding it alien because you are the only person inhabiting it. The opening minutes of 28 Days Later are much more scary for the feeling of finding the familiar transformed into alien than the rest of the film. (I wish I could remember the Australian science fiction film in which a man wakes up from an accident to find that he's apparently the last man alive.)

One of the key things that Rene Magritte's painting explores is the difference between object and representation. Does the statement "This is not Rene Margrite's pipe" apply any less to a prop you can pretend to smoke, than a painting or photo? Not really. The real Rene Magritte's pipe was defined by how it was used, by whom, and where.

So I would argue that presence is necessary, but not sufficient. We can observe swingsets, sandboxes and monkeybars until the cows come home. It doesn't become a playground until someone takes the next step and actually plays in it. Behavior settings and lets extend it to behavior objects are critical to my view of what is needed to talk about "virtual worldliness."

95.

So to that end, here are two different definitions of virtual worlds that might be useful in other contexts.

1) A set of interconnected and interrelated behavior settings, shared between multiple people over time and built around a metaphor of physical place.

2) An online ecology in which agents and objects interact in structured ways that are consistent over time.

96.

Kirk wrote: The real Rene Magritte's pipe was defined by how it was used, by whom, and where.

So I would argue that presence is necessary, but not sufficient. We can observe swingsets, sandboxes and monkeybars until the cows come home. It doesn't become a playground until someone takes the next step and actually plays in it.

Agreed

97.

Kirk>I would argue that the notion that worlds are defined by media is not very useful for some perspectives.

So you agree or disagree with my claim? I take it from your post you agree, but I'm not entirely sure what you mean.

98.

By the way, Kirk, after spending the end of last week at the GLS conference listening to Henry Jenkins, James Paul Gee, Kurt Squire, and others, I'm again curious about your certainty that learning is technology independent. It certainly seemed that some pretty bright people were taking positions that differed from your posts. Also, by your lack of answer to the appendectomy question, I presume that your choice would be to learn from a book (or that all of the options would equally prepare you for surgery if you only had access to one)? Anyway, since I'll be at a conference in Indiana later this fall, hopefully I'll get a chance to see your research and hear your thoughts over a beer.

99.

Separately, had a chance to talk about this topic at both Supernova and GLS with lots of people -- some of who have posted on this thread -- and a parallel question or two emerged that probably act as predictors to how you would answer the original post:

1) Do you think that the experience of using graphical worlds compared to text worlds is greater or less than the same comparison between movies and novels? Should either of the differences be classified as fundamental?

2) On the continuum of complexity (or, perhaps, freedom of expression), is it a fair statement to say that movies offer more complexity than text and that games offer more than movies? (Nota bene: "games" in this context includes text and graphical games)

100.

I about 80% agree but feel that we need to take it just a bit further beyond just the "observer." If we stick with the quantum mechanics metaphor, it's impossible to "observe" without "interacting with" in ways that perminantly destroy the original state of the system.

The comments to this entry are closed.