The Theory Of
Here you'll find discussion of all manner of topics relating to the theory of multiplayer games. As I see it, anyway. A note to commentors: if you stray off-topic or if your reply contains ad hominem attacks, your comment will be deleted.

Synopsis: A system for addressing internet anonymity is discussed.  The system relies on the act of assigning an internet-wide identity to people.  Like a character in a game, how people develop their internet identity's reputation will impact their ability to get along with others online.  Features of the system beyond the scope of games is also discussed.

Internet anonymity gets a bad reputation because of the antics of people online when they can be anonymous.  Lots of people say "If people knew who these antic-prone people were, they wouldn't behave that way."  I happen to be one of the people saying that.  I believe that if everyone knew who I was, I'd spend a little more time thinking about the consequences of my actions.  My public identity would be hanging in the balance.  Just as I have incentives to stay on good terms with my neighbors, I'd have incentives to stay on good terms with people online.

The nice thing about internet anonymity is that it goes both ways.  I can contribute on the web and in games without some random person getting my identifying information and doing something like calling me at home, or sending threats by paper mail or some such lunacy.  The overwhelming majority of people are perfectly reasonable and are uninterested in performing maliciousness, but we worry over that tiny fraction of immature or unbalanced people who would choose to seek us out.  When I was a kid, we never locked our doors.  Until a thief started entering people's houses and stealing silver.  It only took one thief to get everyone to start mistrusting strangers.

So what could be done to instill some respect in people while still protecting everyone from the online predators?  The solution that I suggest is the same one that is used on countless web sites, only I take it to another level.

Instead of having an identity per web site or game, create a single identity repository on the internet.  Much as web sites register their information with one central website repository, each user on the internet registers with one central user repository.  Unlike the website repository, not everyone can wander around in the central repository's information.  The secrecy of identity information is critical.  However, once you are registered there, you can go to any web site, game or other online experience and use that central identity to establish who you are.

I would be user 3920578.  You would be user 4495874.  That's the only information that we have to volunteer about ourselves.  Everything else is private.  So when playing World of Warcraft, you are known to be a certain user.  If somebody wanted to, they could click on your avatar and ask Google to list stuff that you've written on MMORPG.COM.  That is part of your internet identity - what you say in public.

Ideally, Google will record the identity of the person who made the query.  You should have the opportunity to find out who asked things about you (including government agencies).  So if user 1987458 does a search on your information on MMORPG.COM, you would know.  You could then do a search on their information, but that would inform them that you did that search.  And so on.

So what's in here to stop some griefer from getting your home phone number and calling you in the middle of the night?  There is no public connection between your user number and your phone number.  If you use your real name in a post on one site and you include your home town somewhere else on the internet, somebody could put them together to find you.  So you don't want to publicly post personal information.

What's then needed is a way for you to grant permission for people to get personal information about you.  Sometimes you want to share that information.  But you don't want to share it with everyone.  Just a specific person.  If you want somebody to email you directly, you could send them your email address.  But how could you do that when you don't know their email address?  You tell the central repository to let that other person see your email address.  Now they can see it and send you an email.

Let's get back to gaming.  You see me in a game today and mark me as "Gongo", because that's the name my character is using.  We group, we chat, we finish some quests together, and we graduallly part company.  Years later, we bump into each other in a new game.  You are told that this guy flying the spaceship next to you was "Gongo" back in that fantasy game all those years ago.  We can renew our gaming friendship.  We aren't obligated to swap names, emails or phone numbers.  But we certainly could if we wanted to.

What about the idea of curbing annoying behavior?  Well, consider that a jerk will be pegged as a jerk as soon as he behaves like a jerk.  Instead of marking me as "Gongo", I might get marked as "Jerk".  From then on, anytime the player who marked me as "Jerk" sees me - anywhere on the internet - they will know that they didn't care for my behavior sometime in the past.

That definitely puts a curb on the enthusiasm behind being a jerk.  It's no longer enough to do something annoying in a game with one character, have some mischievous fun at others' expense, then simply create a new character and move on as if it had never happened.  With the internet identity in place, your antics follow you wherever you go.

Now toss in a social networking structure on that central identity repository.  Declare who you like and who you don't like.  You meet somebody new on the internet.  You don't know them, but you know that a friend of yours likes them.  So you decide to trust them.  Or vice-versa.  You don't know them, but you know that a friend of yours dislikes them.  So you decide to be wary.  Or yet again, you don't know them, but you know that an enemy of yours likes them.  You may decide to be wary again.  Or you may just decide to see for yourself.

Roleplayers might think that this is a problem for them.  Sometimes they want to roleplay a jerk.  That's still okay, assuming that everyone understands that it's all roleplaying.  When someone decides to roleplay a jerk and they do it to the wrong person, they'll get tagged by that person as truly being a jerk.  This serves to remind us that we're not just avatars in games.  We remain real people, impacting other real people.  If we're going to roleplay a jerk, we make sure that everyone understands that we're just roleplaying, and we'll take further steps to ensure that others are accepting of that.

Back to general cases.  Suppose you wanted to sign up for a new game.  Normally, you go through that annoying task of giving them your credit card information.  But if the central identity repository knows about your financial information (just as you have it spread all over the web today), then you can give the game publisher authorization to see a particular credit card and to use that for your subscription.  And that's all they get to see.  Properly done, the company simply needs to make sure that you can pay the subscription through the central repository and then bill through the central repository.

If a game wants to ban a player, they can look at their list of banned identities and continue to refuse admittance to those people.  No more of this nonsense of griefing under one identity, being banned, and then creating a new identity and getting right back into the game to grief some more.  No more banning of IP addresses, credit cards or other partial forms of identity matching.

What's the downside to all this?  Clearly, we'd lose our anonymity.  Companies and people would be able to spot us once they had seen us before.  In games, it would be worst when you're trying to just play the game when you stumble across somebody that you know in real life and they want to talk to you about non-game stuff.  "Just leave me alone, I'm trying to play a game here."  But you don't want to be rude, so you're stuck talking to them.  It might even be something related to work.  Ick.

Clearly, there would have to be means to broadcast to others that you really want to be left alone.  Games would make it non-trivial to get to the identity information of a player.  So instead of having all your identity information floating over your head, a game might only show your identity number if explicitly asked, and then the interested player would have to go to a web site to look up that number.  Only people who were motivated would take the time to find out.  Just getting to the identity information in the first place might tell the target that their identity number is being asked for.

How about all this "Is that female character really female?" nonsense?  Never fear.  Your gender would be part of your anonymity.  If you're concerned about being a girl playing a game, you can remain anonymous to a degree.  Somebody can still look you up on the web according to your identity.  Frequent postings on women.com might be a giveaway, of course.  But if somebody decides to harass you over your gender, you just set the internet police on them.  You can identify them perfectly because their identity is known to you.  Not their name, age, address or gender, but you can say "That one, officer."

Those police aren't just internet-based, either.  Now that you have been fingered for harassment, the real police can use the identity information to go to your house and charge you with a crime.  For the paranoid, the police can't see any personal information without proper authorization.  And when they do look at it, the police officer who is looking at it gets their identity recorded.  The idea is to avoid any secrets from the central repository.

For the truly paranoid, yes, the central repository would be constantly assaulted by every criminal element on the planet.  Just as they are constantly attacking all the various and sundry company- and web-specific repositories out there.  Instead of trying to defend myriad repositories, a single repository could be enforced like Fort Knox, growing ever more protective of that information.

For gamers, the goal of the system is to make sure that it's possible to recognize a certain person.  If I am significantly nice or mean to you, you can push a button to record that.  If you see me another time, you can be told of your opinion of me.  That can be a valuable technique for managing online interactions in avatar-based games.  Or in web forum discussions.  More broadly, the system can be used for almost everything under the sun.  A cynic sees George Orwell's 1984, while a more optimistic person would see the opportunity to take something away from the less-than-altruistic members of our society.  Both online and off.

t0nyd just observed how odd it was to publicly suggest that a certain mechanic be part of an MMORPG, only to have people respond that he doesn't want an MMORPG, he wants an FPS.  I commented that a part of the problem is the MMORPG acronym itself.  I wanted to expand on that idea here because it's important.

Ask 1000 people on this site what an MMORPG is, and you'll get 1000 different answers.  The acronym MMORPG is pretty much useless because all it does is refer to a broad space of software application that people use online for entertainment.  MMO is even more broad, and even more useless, even though I use that acronym myself quite a bit.

Massively

Multiplayer

Online

Roleplaying

Game

If you look at each word in the acronym, you and your friends could spend weeks arguing over what each one means.

Is "massively" pertinent to simultaneous connections to the server, or just the number of people who can connect to it.  Must they be able to interact with each other, or just connect at the same time?  Is 100 "massive" or must it be 500 or more?  A thousand or more?

Just the word "game" can cause conniption fits among the intelligentsia of the gaming world.  Does a game require competition?  Must there be defined rules, such that a sandbox implementation (whatever that means) cannot be termed a game?  These debates just go on and on and on.

I tend to try to dodge the labeling wars by going with generic words and phrases such as "entertainment software".  It's rather more boring terminology than using acronyms such as MMORPG, but it permits people to think about a bigger picture than that afforded by the more restrictive acronym.

Is it important to say that an MMORPG is a program that lets 1000 people or more simultaneously connect to a server where they can move a humanoid avatar around that can own virtual objects, advance in levels, attack computer-controlled avatars and chat with other players who are connected to the same server?  I say that it's not important the least bit.  Because if I then want to create a game that is a variation on that theme and that players would love to play, do I want to run into a stone wall simply because people won't slap the MMORPG label onto it?

Consider Will Wright's upcoming game Spore.  Is it an MMORPG?  I don't know.  I suspect most people won't label it that way.  Does it matter?  Not a whit.  Is it fun to play?  That's the important thing.  Will it be fun because it can or cannot be labelled MMORPG?  No.  It will be fun to play because of the structure that Will Wright and his talented developers and artists put into place.

Arguing over whether or not something is an MMORPG or FPS is ludicrous.  Argue over whether or not a given construct or combination of features is fun.  If you don't think it would be fun, just say as much and move on.  Don't claim ownership of a meaningless acronym as a means of attempting to dismiss somebody else's idea.  If you have the design talents, tackle the challenge of figuring out what segment of the population out there would be interested in playing a game with that construct.  Then provide feedback.  Going a step beyond, work with the original idea to come up with something that works perhaps even better than the original idea.

Synopsis: Advances in graphics are producing photorealistic environments, outpacing other technologies that make up games (physics, artificial intelligence, animations, etc).  This produces a jarring disconnect between what we see and how everything moves and behaves.  The game industry has a decision to make about how to deal with that disconnect.

I just watched a bunch of YouTube videos on DirectX 10.  There's no question about it, the graphics have become pretty close to photorealistic.  I saw one clip showing graphics from Age of Conan, Crysis, Bioshock and others.  Lighting, textures, details, it's all there; just gorgeous graphics.

I've often commented about what a waste it is for games to crank out such beautiful graphics when the gameplay itself remains unchanged.  Crysis is a first person shooter, just like any other.  It's a remarkable technical achievement, but it's just a shooter.  Put the pipper on the enemy, click the mouse - boom, he's dead.  There's nothing wrong with it being a shooter.  I'm simply taking exception to the fact that new games have primarily been exploring avenues of better graphics.

Certainly we see that in MMOs.  The most significant achievement in MMOs was World of Warcraft and its ability to polish the package to a fine luster.  It's the same gameplay that has come before, but it's put together very well.

While I was flipping through YouTube videos, I happened upon a bunch of videos demonstrating Crysis AI oddities.  My favorite was the video of two enemy soldiers talking.  The player got the drop on them from a sniper position.  He calmly shot one of the pair, which promptly ragdolled to the ground, dead as a stone.  The reaction of the surviving soldier?  He just as calmly continued talking to his dead compatriot.

Other videos illustrated other rather dumb traits of the game AI.  Certainly MMO AI is simplistic.  I'm sure that players have been picking up on game AI goofs and blunders for a while now.  Certainly Far Cry exhibited some of the same oddities that Crysis does.  Older games had even simpler opponents.

The funny thing about improving the realism of the graphics is that I think a greater percentage of players are going to notice the non-realism of the rest of the gameplay.  Developers understood almost immediately that the physics of a game needed to be improved if they were going to take the graphics to a photorealistic level.  It's just too jarring a departure from what the player is seeing if it doesn't behave the way that it is displayed.

Physics has become something of a hot topic in gaming.  I would claim that's because of the requirement that the simplest behavior of objects must match their appearance.  A box knocked over should tumble and bounce, not just slide.  If it only slid, then the player would be reminded that there isn't actually a box there, but just some triangles and some code.  It breaks the suspension of disbelief that the player is maintaining when seeing those beautifully-rendered scenes.

Organic movement is also important.  A photorealistic character standing stock still may provide great material for screenshots, but as soon as the character starts moving, it had better be natural.  If it's not, the delicate fabric of realism created by the graphics is torn apart.

The highest-order sense of realism that improved rendering calls for is the decision-making of the computer-controlled characters.  The artificial intelligence.  If a photorealistic monster both moves in a natural and realistic way and interacts with its environment in a physically realistic way, it would become rather disappointing to have it visibly tense its muscles, roar at the player in the tight confines of an underground corridor (photorealistically textured, of course), raise its double-bited axe with the nearby torchlight glinting off its razor-sharp edges - only to have it smack into the ceiling.  Courtesy of the physics engine, the ceiling would chip from the blow, and the chips would bounce off the shoulders of the monster.  Courtesy of the graphics engine, it would be gorgeous to watch.  But because the AI says that the axe must be raised behind the head of the monster, the axe will keep bashing into the ceiling of the corridor and the monster will never get around to swinging its axe.

The AI didn't match the realism of the display or the physics.  It's tough to get all these elements realistic, especiallly when organic beings are involved.  With that character stuck there, we're reminded that it's a computer game instead of being immersed in the notion that we're facing mortal danger in the form of an axe-wielding monster.

So I may cast doubts on the value of advancing the quality of the graphics in games while leaving the rest of the games in the stone ages of technology, but I figure that the advance of the graphics is going to be a flashpoint for players to insist on higher fidelity physics, animation, artificial inteligence and every other technology involved in the gaming experience.  Graphics have outstripped the rest of the field, and developers are going to have to hustle to move the rest of the technologies forward.

The alternative to this rosy technological scenario is that developers may do what the authors of Eve Online have done, which I thought was immensely clever; they went into space.

Games that take place in space require spaceships, not people.  There are no swinging arms, running legs, flowing cloaks and hair.  It's just cold, angular plates and mechanisms that swivel and hinge on a couple axes.  The spaceships don't need to path around trees, walls and pits.  They just need to travel in straight lines with the occasional concern of striking a planet or another ship.  The realism found in Eve Online is accomplished rather easily because of the dearth of organic complications.

So the industry has at least two ways to go.

1. Continue to push the technologies in order to fill out the realism of an organic environment.

2. Produce an environment that is better suited to whatever technologies are available.

A third way worth mentioning is:

3. Reduce the quality of the graphics such that realism is not the goal, permitting the player to think of the game as a toy instead of an alternate reality.

The second and third choices are not unlike each other.  Number two states that there are environments that can look great but not place significant demands on other technologies (e.g. space MMOs).  Number three states that there are environments that don't place significant demands on any technologies.  The latter would be for games that focus on deriving their entertainment more from the gameplay mechanics than from the realism of the gameplay environment.

The funny thing that happened on the way to photorealism is that there is now a demand for realism in all aspects of these virtual environments.  It will be interesting to see how the industry reacts to that demand.  Will they attempt to reduce the demand by entertaining us with environments that don't attempt realism or will they attempt to continue the trend towards realism?  The former requires clever gameplay constructs to entertain us.  The latter requires the sort of engineering that gave us DirectX 10.