May 8th, 1998

So Let's Get Practical...

"Killers may have a right to play the game their own way. They don't have a right to inflict their way of playing to the other groups." - from a posting by Loic Talecaster late last night
"Look dude,, dieing and parting with your stuff sux,, but it's part of the game, live with it.. role play it." - a reply to Loic by Dreadnaught
"If you don't want to get killed, stay in town." - a lot of posts this last week
Lynx hmms... So you're saying you want to randomly kill people?
Arion says "yes! randomly but with the same chances of dying!"
Ack says "we want to be the few... the proud... the outlaws!"
Lynx should note you are probably despised by a great many people, which is hardly helpful.
Ack says "that is the best part about it!!!"
Arion says "that's the way we like it!"
Lynx points out some people have ideas of fun that other people do not agree with, and these people do not have to put up with it.
Arion says "then let them stay in the havens that are usually there... [safe] rooms, the town square..."
- from "The Black Rose Incident", which occurred on Islandia 8 years ago

As one would expect, the last two essays have generated an amazing amount of discussion and controversy in the Ultima Online community. There have been numerous interpretations of what I said, of course; some feel I am acting as apologist for the playerkillers, others feel that their positions have been vindicated, etc. Just to state it clearly: there are too many serial killers in the world of Ultima Online and they need to learn to get along with the rest of the populace--but we don't want to exterminate them completely anymore than we want to make rattlesnakes, black widow spiders, and sharks extinct, because they fill a valuable role in the virtual ecology.

So let's turn to the problems of actually shaping a virtual society. There are a couple of key issues here that present really important problems.

The more things change, the more they stay the same...! Those quotes above, some freshly minted and the other hoary with age, help demonstrate that the issue of playerkilling, of policing the virtual world, is an old one, and that it is more a matter of the psychological approach the player takes to the game than anything else. One thing that is hard to come to grips with in that realization is that it means that playerkilling does not require a combat system. The problems that players wrestle with in dealing with harassment are exactly the same; the difference is the means of exercising power that is used by the aggressor. Consider the following quote from a post made late last night on the UOVault by Loic Telecaster:

[Yesterday's essay] says that without killers, virtual worlds stagnate. But there is an entire class of virtual worlds, MOOs mostly, that don't even implement a combat system, and they thrive. Thus how can it be said that Killers are "necessary"?

Those of you who read the additional links supplied yesterday know that the Julian Dibbell article on "A Rape in Cyberspace" in fact took place on a MOO. The "kill" command was never typed; nobody entered combat mode. Yet it was every bit as traumatic as a repeat playerkilling is to people today in UO. When the issue boils down to the exercise of power, any tools will do. Failing combat, they will use words. Failing words, they will follow you around and interfere in your actions... and maybe, just maybe, if you supply a method for them to get into a political structure within the virtual world itself, they will play politics instead of killing, since politics is after all the ultimate human expression of the desire to exercise power over others. (No offense to the politicians!)

So one key problem to surmount is the fact that changing the medium of attacks will not prevent attacks from occurring. You'll find the Killer on IRC, on a web board, in chat rooms, in Ultima Online, and in muds everywhere--pk switch or no pk switch.

There are other thorny issues to wrestle with. A common call is for community policing--this is a position that I myself have often advocated. But it must be admitted that a virtual community is sorely lacking in one critical concept to be able to effectively police: identity. On the Net, what identity there is is very fluid. Whereas if you identify a criminal in real life, you can jail him, in cyberspace he effectively can become someone else entirely, leaving you holding an empty shell in your jail cell. A mule. A dummy character. An abandoned persona.

And that's assuming you can catch him--for how do you know who he is? One's anonymity in cyberspace, as we discussed yesterday, is a great empowerment. It's also a great problem for those who wish to track the behavior of repeat playerkillers.

Ultima Online originally was designed for full-bore community policing. We made safe towns, and originally supplied no other tools whatsoever. But those who sought to police accurately pointed out that since they could not track those who did evil deeds, the server code would have to. Hence the notoriety system. And now we are moving to a more precise and specialized system, because notoriety's key flaw was that it measured different types of behavior on the same scale, which rendered it highly problematic as a method of identifying criminals. The reputation system purposely tracks only one type of behavior for the purposes of flagging someone as "red," because that way it can serve as a more accurate tool for curbing that one type of behavior. Will it curb all methods of "attack"? No, because it is a specialized tool.

The idea of behavior tracking systems is not new. We of course have the concept of criminal records in the real world. On the EBay auction website, users of the site can award "stars" to other users, and you can get a sense of how trustworthy a person is before engaging in commerce with them. In many muds, people are flagged permanently as "thief" or "murderer" after one instance of thievery or murder.

The key issue behind having such systems is of course, who judges, and who punishes. The quote above from the "Black Rose Incident" is often used as an example of why consensus government fails in virtual settings, and a major reason why it fails is because the social mores of the playerbase are being dictated (or attempted to be dictated) by the game administration, rather than by the playerbase. All forms of compromise suggested by the admins in the incident fail to satisfy both parties, because they are not solutions offered by the parties. A similar dilemma arose in the incident described by Dibbell: in the end, the populace of the game felt themselves powerless, organized a government, and it meant nothing: the final action taken had to be taken by a "god."

It is no accident that in virtual communities, admins are often titled gods, wizards, and immortals.

This is of course an essentially paternalistic structure. One has to ask the very tough question: can an online community ever truly flower if it always has to run to Dad to deal with problems? The reason this is a critical question is because the presence of an all-powerful being is not a philosophical question in a virtual setting. In final analysis, it's the guy with the ability to flip the power switch on the server. In the case of the virtual rape on LambdaMOO that Dibbell described, the head admin came back and adbicated his powers to the populace.

Now, there's clearly a whole can of worms there regarding religion in a virtual setting that I am not going to open! However, the implications in terms of the development of online governments are very interesting and very important. Our challenge is that in UO, we have established what is to my knowledge the virtual setting with the largest scope of possible player actions and activities yet given to people without godlike powers. (On many MOOs and the like, the common player has godlike powers as a matter of course, which is a different big can of worms...!) With UO we--no, more precisely, you, the players of UO, have a unique chance to actually make a virtual world that sustains a virtual government that matters. No mean feat.

I say "that matters," because in the end the head admin at LambdaMOO had to take his powers back, and it is back to Dad as usual there. Just as we in Ultima Online had to retreat from the original design of full-bore player policing and add back in greater game admin involvement. But our intent is still clear: this is going to be your world. So if the tools do not suffice to handle the problems of anonymity, lack of accountability, binding to identity, and non-conbat means of attack, we need you to tell us what tools will. Because while we may have built the world, we don't want to be your parents anymore than we want to tell you what Virtues you must follow. That is a matter for your conscience and your free will, which try as we might, we could not take away even if we wanted to. So I look forward to fruitful discussion on the boards of things like townstones, locally defined "laws," methods of supporting player militias and towns in code, etc. It may take a while to come to fruition, but come it will, because despite what some may say, we, the developers are not saying "you're on your own." We're saying that maybe you might want to take off the training wheels someday, because bikes with training wheels get to ride on much more interesting terrain.

And after that, maybe we can see about changing that "just because we built the world" thing too. :)

-Designer Dragon

The full log of the Black Rose Incident makes for great ancillary reading.

Elizabeth Reid's master's thesis on Cultural Formations in Text-Based Virtual Realities is fairly academic, but fascinating reading. Highly recommended. A sample quote that seems relevant to today's discussion:

Cyberspace--the realm of electronic impulses and of high-speed data highways....--may be a technological artifact, but virtual reality is a construct within the mind of a human being... Virtual worlds exist not in the technology used to represent them, nor purely in the mind of the user... The illusion of reality lies not in the machinery itself, but in in the users' willingness to treat the manifestations of their imaginings as if they were real.
In other words, maybe--for those who said on Usenet yesterday that UO isn't a community--clap if you believe in fairies. If you believe in them, they will be real. And if you see a community there in UO--well, then, there is one. But if you refuse to believe, well, you'll never get to see the magic. Which would be a real pity.
Of course, you should read Bob Hanson's excellent and detailed Reputation System FAQ, which is our latest method trying to empower player policing of their environment.
Remember that head admin at LambdaMOO? His name is Pavel Curtis, and he has also written on the subject of virtual communities. An interesting, off-the-beaten-path essay to read is his take on transforming a MOO into a virtual professional community. You may want to seek out his work. A good introductory essay is Social Phenomena in Text-Based Virtual Realities.

I have received MANY wonderful letters since starting this regular essays column. But yesterday's letter from Postman77 of the Anti-PK Unified Council was one of those letters that makes your life's work all worthwhile. Thank you for your eloquence, Postman77. It made our day.