4 This is does not say, however, that no generality exists here. How any one of us learners throws light, both by comparison and contrast, on how others learn. Learning is not infinitely variable and there are patterns and principles to be discovered, patterns and principles that ultimately constitute a theory of learning. Indeed, what I am offering here is a case study meant to offer suggestions for a theory of how deep learning works see, also, Barsalou 1999a, b; diSessa 2000; Glenburg 1997; Glenburg & Robertson 1999). In the end, I hope to convince you that todays young people often see deeper and better forms of learning going on in the games they play than in the schools they attend. Though some of the information below is personal, I intend and hope readers will think about the comparisons and contrasts of my learning experience with RoN to the sorts of learning that goes on in schools. Ironically, perhaps, a baby-boomer trying to learn a modern computer or video game is not, in some respects, unlike a child in school trying to learn science or math. Both parties are being asked to learn something new and, in some respects, alien to their taken-for-granted ways of thinking.
PREPARATION FOR LEARNING: BEFORE RON By the time I started up RoN I had played lots of computer and video games. They had taught me new ways of learning and new things about myself as a learner (Gee 2003). However, I had not had good experiences with RTS games. I felt overwhelmed by their many details and by the pressure of competing in real time. I had watched my twin brother play RTS games at a high level and was amazed by the number of details he had mastered and the speed with which he had acted and thought in the games. I had watched my seven-year-old play the wonderful Age of Mythology and was stunned that he and his friends could play such a complicated game so well. Far from giving me
5 confidence these experiences just made me think that I was not suited for the micromanagement and on-the-spot decision-making RTS games demanded. In regard to RTS games, I was an at risk learner, at risk for failing to be able to learn and enjoy these sorts of games. Though timid about RTS games, when WarCraft III came out, I tried it, prodded by my brother who loved the game. I made some progress in the single-player campaign, but eventually found the game too hard. We should pause a moment, though, at this phrase too hard. WarCraft III is a superbly designed game. In fact, it is well designed to get itself learned. So when I say it was too hard, what I really mean is that I failed to engage with it in a way that fully recruited its solid design and learning principles. Good games are never really too hard. They fail, for some players, either because their designers did not use good learning principles or because players have, for one reason or another, failed to engage the good learning principles that are built into the games. So something has to come even before good learning principles. What has to come before is motivation for an extended engagement with the game. Without a commitment to an extended engagement no deep learning of a complex domain can happen (diSessa 2000). So what made me motivated to offer such extended engagement to RoN and not earlier to WarCraft III? Well, as good as WarCraft III is, RoN is yet better at allowing newcomers to learn it. But, more importantly, and ironically, perhaps, my failure at WarCraft III motivated me to try RoN. I had liked WarCraft III. It had made me feel that RTS games were important and worth playing. Though I had had limited success with the game, I had had some small success that made be feel that at another time and place, perhaps, I would do better. It had led me to read about RTS games and reflect on them. WarCraft III, it turned outthough I realized this fully only

6 when I started RoNhad prepared me for future learning (Bransford & Schwartz 1999) of RTS games. When I started RoN, I realized that I already knew something, somewhat more than I had thought. I felt I had a small foot up. In a school setting, my experience with WarCraft III would simply have been seen as a failure as I received my low or failing grade. In reality, it was not a failure, but an important precursor for later learning. My experience with WarCraft III is what I will call, following the work of Stan Goto (2003), a horizontal learning experience. Vertical learning experiences are cases where a learner makes lots of incremental process on a scale from low skills to high skills, as if moving up a ladder. Horizontal learning experiences are experiences where one does not make a lot of progress up the ladder of skills, but stays on the initial rungs awhile, exploring them and getting to know what some of the rungs are and what the ladder looks like. Horizontal experiences look like mucking around, but they are really ways of getting your feet wet, getting used to the water, and getting ready, eventually, to jump in and go swimming. They may, in one form or another, be essential to learning, or, at least, essential for learners who are at risk. So, is there a contradiction in saying that when I started RoN I was still an at risk learner, but that my experiences with WarCraft III were important preparation for future learning? No. All that my being at risk meant, in the end, was that if RoN had failed to reward my preparation for future learning (the future was here with RoN) or had been a bad learning experiencea real failurethen I may have given up on RTS games forever, assuming I was too dumb to learn them. This is all at risk needs to mean in schools, too, though there it often means giving at risk learners a special dumbed-down
7 curriculum meant to catch them up on basic skills, a curriculum that all too often is a bad learning experience for these students. Computer and video games have a built in advantage in the creation of motivation for an extended engagement. Human beings feel that their bodies and minds extend, in a rather intimate way, to the area around them over which they have direct control, usually a fairly small area (Clark 2003). Thus, as I type, I feel that my keyboard and mouse seem almost like extensions of my fingers, just as blind people often feel that their cane is an extension of their hand. The space closely around my body seems to be connected to it in such a way that I can feel that it is being invaded by others. When humans can manipulate something at a distance, for example controlling with a keyboard a far-away robot seen on a screen, they get an uncanny feeling that their minds and bodies have been vastly extended (Clark 2003; Goldberg 2001). When people are playing a computer or video game they are manipulating a character (or many different things in a RTS game) at a distance in a very fine-grained wayin this case a virtual distance. They feel that their minds and bodies have been extended into this virtual world. This process appears to allow players to identify powerfully with the virtual character or characters they are playing in a game and to become strongly motivated to commit themselves to the virtual world the game is creating with their help. When students are learning a content area in schoolsuch as some area of sciencethis domain could be seen as a special world of its own, the world of doing science in a certain way and acting with certain values. Students could be encouraged to take on identities as scientists of a certain sort, to see and think about themselves and their taken-for-granted everyday world in new ways. In this case, school would be

8 functioning more like a good game than traditionally schooling which stresses knowledge apart from action and identity.
RoNs TUTORIALS: FISH TANKS Lets begin to explore what makes RoN a good learning engine. When a player starts RoN, the designers immediately have two problems. First, learners are all different and the designers dont know what each one already knows, nor what their favored style of learning will be. Second, learners dont necessarily themselves know how much they do or do not already know and what their best style of learning will be in a given situation. Schools tend to handle these problems by assessing the learner and then deciding for the learner how these problems ought to be dealt with. RoN, like many other good games, solves the problem by letting learners assess themselves and learn things about what they do and do not know and what style of learning suits them here and now. Learners then decide for themselves how they want to proceed. Of course, RoN is designed to assist learners in this task; they are not left solely to their own devices. By the time you have interacted with RoNs tutorials and skill tests and played your first few real games, you know a good deal about yourself as a learner, in general, and a learner of RTS games, in particular [In this paper the games called Quick Battles in RoN are what I refer to as the real game; a game called Conquer the World is also part of RoN , but I do not discuss that game in this paper. Conquer the World is composed of Quick Battles and other elements]. When RoN starts, you see a screen with the following choices (the numbers on the right are dates, ranging from 60 AD to 1940):

9 Tutorial

Learn to Play -Quick Start

Bodicia

Beginning Player

Alfred the Great

The 100 Years War
Experienced Real-Time Strategy Player

Henry VIII

Battle of Britain

Advanced Topics

Right away the learner sees choices: jump right in (Quick Start), learn step-bystep (moving from beginning player to experienced player to advanced topics), start with the experienced or advanced topics (thereby testing ones own assumptions about ones previous knowledge), or skip the tutorials altogether. Choice is built in from the beginning. Notice, too, there is no remedial in this learning world. You begin where you begin and move to advanced when you move there. None of this is timed. There are no invidious judgments based on ones previous failures. When the learner places the mouse on each choice above, a box is displayed at the bottom of the screen detailing just what historical event each choice will deal with and what skills the learner will learn by making that choice. Table 1 below shows each choice and what is displayed in the box when the learner places the mouse over that choice:

Quick Learn Learn-As-You-Play Introduction -One-on-one battle -Hints and suggestions as you play Bodicia -Beginning Player 60
Bodicia - Tutorial 2 Help a queen fight off the Romans to reclaim her nation - Unit selection - movement - map scrolling - help text - basic combat Alfred the Great -Beginning Player 878
Alfred the Great - Tutorial 3 Turn back the raging Viking hoard - Constructing and using buildings - Training units - Minimap The 100 Years War -Experienced Real-Time Strategy Player 1337
The 100 Years War - Tutorial 4 - Library research - Food, timber, and metal gathering - Capturing cities - Repairing buildings - Unit combat advantages and disadvantages - Transporting units across water Henry VIII -Experienced Real-Time Strategy Player 1513
Henry VIII Tutorial 5 Defend against Scottish raids - City construction - National borders - Knowledge and wealth gathering - Merchants and rare resources Battle of Britain -Advanced Topics 1940
Battle of Britain Tutorial 6 Battle the Germans in Britains finest hour - Diplomacy - Air combat - Generals - Oil - Enhancement buildings - Formations
Table 1: Tutorial Screens
All is not as it at first seems here, though. What you see in the boxes are by no means all or even the majority of the skills you need to play RoN well. They are the basic skills you need to play the game, but basic in a special sense: they are the skills that allow you to actually start playing and learning from playing. I will point out below that the designers of RoN dont just take it for granted that players will be able to move from the basic skills in the tutorials to learning by playing. Once the player actually starts the real game, they ensure that this transitionfrom basic skill learning to learning by playingwill happen. But before I tell you how they do this (its all about players being able to customize the game to their own desires and goals), let me finish my discussion of the tutorials. If we look back at the terms experienced real-time strategy player and advanced topics in Table 1 we see something interesting. Experienced and advanced mean something quite different here than they do in places like schools. The skills taught in the tutorials, as we have said, are basic (in the sense defined). They are not the deeper skills required to play RoN or any other RTS game well, skills like time management, speed, micro-managing many details at once, and strategic thinking. So it may seem odd that terms like experienced and advanced are used. But experienced and advanced here mean what players need to know to begin to take yet greater control over their own learning by discovery through playing. They dont mean at the top of the vertical ladder of skills (or you get an A in this subject). The player is experienced and advanced in the sense of being prepared for future learning on site, not in the sense of necessarily being an expert. Each tutorial places its basic skills in a scenario that is just a simplified version of the real game. This allows learners always to see how these basic skills fit into the game

12 as a whole system and how different skills integrate with each other. In school, on the other hand, very often these days children are exposed to basic skills one-by-one, stepby-step. For example, in early reading instruction they are taught first awareness of the sounds that compose words, then the decoding of letters, then reading aloud to attain more fluent decoding, then comprehension skills (Coles 2003). Then and only then do they get to play the real game of reading, namely reading for meaning and to carry out their own purposes. In schools, too often, skills are decontextualized from the system (the game) and from each other. This never happens in RoN or any other good game. As an example of what I am trying to get at here, consider the tutorial labeled Alfred the Great (see Table 1 above). When you click on this tutorial, while the scenarios is loading, you see the following in print, while listening to the same thing (my own remarks below are placed inside brackets):
Eight hundred years after Bodicia rebelled against the Romans [this event was dealt with in the preceding tutorial labeled Bodicia], Britain was savaged by repeated Viking attacks. Alfred King of Wessex has been paying tribute to stave off the raiders, but in 878 the Vikings prepare for conquest. After a defeat, Alfred retreats to rebuild his forces and drive the Vikings away.
Once you press Start to start the scenario, you see the Vikings attacking the British town of Ethandum and hear the following:
Alfred suffers a stinging defeat when the Vikings attack in battle. The Norsemen loot the town and Alfred is driven back to his stronghold in Carlisle. Alfred must rebuild his forces and attempt to retake Ethandum.
Here we see that the scenario opens with a short context within which to understand and make sense of what one is going to do. After the Vikings victory, the scene changes to the British town of Carlisle, the place to which Alfred has retreated. This is where we will play out our tutorial. We dont start from scratch, though. We start in the Classical Age, the second of RoNs eight ages, not in the Ancient Age where real games start. We also start with a large city, granary, lumber mill, market, and fort, as well as several citizens and their farms. While the game always starts with a (small) city and some citizens, the rest of these things players would normally build for themselves. Furthermore, while players in the real game always start with a library where they can do lots of different types of research, including research that leads to new ages, this scenario has no library, because we are not going to use it. The setting of the scenario has been designed to be a minimal game setting with no more and no less than we need to learn at this point, but with enough to see how things fit together as a system. I will call this a fish tank tutorial, because a fish tank can be, when done right, a simplified environment that lets one appreciate an ecosystem (e.g., a river, a pond, or reef in the ocean) by stripping away a good deal of complexity, but keeping enough to bring out some basic and important relationships.

This is a preformed scenario where you can play the game at your own pace. Try to capture the Barbarian capital or conquer 70% of the map. Therell be hints and reminders to help you as you play.
The Quick Start scenario is actually the real game set an easy level of difficulty with copious comments and hints. There is an opponent (in the real game you can have multiple opponents), but the opponent builds up slowly and does not make the smartest choices. The player gets a real sense of being in the game, even a sense of urgency, but cant really lose or, at least, lose at all early before having put up a very good stand. Let me just show you just the beginning of the Quick Start tutorial, so that you get the flavor of what is going on. The material below deals with how I operated in the Quick Learn tutorial. Here, once again, I print my own remarks in brackets:
[Voice:] The leadership of your fledging tribe has fallen on your shoulders. The first task is to unify a new nation under your rule. Youre free to build your
18 nation at your own pace. Occasionally you may receive advice to help keep things moving, but otherwise its all up to you.
[If you wait, eventually you will read and hear hints about what to do. But there is wait time here to allow you to explore the screen and click on whatever you like. I clicked on the scout. When I did so I saw the box printed below and simultaneously heard the remarks listed below that:]
Scout: Currently selected (Hotkey )
.Scouts, Ancient Age [picture with hotkey] fast, but unarmed; good for exploring the map and finding enemies.Can spot hidden enemy units, such as spies and commandos.Can also destroy enemy spies.Strong vs. spies; Weak vs. Archers, Gunpowder Infantry
[Voice:] This is your scout. Use him to discover rare resources or locate the enemy position. Scouts are very fast and can see farther than most units, but cannot attack. You can move your scout around the map manually or click the auto explore button to have him explore on his own.
[After a few moments, I saw the message printed below on the top left of the screen and simultaneously heard the words below that:]
19 [Top Left Corner:] Create citizens to gather more resources
[Voice:] Your first priority in the Ancient Age is to create citizens and gather resources. Click your capital city and click Create Citizen to add to your work force. Put as many of your people to work gathering food and timber as you can. If youre running low on resources, you can always build farms to gather food and fill up woodcutters camps with citizens to gather timber.

[During another wait time, I clicked on the library and then clicked on a red button that lets the player research military technologies. Once this research is finished, the player can build military buildings. After clicking on the red button, I hear the following]:
[Voice:] Now that you have studied the first red military technology, you can build a barracks and beginning training troops to protect your nation.
[After a few moments, I hear the following:]
[Voice:] If you want to see more of the map, you can always zoom in and out by using the mousewheel or pressing Page Up and Page Down on the keyboard.
20 The Quick Start tutorial goes on this way for a while. If the player explores and does things, the tutorial confirms these acts and explains them. If the player waits, the tutorial prints a hint about what to do on the top left of the screen and says the hint orally and explains what it means. There are also, from time to time, remarks about how the game works, for example, the remark above about how to see more of the map. The tutorial is a nice dance of the players actions and designers guidance and instructions. Midway through the Quick Start scenario the following box pops up:
MID GAME At this point you should be having fun exploring the game and following some of the prompts that appear in the top left of the screen. If youre not having fun, you may want to try one of the following options.
Im having fun I want to continue playing I need to know more basic information, take me back to the tutorial screen The game is too slow. Let me start a Quick Battle.
This box is an excellent example of alerting players to the fact that they need to assess their own progress, desires, and learning styles. They need to be proactive, make decisions, think about what they are doing and learning, and take control of their own learning.
21 When I started RoN, I started by doing the Quick Start tutorial. I did this for a rather perverse reason. I was so sure I would fail that I wanted to reconfirm my own view that actually playing the game would be too tedious and complex for me. What happened was that I got excited, feeling, Wow, Im actually playing a RTS game and winning, to boot! (of course, this may remind you of the great scene in the movie What About Bob? where the ever fearful Bob is lashed to the sail of a sail boat and yells to his friends, Look, Im sailing, Im actually sailing!). The Quick Start tutorial is a sandbox. The sandbox feels like the real world to a child, but is guaranteed not to destroy the childs trust and ego before the child is strong enough to face more significant challenges. But this tutorial is a specific type of particularly efficacious sandbox. It is a sandbox with a wise parent present to guide and confirm efficacious play in the sandbox, in the case proactive game designers. Lets call this a supervised sandbox. Once I had done the Quick Start tutorial, I was energized to learn more, but, of course, I could not remember all the details the tutorial had introducednor was I meant to. Now I could turn to the specific fish tank tutorials and make each of these details, through focused practice, a part of my embodied intelligence and not just the caprice of my risky verbal memory. But I also knew now how these details fit into the larger scheme of the whole game, remembering that even in the fish tank tutorials skills are also introduced in terms of how they relate to other skills and to a simplified game system. Of course, other learners might do the fish tank tutorials first and use the supervised sandbox of the Quick Start tutorial to assess their learning and readiness to jump into the real game. There is one last important point to make about the Quick Start tutorial. What it does, in addition to what we have already surveyed, is introduce the genre of RTS games

22 to players who may have not played such games before. Genre just means what type of thing a thing is, for example whether a novel is a mystery, romance, science fiction, etc., or a piece of writing is a story, report, essay, and so forth. RTS games are one type of computer/video game (there are many others, e.g., shooters, adventure games, roleplaying games, etc.). They involve typical actions, rules, and strategies that are different from those involved in other types of games. Schools often try to teach kids to read and write, rather than read or write specific types of things like stories, reports, field notes, essays, or expositions. But, just like games, these different types of reading and writing operate by different principles and are used to carry out different types of actions. Good learning always involves knowing early and well what type of thing we are being asked to learn and do (Christe 1990; Cope & Kalantzis, 1993; Martin 1990). Learners need to see this type of thing in action, not to be given static rules, if they are really to understand. In fact, for most types of things like types of games, writing, movies, and so forththere are no clear and static rules that define different types. Each type (e.g., a RTS game or an essay) is composed of many different instances that are variations around a theme. The only way to learn is to see some instances and live with them concretely. Sure, there are some things you need to learn that help you to play most games, regardless of their type (e.g., moving and clicking a mouse), but these are the tip of an iceberg compared to what you need to know about how different specific types of games work. Thinking a RTS game is a shooter will make you a particularly bad learn of the RTS game or, at the least, will make you disappointed with it and not like it. The same thing is true of writingthere are some basic all-purpose things to learn (e.g., where to
23 put commas and periods), but they, too, are but the tip of an iceberg, and writing an essay thinking it is supposed to be a personal narrative wont work.
RoN: UNSUPERVISED SANDBOXES We are now readyas the player isto leave RoNs tutorials and start the real game. I said that the skills RoNs fish tank tutorials taught were basic skills in the sense that they are the skills that will allow you to actually start playing and learning from playing the game. The designers of RoN have ensured that these skills, once you learn them, will function just this way by building certain devices into the game play itself. When you leave the tutorials and actually start playing, there is a pause key that will stop time. This allows you to explore what icons on the screen mean and think about what you want to do. When time is paused, your opponent(s) do not continue building and so you do not have to worry about falling behind. Furthermore, you can set the game at one of two easy difficulty settings (easiest and easy) that greatly decreases the pressure of time. On these settings, opponents move slowly and not always in the smartest fashion. Finally, you can turn on (of off) hints that appear from time to time to remind you of what you have learned in the tutorials and teach you new things. What all this means is that the player learns in the tutorial just enough to move on to learn moreand more subtle thingsby actually playing the game, but playing it in a protected way so that deeper learning can occur through playing. The player can customize the game play to be, in fact, another sort of sandbox, in this case what we might call an unsupervised sandbox. The player is protected to explore and take risks, but, aside from the small hint notes that can be turned off and on, there is much less guidance and direction from RoNs designers.

24 We see, then, that in RoN there is no clear division between the tutorials as a learning space and the players first real games with difficulty set on easiest or easy and use of the pause key. These first real games are actually hidden tutorials which assist players in teaching themselves how to play RoN, not as a set of discrete skills, but as strategic thinking using an integrated system of skills. These unsupervised sandboxes make for a smooth transition between official tutorials and really playing the game (set on normal or a harder level).
LEARNING AND PLAYING But there is a yet deeper principle at work here than the smooth transition between tutorials and playing. In a good game like RoN there is never a real distinction between learning and playing. The tutorials are simplified versions of playing the game. The game itself has a number of difficulty levels and at each level players must refine their skills and learn new ones. Players can also play other players in a multiplayer form of RoN on the Internet, getting into games with others whose skill levels are equivalent to their own. They can move up to play better and better players as their own skills progress, and, in doing so, will constantly be learning new things. When learning stops, fun stops, and playing eventually stops. For humans, real learning is always associated with pleasure, is ultimately a form of playa principle almost always dismissed by schools. There is one crucial learning principle that all good games incorporate that recognizes that people draw deep pleasure from learning and that such learning keeps people playing. Good games allow players to operate within, but at the outer edge of their competence. At lots of moments, a good game feels highly challenging, but
25 ultimately doable. Perhaps the player fails a few times at a given task, but good games show how much progress the player has made on each try and the player sees that this progress is increasing each time he or she fails. Eventually success comes. This feeling of being highly challenging, but ultimately doable, gives rise to a feeling of pleasurable frustration, one of the great joys of both deep learning and good gaming. Good games, however, do not at all points operate at the outer and growing edge of the players competence. This is because they also recognize another important learning principle, what I call the principle of expertise, because it is the foundation of expertise in all significant domains (Bereiter & Scardamalia 1993). When learners learn a new skill set/strategy, they need to practice it over and over in varied contexts in order to make it operate at an almost unconscious routinized level. Then they are really good at it. But they are also in danger of resting on their laurels and learning nothing new. At this point, a good game throws a problem at the player where the routinized skill set/strategy wont work. This forces the player to think consciously again about skills that have become unconscious, taken-for-granted, and routine. The player must integrate his or her old skills with new ones, forming a new and higher skill set/strategy. Now, in turn, the game will let this new skill set/strategy get practiced until it is routine. The player has moved to a new level of expertise and will then eventually face a yet harder problem that will start the process all over again. Thus, good games cycle through times where they operate at the outer edge of (but within) the players competence and times where they allow players to solidify their skills. The times where players are solidifying their skills to the point of routine and taken-for-granted application give rise to another form of pleasure, the pleasure of mastery. Games cycle through periods of pleasurable frustration and routine mastery, a cycle of storm and calm.

26 These cycles are actually clearer in games like shooters (e.g., Return to Castle Wolfenstein, Deus Ex, Unreal 2, etc.) than they are in RTS games like RoN. In a game like RoN they are partially under the players own control through the ways in which players can customize the game to their own skill level and interest. Players can themselves choose periods of skill solidification and high challenge, though the game gives them plenty of feedback as to when things are getting too easy or too hard. But how do players know when they are prepared to move beyond the unsupervised sandboxes they can create by playing the game on easy difficulty levels? How do they know when they are ready to move on to the more rigorous challenges of the normal difficulty level and harder levels, as well as multiplayer play? As it happensas happened with me, in factthe player can certainly tell the game is becoming too easy by how fast and thoroughly he or she gains victory over the opponent(s). However, I found that when I moved on to the normal level, it was, at first, too hard, harder than I had thought it would be, given my swift victories on lower difficulty levels. The problem, of course, was that I had not properly evaluated my skills. I did not realize that my skill sets/strategies were not fast and efficient enough to take on harder challenges. RoN does two things to speak directly to this problem. First, it offers players a whole set of Skill Tests. I list the skill tests in Table 2 below. Note that some tests are defined in terms of skills (e.g., mouse clicking) and others in terms of strategies (e.g., getting to the Classical Age fast). As we have said, in games, skills are always seen as strategies.
Skills Tests 1. Aging Madness Age 2
How fast can you get to Classical Age? Find out if your resource management skill is good enough.

Aging Madness Age 4

How fast can you get to Gunpowder Age? Find out if your resource management skill is good enough.

Aging Madness Age 8

How fast can you get to the Information Age? Find out if your resource management skill is good enough.

Raiding Party

Take your bloodthirsty Mongol horde and pay a visit to some enemy towns in an exercise of micromanagement.

Hotkey Handling

Do you know your hotkeys? This is a test of hotkey knowledge.

Protect the Wonder

Protect your Wonder from jealous enemies in a exercise of defense.

Tactics

Defeat the enemy troops to take control of a valuable resource without losing more than half your army in this test of generalship.

Whack the General

How fast can you click your mouse? This is a test of clicking ability.
Table 2: RoN Skill Tests These skill tests allow players to assess how well their skills fit into an efficient strategy sethow well integrated with each other and with the game as a system they are.

28 The skill tests are, as they often are not in school, developmental for the learner and not evaluative (judgments carried out by authority figures). Furthermore, they are tests of what skills mean as strategies, not decontextualized tests of skills outside contexts of application where they mean quite specific things. The second thing RoN does to solve the problem of letting players know where the cutting edge of their competence is is to render the whole matter social. Sadly, I failed my very first skill test several times. But I knew just how to increase my learning curve so I could pass the test. Every player knows there are an immense number of Internet sites and chat rooms from which loads of things can be learned and to which lot of questions can be directed. One very effective thingthough there are a great many othersthat players can do is download recordings of RoN games played by players at different levels of expertise. Players can watch these to learn new things at ever increasing levels of expertise. Players can also easily record their own games and review them. They can also pit the computer against itselfat whatever level of difficulty they chooseand watch how things are done. On line, there is a world-wide university of peers and experts available to any player all the time. RoN lists its own web site on its program file, a site with much information, chat rooms, and links to other sites. There are also published strategy guides and many game magazines that will discuss games like RoN, offering hints, guides, and other sorts of helpful information. This social aspect of RoN, and games in general, makes RoN and other games the focus of what I have elsewhere called an affinity group (Gee 2003). An affinity group is a group of people who affiliate with others based primarily on shared activities, interests, and goals, not shared race, class, culture, ethnicity, or gender. The many sites
29 and publications devoted to RoN create a social space in which people can, to any decree they wish, small or large, affiliate with others to share knowledge and gain knowledge that is distributed and dispersed across many different people, places, Internet sites, and modalities (e.g., magazines, chat rooms, guides, recordings, etc.). Distributed and dispersed knowledge that is available just in time and on demand is, then, yet another learning principle built into a game like RoN. Too often in schools knowledge is not shared across the students, is not distributed so that different students, adults, and technologies offer different bits and pieces of it as needed, and is not garnered from dispersed sites outside the classroom (for a case where it was, see Brown 1994). RoN has no such problems.

34 Cope, B. & Kalantzis, M., Eds. (1993). The powers of literacy: A genre approach to teaching writing. Pittsburgh: University of Pittsburgh Press.
diSessa, A. A. (2000). Changing minds: Computers, learning, and literacy. Cambridge, Mass.: MIT Press.
Gee, J. P. (2003). What video games have to teach us about learning and literacy. New York: Palgrave/Macmillan.
Glenberg, A. M. (1997). What is memory for. Behavorial and Brain Sciences 20: 1-55.
Glenberg, A. M. & Robertson, D. A. (1999). Indexical understanding of instructions. Discourse Processes 28: 1-26.
Goldberg, K., Ed. (2001). The robot in the garden: Telerobotics and telepistemology in the age of the Internet. Cambridge, MA: MIT Press.
Goto, Stan (2003). Basic writing and policy reform: Why we keep talking past each other, Journal of Basic Writing. 21: 16-32.
Kelley, A. E., Ed. (2003). Theme issue: The role of design in educational research, Educational Researcher 32: 3-37.
35 Kress, G. & van Leeuwen, T. (2001). Multimodal discourse: The modes and media of contemporary communication. London: Edward Arnold
Martin, J. R. (1990). Literacy in science: Learning to handle text as technology. In F. Christe (Ed.), Literacy for a changing world (pp. 79117). Melbourne: Australian Council for Educational Research.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge, Mass.: Harvard University Press.