Softworlds — An Alternative Approach to Education

This was written a few years ago. While some of the references are a little dated, the ideas are not. It’s reposted here as the anchor article for the website.

I grew up in the San Fernando Valley, not really a place but rather a collage — oh, let’s not be fancy, more like a neglected bulletin board — of cars, auto dealers, roads, freeways, and parking lots. My teenage fantasy, as with many of my friends, was to become a race-car driver.

My friends and I never grew tired of jabbering about fast cars and auto racing. Feeding our obsession was that this was the mid-sixties, the era of cheap gas and humongous engines. In 1964, Pontiac stuffed a big engine into a small chassis and invented the “supercar,” a design which reflected the character of the teens who coveted it: too much engine, not enough brakes. Pontiac’s introduction of the GTO was followed by a slew of similar models from Detroit and the trend didn’t abate until a decade later when the Mid-eastern oil cartel doubled prices at the gas pumps.

While I admired the big American supercars and came close to buying one, my real love was genuine — that is, foreign— sports cars. Supercars were one-dimensional, two-ton, overpowered Godzillas; I preferred European models known for their balance of engineering, which included superior brakes, nimble handling, and lightness. After years of yearning, I bought a used Porsche 911.

The problem with owning a car such as a Porsche is that it’s designed for the Autobahn, the German highway without a speed limit. It’s as at home on American roads as Barack Obama at a Tea-Party rally. So, being 20, I carefully and maturely weighed my options. Hmm… safety for myself and others or cheap thrills? No contest: as with most young men, I couldn’t fire two synapses in succession without having them short out from hormonal fluids. I drove residential streets as if they were drag strips, negotiated mountain passages as if they were a road race course, and treated icy roads as a rally event.

I came to my senses, eventually, and decided to sell the Porsche and spare the lives of fellow drivers. Just kidding. I’d like to tell you that I wised up but the only way that was going to happen was if I was involved in some Back to the Future style story, where the future, mature me would knock some sense into the mindless, adolescent me. Nope, I gave up my stupid driving tricks because they weren’t adequate to satisfy my race-car driver fantasy. There just wasn’t sufficient compensation to make up for the hardships of owning a German sports car on a gas-station salary and night-school time table. The car had to go. What the heck; I made a profit on the sale and bought a relatively conservative Alfa Romeo sedan. All was for the best: my race car driver fantasy was nearly dead and I was still alive and, to my parents relief, I could begin paying attention to a real career — flunking out of college.

Three decades later, my driving priorities have evolved to old-fogeyism: my idea of a great car is one that gets where I want to go, with minimum stops at the gas station, and keeps my family safe along the way. But if I was 19 today, and wanted to race cars, I could do that for a fraction of the expense, none of the danger, none of the air pollution, and a chance at real satisfaction — all in the comfort of my own home.

Auto racing simulations (sims) have been around since the early days of video games, but Papyrus was the first game developer to push the genre beyond driving skills. Until Papyrus produced their first racing sim, Indy 500, your role in a racing game was to sit in the cockpit and drive; the rest of the racing team — engineers, mechanics, and so on — resided in your imagination. Papyrus expanded the nature of racing sims to include a mechanics’ “Garage” where you were able to make dozens of adjustments to the tires, suspension, spoiler, steering, and fuel allotment in order to meet the demands of a particular race track. In Indy 500, and nearly all subsequent race simulations, you’re not just a driver, you’re also an engineer.

Computer-generated virtual reality has long surpassed my fantasy-generated virtual reality. Compare a Papyrus racing simulation, or that of its later imitators, with my pathetic attempts to turn my Porsche ownership into a kind of race-car game.

When do we get to the educational bit?

Fun, racing sims may be, but this is about the future of education. Hence, imagine a race-car simulation that, in addition to the above, includes the following modes:

Mechanical: This module allows you to tune your engine’s fuel and electronic systems for best performance. Not enough? If your engine blows, you must rebuild it.

Engineering: While it’s nice to be able to adjust the suspension components of your car, what about modifying or designing your own while adhering to NASCAR guidelines?

My race car sim has four characteristics that make it a superior educational product:

First, it’s a design that could not be done without computers. Much of educational software consisted of glorified — not to mention expensive — flashcards, books, movies, calculators, markers and paints, typewriters, and poster-boards. No surprise that educational software, widely know as “edutainment,” has mostly disappeared. If we are going to spend the money, time, and effort for computers in education they should be used in a manner that the educational design could not be accomplished without computers.

Second, it is domain based. Auto racing is the context for the underlying models of mechanics, applied math and physics, and business.

Third, it uses a layered design often characterized as peeling the onion. The outmost layer is the most hands on, that is, the most concrete, while each successive layer is increasingly abstract. You don’t need to know how to spell Piaget to understand that the natural form of learning progression goes from the concrete to the abstract. Yet children usually are expected to dive right into abstractions such as math, physical science, and biology without a clue as to how it relates to anything in their lives.

Fourth, it is a programmable environment. It is the mutable nature of computers that attract many users and their very programmability that makes them learning machines.

Below, I’ll expand on the four characteristics.


What if I gave you the following assignment? You’re to gather, analyze, and compare statistics on warfare among four countries. You’re to comb the world, explore every nook and cranny, and overturn every stone in your search for relevant information. When you’ve exhausted your analysis of collected data you’re to seek more. You’re to use this collection of data to populate your humongous spreadsheet and massage it until the analytical capabilities of your software is exhausted and then, in pursuit of never-ending analysis, upgrade to the latest, fancier version of your spreadsheet. Price no object. You’re to remain insatiable. Sound like fun? Put that way, not to most of us, but I’m characterizing the actions of millions of players, worldwide, of the strategy card games that began with Magic the Gathering in the 90s, and describes the play of many collectible card games currently in vogue.

So an activity this addicting must have been designed by Disney? Spielberg? Columbian drug cartel? Nope. Collectible card games are designed by those wild and crazy mathematicians: if you take away the dress-up graphics and fantasy story worlds, Magic the Gathering and similar games are just contests of comparative statistics.

So why does everybody recoil in horror when faced with a study of comparative statistics but expend limitless energy and money on these games? Because the handful of rules for comparing the respective statistics, the hand-drawn original art, and the immense support system of information (contained in books, magazines, websites), along with the camaraderie of fellow users, work together to give you a world that is not a dry abstraction of statistics, but a world of captivating fantasy, adventure, intricate strategy, and competition. The popularity of card games such as Magic, as well as computer games with similar attributes — games dominated by strategy rather than just reflexes — such as Warcraft, Civilization, Age of Empires, and their near clones, lend support for the idea that children not only don’t avoid hard mental work but will pursue it with all their hearts if it is in a context that is personally significant. Seymour Papert, designer of the child-oriented programming language, Logo, coined a term for these games: hard fun.

The case for context-rich education stretches to anything in a child’s life. My niece’s love of horseback riding developed a sustained interest in biology. A trip to Washington for the six year-old daughter of a friend turned into, in order, an interest in U.S. Presidents, a fascination with the kings and queens of England, and, by the age of seven, watching and reading the plays of Shakespeare. A pair of guinea pigs for my daughter has resulted in… well… lots of material for the compost bin. An involvement with the concrete doesn’t always develop into something deeper but, on the other hand, the opposite is almost never true.

The tragedy of education is how meaning, and therefore interest, got stripped out of knowledge. How did it happen? For the most part, when we weren’t looking. Knowledge was increasing at such a rapid pace that experts — those who would focus on a tiny area of knowledge — had to be created. Stripping context from knowledge was an unintended consequence. Can we put meaning back into knowledge? Yes, by reversing the process through eXtreme domaining.

eXtreme domaining

In the Introduction, I talked about my passion as a teenager, auto racing. Because passions are always the best contexts for learning, students are going to be enthusiastic learners when the knowledge touches on their passions. And teachers are going to be enthusiastic teachers when the knowledge touches on their passions. Extreme domaining starts with a current passion about pretty much anything. It doesn’t much matter where you start; what matters is pushing past the surface to see where it takes you. Above, I talked about where an interest in auto racing can lead. That was then. Now I live in Seattle, so naturally, I prefer coffee to cars.

Edmonds Cafe

Now I want to talk about my cafe. It’s an imaginary cafe, which has the advantage of needing less capital. My cafe will reside in Edmonds, a town of about 45,000 people just north of Seattle. When my wife and I moved to Edmonds, in 1985, the only good coffee was in the expensive waterfront restaurants, but there are now several places in town to get a fine cup. To be successful, my cafe is going to have to have its own niche.

My cafe will be housed in 1000 to 1200 square feet. Besides coffee, it will serve the usual costly, bad-for-you, bakery food, but I’m also going to have heartier cold pastas because, besides yuppies and college students coming to work or study with their laptops, I want teens, young adults, and people my age (boomers) coming to play, to learn, to have fun.

As I said, this is an imaginary cafe, the right product for those with no money; but I want to bring Edmonds Cafe to life (so to speak) virtually, as the Edmonds Cafe Sim. The Edmonds Cafe Sim is a domain for learning about, as I will demonstrate, a whole lot of things.

Foremost, Edmonds Cafe is a business simulator. We do it virtually so you don’t have to lose your marriage, money, or both. We’ll start with a basic cafe simulator and create the economics of retail positioning (literally and figuratively), obtaining equipment, hiring and managing staff, buying wholesale coffee, tea, and pastries and marking up successfully.

That’s just the start. I said I wanted Edmonds Cafe to appeal to a variety of age groups and interests. Edmonds cafe will have two stories, the main floor and a well-lighted basement. Upstairs, a book and magazine store will surround the cafe interior. Downstairs will feature wireless Internet connection for users’ laptops as well as Internet and gaming machines for rent. Edmonds Cafe will have events in the basement: big-screen videos, gaming contests, music, and computer programming classes — which, by the way, are going to be teaching how to create the Edmonds Cafe sim.

To this point, I’ve described a complex business simulator by adding elements not typically found in cafes or, to be more accurate, usually not found all in the same place. Occasionally, you’ll find one of them but not all. But, since this is software, Edmonds Cafe will be like most sims sold as games: start with the basics, in this case, a coffee-and-pastries shop, and add whatever elements you wish. Each element must have its own business formulas, which will have to be integrated into the larger cafe business.

I did not add the additional business elements — books, gaming, and so forth — merely to add quantity to the financial model. That would be useful, primarily to learn about how different businesses work, but what I’m after is a new dimension that goes beyond a profit-and-loss statement. I want a sim that recognizes that a businesses exists within its culture. Call it, coffee anthropology.

Everything I know I learned in the coffee simulation

I buy green coffee beans online, 20 pounds at a time, and roast them in my almost smokeless machine. For best results you have to match your green beans to the characteristics of your home-roaster, and for mine, I strongly prefer Kenya AA. There’s no second choice. But there’s no such thing as labeled fair-traded, organic, or shade-tree Kenyan coffee. Dilemma here. I fear for the planet, global warning and all that, and I believe that people should earn a living wage or profit. But I’m an American baby boomer and, as such, fear not getting my needs met nearly as much. End of civilization or inferior coffee. Pretty balanced, from my perspective.

At the risk of being less glib, if I was convinced that buying only coffee labeled organic, shade grown, and fair-traded would contribute to good news for Earth and better prices for the small farming co-ops of Kenya, I’d get used to some other kind of coffee. But, according to coffee expert, Kenneth Davids, the issues are complex:

Coffee professionals often pit these two solutions (fair-trade solution and the snob solution) to the coffee price crisis against one another, the coffee snob side claiming that fair-trade coffees are a sort of socialist solution that eliminates market incentives for quality, while some fair-trade adherents implicitly knock the coffee snobs as obsessives who care more about aroma and mouthfeel than they do about human beings. What I find interesting about the results of the Pursuing Coffee Quality at the Supermarket cupping is how they suggest that these two solutions are not as contradictory as they seem at first. In terms of quality, the fair-trade Guatemala from Bucks County turns out to be every bit as distinguished in the cup as the three Kenyas. On the other hand, all of the finest Kenya coffees are produced by cooperatives of small holding growers working through an auction system, so it can be argued that an extra dollar spent for a good Kenya, regardless of the reason for that expenditure, is every bit as well-spent in social terms as a dollar spent on a fair-traded coffee.

Accurate appraisal or a terrific rationalization? I don’t know. I’ve followed lots of Internet links and have yet to come to a conclusion. My point isn’t to answer this question but to point out how complex — politically, economically, and ethically — such a simple decision can be, especially when the simple “feel good,” choice (organic, fair-traded, shade-tree grown) is unavailable. And that’s just for a cup of coffee. Alan Kay contends that fantasy is a form of simplification, that the attraction to it is that simplifying is necessary for control. We’re trading the reality of complex systems for the illusion of control.

I’m not against simplification; I’m against taking the word of others in how that simplifying was done. When I play a sim, I want to know what decisions went into the formulas that decide consequences. And when I design my own sim, I want to give players that information and let them do the research and make the decisions on consequences. More and more games are being designed to let players modify data and rules. My sim will make player control of data and rules paramount. I’ll talk more about that in the section, Layered.

If you follow the Internet links on the coffee market, you’ll also run into the economic and political issues of the World Trade Organization (WTO), the question of fairness regarding government subsidies, other effects such as corruption, large versus small farmers, water shortage (for irrigation), poverty, and so forth. If, to your simulation, you add music, books, and gaming, you could follow economic and social issues to the end of time. Just a few examples: consumer digital rights versus big publishers; big publishers versus authors and artists; distribution methods; recycling; who controls the Internet; violence in video games; gaming and learning; and the politics of all of the above.

Not interested in a coffee simulation? How about a baseball simulation? There are many really good statistical baseball games, both board, and computer. But what about the economics of a monopoly (baseball has a congress-granted exemption), the economics of municipal subsidies for stadiums, the history of and current race relations in sports, mass psychology (why do we root for strangers?), physiology and conditioning, sports psychology (motivation), medicine (from tendon replacement to steroids abuse) and the physics of baseball. Start with any passionate interest — painting, motorcycles, photography, weaving, horseback riding, computer games, or rap music — and it will lead you way beyond your starting point. Any domain which involves living beings will be sufficiently complex to make a learning domain.

A domain is a horizontal context — the boundaries are arbitrary. As you can see from my examples, a domain can sprawl like a suburban landscape. But computer simulations don’t sprawl, in part, due to the economics of selling games but also due to the difficulty of designing such a beast. Microsoft Office (speaking of sprawl) is not a good model for computer games. It’s easier to separately develop and make money off an action baseball game and a management baseball game than to combine them in a single game or as a main game and add-on module. The solution may be to design software with the ability to plug in modules rather than as a monolith.


“I want to get rid of the hood.”

-Alan Kay, speaking of his programming language/environment, Squeak

Pop star Madonna’s rise to fame in the 80s was aided by her unique approach to dressing for performances: she wore her underwear on the outside of her clothes. But long before Madonna made the scene, hobbyist auto mechanics invented a style that, because it wasn’t contrived, was far sexier than anything Madonna could muster: they exposed the private parts of their cars. What must have started as a convenience — hey, let’s leave the hood off until we’re done working on the engine — became part of the act. Eventually, the entire engine enclosure was cast aside leaving the hot rod’s engine not merely accessible but downright unavoidable. The hot rod’s design was an invitation: take me apart, examine my workings.

As a young adult, I worked at a gas station with some of those hot-rodders and saw them take to motors as if they and their cars arrived in the same litter. How did they get started as mechanics? By tearing apart and reassembling the engine of the family car? I’m sure these guys could have done it right, even on the first try. But not everyone is a natural. If I touched the family car, I’m sure my father would have suspected malpractice even if I carefully hid the leftover parts: Hmm… son, the car can’t seem to get up the driveway anymore.

I, too, wanted to learn about engines and I wanted both the car and me to survive the experience. Ideally (if it had existed at the time), I would have started with a computer simulation. A software “car engine” would have allowed me to experiment and learn without worrying that I’d damage something. It’s the difference between I ruined my Dad’s car and he’s going to kill me. and Click to Undo. My experience is not unique. Lots of people (even us old people) want to learn about and fiddle with complex stuff without having to worry about breaking it.

Along with an undo safety net, software simulations have another nearly incomparable advantage over the real thing. Any software system can be designed to be easily inspected. If you can design a simulation to model science systems, a nuclear power plant, a baseball team, and a health-care system, then you can design it in a way that the assumptions that went into the design can be revealed. From SimCity to Warcraft III, computer strategy games are just glorified spreadsheets (to be sure, really, really glorified). These simulations may model countless complex, ever-changing relationships, but, aside from the pretty graphics, they’re still, like those seen in an Excel spreadsheet — formulas with variables. So there’s no technical reason why software simulations can’t lay bare their assumptions. And any software that claims the tag “educational” should issue the same invitation as the hot rod — open my hood, toss it aside, have at my innards.

Faith-based Living.

As culture and technology get increasingly complex, my life has become increasingly faith-based. By “faith-based,” I don’t mean I live my life in accordance with a religious belief. I’m referring to having to accept products, ideas, practices, and everyday events — everything, from groceries to political positions — at face value. The food I eat and the politicians for whom I vote are equally processed, packaged, and, just in case I want to know what they’re made of, labeled.

I shop for groceries at Trader Joe’s at least once a week. I grab packages off the shelves and, if I’m lucky enough to find my reading glasses, try to view and comprehend the ingredients. I can barely make out the words on the food labels, but if I can I have no idea what they mean. “Thiamine,” “trans-fat,” “diglycerides”? Paging Dr. Weil.

And even when the label is a no-brainer, as in “free-range” eggs or chicken, I only think I know what it means. I’ve been buying “free-range eggs” for the last couple of years, not minding paying a little more because I can envision happy (well, happier) little hens hanging out with their hen friends at Starbucks. (We Americans love the fairy-tale of no-pain “sacrifices,” such as buying hybrid SUVs.) In an interview from, Author of The Omnivore’s Dilemma, Michael Pollan, explodes my happy delusion:

I read a lot of labels and I’m still a sucker for it. Free-range chicken, for instance, can mean nothing more than a 20,000-bird shed with a tiny little lawn and a little door that’s opened two weeks before the hens are slaughtered. These little yards are purely symbolic. Chickens don’t use these little yards because they have no experience using them. They’ve never been outside before; there’s not enough room for all of them and they’re a flock animal. So it’s a ploy to appeal to the consumer. When you see “free-range,” it’s not happening, but if you see “pastured” chicken, which you sometimes will at a farmers market, that’s real.

It’s comforting to rely on labels such as “organically grown,” “fair-trade,” “free-range;” to put your trash in the recycling bin rather than the garbage can; to drink green tea and other “anti-oxidant” laden foods; to know if you get sick you can be treated by (as one local radio spot put it) “a dream team of cancer treatment professionals.” But what the hell does any of this mean? To most of us, “anti-oxidant” means won’t get cancer (but if you do the “dream team” is on the case). So, not only do we drink green tea, but when we go to Costco we buy green-tea pills. Oops:

Many people drink green tea in hopes of warding off cancer, heart disease and other ailments, but people who take concentrated green tea supplements with the goal of multiplying those benefits should think twice. A study completed by UM researchers indicates that extremely high doses of green tea extract actually may activate rather than shut down genetic mechanisms that help certain tumors survive and grow. (Source, no longer available.)

Multiply the green-tea and “free-range” delusions by the zillions of misinformed decisions with which we run our lives and you’ll get what I mean by “faith-based” living.

If you watched CNN 24/7, poured through books, magazines, and newspapers, drove Google’s search engine like an Italian sports car, and laid all the bloggers end-to-end, you still wouldn’t reach a fraction of the knowledge and expertise that you’d need to acquire to understand all that affects you. Just maybe part of the reason we need all that Prozac and vino comes from the anxiety over having to trust what most of us don’t trust. We do our best to sooth ourselves with drugs. Who cares that we know nothing about what goes on behind the scenes of every important aspect of our lives.

While food corporations may not really want me to look at or understand their tiny labels, politicians are another thing. They know that thinking makes my brain hurt so they (or their detractors) give me concept labels such as “Christian right,” “compassionate conservative,” “liberal,” and “fair-traded” (guaranteed to give lobbyists their money’s worth, as opposed to “free-range” politicians who vote according to poll numbers). Political labels are marketing devices designed to turn off your thinking and get you to vote solely with emotion. At this point, I don’t think I have any ideas to offer on how software can directly influence the political process — Diebold cornered that market years ago — but I hope that, over time, using software simulations to dig into other areas of our lives can help create a relentlessly curious culture; that software simulations will help us develop habits of inquiry that lead us out of our faith-based lives.

So a domain provides the interest, layering provides depth, simulation add-ons can provide breadth. Start with the café, then attach a bookstore, then a coffee farm. An educational context becomes what it should be: start any place that draws your interest and go down, up, or sideways. This is the natural way of learning. Lectures and books — lineal by their nature — may have been our best tools in the past, but they are not any longer now that we can create software simulations of anything. Combining the attributes of domains, layering, and breadth can free us from the contrived separation of ideas, which leads to boredom. You might hate math, but if you’re a race-car driver, you’re going to be a practicing physicist.


Students not having users is the problem, and hardly stands as a convenient excuse. Can you imagine having a bunch of MDs running around who never had “users” during the course of their education? Of course, students should be required to have clients and end users. No wonder they are so clueless.

— Bruce Tognazini

I’ve made a case for simulations as primary learning tools, and it’s now common for future-oriented educators and computer scientists to state that the future of education should and is going to come via using computer simulations. But if we stopped there, this would miss a major opportunity: better than learning by using simulations is learning by creating simulations. The skills that go into creating a simulation are so vast that they should cover most anything traditionally taught in school and so much more. I will go as far to say that if there was a choice between learning with simulations and learning by creating a simulation, and at the end of the creation of that simulation it was destroyed, the latter

experience would still be far more useful.

How students can create their own simulations

Think mentors not teachers, and just-in-time-instruction of concepts and knowledge necessary for the development of the simulation. Mentors could be comprised of a team of college graduate and upper-division students, artists, animators, programers, writers, and librarians, orchestrated by a traditional but highly experienced teacher versed in both project development and teaching. Students who have completed a simulation would be moved into mentoring roles for the next generation of young simulation developers.

Only the most naive would believe that any dramatic change in the public school system has more than a slingshot-and-stone chance of displacing the traditional classroom anytime soon, but it’s something that can be seeded and grow, starting in charter schools, libraries, college labs, and club houses. Imagine a group of students researching a model in the library and on the Web, questioning participating teachers, college professors, grad students, and librarians, being mentored in relevant statistical models, in animation and art, in programming, and anything else that goes into creating a software simulation. Imagine the subsequent curiosity, community, and ownership that would emerge from this sort of experience. Imagine the subsequent demand, by these young people, for evidence and transparency from experts, politicians, and marketing departments. Imagine the knowledge, creativity, and industriousness these students would bring to their community.

Finally, I don’t really advocate destroying finished simulations. On the contrary, they could be studied, modified, expanded and, I hope gathered into a dynamic Wikipedia or even merged in the current static version to create the Simpedia.


Student-created simulations aren’t going to happen without “user programming” in the sense of HyperCard-for-simulations. Unfortunately, Web technology has evolved in a manner that user programming isn’t going to happen there: learning three programming and two mark-up languages to create a rich web app is for web-programming repeat offenders, not for those learning to develop a one-off simulation. I suspect that a client-server model will be the best solution. Nothing says user like mobile, so I look for smart-phone and tablet as the major client platforms.