kip/bot/blog

The state of Idaho is committing to a state-wide program of chess instruction for 2nd and 3rd graders.  As someone who loved playing chess as a kid, worshipped the sane Bobby Fischer and dreamed of a comeback, and a proponent of constructivist approaches to education, I normally love this kind of thing.  (There’s an argument that chess isn’t external or project enough to be constructivist, but there’s some alliance there.)

Now, I’m starting to wonder.  Is chess nothing more than a measure for under-funded schools — short on textbooks, lab equipment, notebooks, teachers — to bring something to these kids?

On the upside, they’re using the web to bring a chess expert into the classroom and it does attempt to bring some history and math into it.

In grad school, I got footnoted in a couple of labor economics articles. Since it was my first (and ultimately only) year in grad school, getting footnoted was a big deal . . . enough of a big deal, in fact, to get a bitchy comment from the TA.

So why write about it? Cuz I owe that footnote to playing Dungeons and Dragons.

D&D and grad school: Labor economics is all about mathematical models: functions that describe labor supply and demand, or which model the elasticity of the labor market, or combine all the factors (in a weighted relative, interacting fashion) making up labor decisions. This is probably true for all economics.

Anyway, a labor econ class has a lot to do with walking through these formulas. The prof puts it up on the screen and explains each factor, the math being done to or with it, and gives citations and articles backing up each number and variable. During one of these classes, I caught two mistakes (which were later called refinements in the article). A few of the students, including the pissy TA, were somewhat mystified.

How’d I do it? THACO and D&D formulas.

THACO is the most basic formula in Dungeons & Dragons. It stands for “To Hit Armor Class 0″ and it’s a complex enough math model to be the source of my greatest academic success (a footnote!) in grad school (before I quit, of course, I’m sure greater successes lay ahead had I stuck it out). So let’s dissect:

Armor Class 0 — that’s a person who has no armor, no defensive skills in combat. Armor Class -1 (AC-1) is a person who has the flu and is even easier to hit than usual. Adding armor, or taking theraflu improves, or adds points to AC, while drinking alcoholo or going to sleep in a combat situation decrease points.

To Hit — now we get into dice rolls and probability. If I’m a butch guy, with fighting skillz, I’m very likely to successfully hit a person with AC0. If we describe that as 75% likely, D&Ders will turn that into a dice roll: roll a 20 sided die (d20), and if I get 1 - 15 (75% of the possible rolls) I hit the person. Like the recipient’s AC, the TH can be altered, sickness can reduce my likelihood of landing the punch, steroids can increase if not the likelihood, the power of that punch. If I’m a spaghetti arm, I have THACO of 1, 1 chance in 20 of hitting the guy.
How simple, but rich, is that? We have a model that describes all combat interactions between two people. Labor economics models some basic forces: the likelihood of hiring, the likelihood of taking a job, and attaches a zillion factors to it. Having learned the basics of modeling behaviors, understand the different impacts of multiplication, division, addition and subtraction, and how to use coefficients as amplifying or diminishing forces, I had the basic tools for grad school economics.
D&D, along with typing, AP English, debate, and learning to program the TRS-80 were the most important things I got out of high school.

Nice cuts from  of the XO: 1) reduced chemicals in the battery; 2) low power needs of the machine (1/30 of most devices); 3) quick clip about the non-reflective screen (one slide below):

Interesting Microsoft move pre-reported by TechCrunch this morning:  Bill Gates will announce (or has announced) that (verified) students will have access to MSFT development tools for free.  In a program called Dreamspark, students can get the entire Visual Studio line, Expression, Windows Server, and Game Studio.  It’s a smart way to compete with open source, build the community of developers who work with and prefer Windows, and build brand loyalty at a formative phase of someone’s career.

Channel 8 did an interview with Bill Gates about this initiative which turned into talking about software.  Despite outsourcing and the commoditization of certain development skills, Gates is still quite bullish on the importance of programming:

[People can use MSFT dev tools to] build a career around or build fun software for themselves.

The skill of design, skill of knowing what good code looks like … will be around for the next couple decades.
There’s nothing more fun than thinking about software . . . software for the poor (there’s a lot more work that needs to be done), software to make jobs more interesting, software to help peofle design things in new ways … if you think about the sciences today they’re really driven by software … biology has so much information that it’s really software people who are gong to help find patterns and organize that information.

Roberto and Lizette Greco are plush toy designers (among other things) who designed a plush mascot for a ‘one laptop per child’ campaign’ (Un Computador por Nino, or UNPC) in Chile. The mascot is pretty cute, and UNPC even has a youTube group.

UNPC does not officially support or promote the XO, they are simply lobbying for one laptop per child as an educational initiative. So Pudu, the mascot, is carrying a less distinctive laptop.

They have a flickrstream about their children’s experience with the XO (the first pic does an evaluation of all the software they’ve installed)

They also highlighted two other piece of educational software: Squeakland, which is the inspiration for eToys; and Scratch, a programming language which looks like an interactive game/environment language. Scratch looked pretty complicated to me, but the Grecos say their kids (aged 7 and used really took to it:

I’m blogging about the XO because it feels like most interactive professionals are rushing to judgement (positive and negative) and missing an opportunity to dig into a rich design case study.

When I say “most interactive professionals”, I’m referring to the voices I’ve come across on blogs, twitter, and some searches. It’s not a scientific sample, by any means. However, the rush to harsh judgement, and the lack of any real in-depth looks, makes me suspicious. In our daily work, we spend hours and hours watching users look at slightly varying shades of color, or small pixel level adjustments to improve the performance of a page by .5%. We spend hours and hours speculating about what features the next OSX release might have, and then many more hours evaluating them. But, for the XO, it seems like we have it all worked out in 20 minutes or from the press coverage.

I wish there were people out there who were explicitly evaluating the XO against: 1) the educational approach driving the project; and 2) research addressing how kids approach computers for the first time.
The first point refers to constructivism, an educational theory which can be summarized crudely (to the point of coarse vulgarity) as kids grow cognitively by doing things rather than simply being taught. There’s too much to cover in a blog post, but even Papert’s summary is better than nothing:

The word constructionism is a mnemonic for two aspects of the theory of science education… From constructivist theories of psychology we take a view of learning as a reconstruction rather than as a transmission of knowledge. Then we extend the idea of manipulative materials to the idea that learning is most effective when part of an activity the learner experiences as [the construction of] a meaningful product. [italics mine]

The constructivist point of an XO, and there are other points (such as providing digital textbooks via the internet), is to get kids building and making things with a computer. I haven’t dug deep into the literature, but there are some places that look at the tools underlying this approach and how well they work: Life-Long Kindergarten, the Maine Laptop Initiative, and the robotics-in-school wave sparked by LEGO’s Mindstorms. Experiments like this are, by nature, hard to conduct. The lab is usually restricted to a single classroom, maybe a school district, or, at best, one state. I would love to hear designers and others talk about this.

(Interesting) sidenote: an individual’s early experience with computers seems to be a strong factor in whether people are inclined to like or put the hate on XO. As a kid, I took an 8th grade programming course on a TRS-80 and it created a lifelong fascination with math, science, generative design, computation, digital creativity. When I was eight or nine, I played with a lunar lander program (your craft is falling to earth and you can use direct or rotational thrust to land safely). It was a painfully slow computer (a phone handset was placed in large rubber holders to talk to the mainframe at CMU), but I spent an entire afternoon plugging numbers in, waiting three minutes for a response, recording the results, and backing into the rules driving the game (to say I was backing into the math would be an exaggeration, it wasn’t as formal as that). I was, in effect, “reverse engineering” and learning a complex system.

Slightly younger friends of mine had Commodore 64s with Turtle Art (the epitome of a constructivist software app, shown below) as kids. They also seem to be predisposed to liking the XO. For all of us, there is a sense that these constructivist moments were as valuable in forming our clearly fabulous minds and inspiring us to learn as any formal schooling we had.

Understanding, and actually looking at, the constructivist underpinnings and implementation of the XO seems to be missing from most design discussions of the tool.

The second point, how kids approach computers especially when they are largely undirected, is another area where I think western designers are missing an opportunity (or being sloppy). We’re all very well versed in how comparatively affluent consumers approach websites and other interactive experiences. We also have an idea of how supervised kids approach the web, but do we know anything about how kids learn them for the first time and alone? When we “laugh our asses” off at the operating system, what is it based on? How do we know that the OS is inappropriate?

Again, I don’t personally have a lot of data points, but one project that is repeatedly referenced is the Hole in the Wall, a program where unattended computers are made available to kids in India. From PBS:

From the slums of New Delhi to the coastal roads of Banda, hundreds of poor kids in India go online every day at free, outdoor computer kiosks installed in slums and rural villages to read news headlines, befriend cartoon figures, draw with digital paintbrushes and explore the possibilities of cyberspace.

There are no manuals, no adults to guide the kids, and it’s a Windows machine (originally an English version). The kids (who appear to skew older than the OLPC target by three years), teach it to themselves and each other.

There are a lot of design challenges against new assumptions in the XO, so the question for me is: why aren’t designers trying to learn from the XO or at least do a more informed critique of its design?

I’m a fan of the XO — the project, the goal, the educational ideas behind it. More than that, I’m fascinated by it.

I have a hard time thinking of what mass-market product has been launched since the PC that is more complex. And I don’t think there’s ever been a product launch as transparent as the XO’s. What a case study: a revolutionary piece of hardware, innovative open source software, designed for a market that may or may not welcome it and an incredibly broad audience.

I’ve got a Flickr photostream, with a bunch of screenshots, but wanted to capture some thoughts on the blog.

First, the project is enormous:

- $100 laptop (it’s now $200, and OLPC hopes to get it to $150)

- for children aged 6 - 12. I think this is the biggest challenge. This age range covers pre-literate kids up through pre-teens, playing simple games through programming.

- Integrated into school curriculum, appealing to government agencies

- be a substitute for textbooks (the swivel screen and glare-proof monitor support its use as a Kindle-like device. Textbooks are scarce in the US, and almost completely non-existent or out-of-date to the point of useless in many of the XO’s target markets.)

- EXPRESSIVE - stimulate the imagination (art, computation, narrative)

- APPROPRIATE - sturdy, stable, long battery life, outdoor use, theft-deterrence

- OPEN - the OS and software must be easy to develop, easy to adapt/upgrade, NOT dependent on another company’s development cycle or staff

Designing something for kids aged 6 - 12 is a massive challenge in and of itself. This challenge manifests itself immediately and viscerally in the keyboard:

(Click for larger image and comments on flickr.)

A snap, but fair, judgement to make is that boy, there sure are a lot of keys: quick keys, amplifier keys (CTRL, FN, etc.), the keyboard itself. To make matters trickier, some of the keys have three values assigned to them. Finally, there’s an inactive slider bar on the top and two types of input devices at the bottom (the middle is capacitive, the outer two resistive). So did they get it wrong? Is it, to quote one designer “a shining testament to the disastrous effects of theory-driven- and designer-driven-design”? Let’s look at what the keyboard needs to do:

- be useful to a 12 year old, who will word process, browse the web, play games, draw, and hopefullly program

- support languages with complex character systems and constructions (more than the US qwerty)

- provide quick key, shortcut usage that power-users expect (unless we think 3W kids can’t be power users)

So is the keyboard poorly designed? If so, is it cuz it’s theory-driven or because it strives to do too much? It still comes back to the age range 6 - 12 year olds. In the states, we can buy our kids different electronic devices at different ages, and the market is awash with chip-driven educational/entertainment devices. But this one has to do it all.

The second big design moment that peole confront in the first minutes with the device is the home screen.

This takes a few minutes to figure out … at least for all my interactive friends. The black border area is hidden from the user’s view, until the cursor is move to any of the four corners — it’s like Expose on OSX. The top black band contains quick links to system things like the network, or the home page, the bottom band contains links to all the applications on the machine. The left and right bands, and I like this, is a clipboard area. Anything the user adds to the clipboard is available here. That’s means a user can collect text, sounds, drawings, photos (from the camera) and have them available in eToyz, an authoring program on XO that has a lot of resemblance to HyperCard of old.

This is a busy home page, after the user has opened several applications. (Notice that the black border is missing, when the cursor is moved out of the corners the full screen is restored.) The ring contains all of the open applications in the form of graphic links, and right clicking allows the user to close the app without having to switch to it. What I love about this part of the interface is the way in which it graphically represents a machine whose RAM is full. Rather than a system resources message, the young user can see that things are pretty crowded and that they might need to balance/fix the situation.

There have been some questions around whether the XO should have a cheap version of Windows. As noted above, there was an early decision to go open-source. Part of the reason for open source is cost — there’s a lot of software and talent that can be leveraged with a Linux system. The other reason is that, with Linux, it’s possible to create a light-weight, clean interface like the one above.

I had no idea of the controversy surrounding the XO until I started twittering, flickring, and FBing my excitement at receiving mine in the mail. The range of reactions (and going-in assumptions) to it are wild:

- it’s a patronizing toy
- these kids need anything but a computer
- it should be windows and run like other computers
- it should be part of a guided curriculum
- it should work better without any teacher present
- it’s too complicated
- it’s too childish
- it’s too innovative and gratuitously so
- it’s not innovative enough
- WTF is this? and this? and this? and THIS?

One designer at Behavior, wrote:

We installed the OLPC Sugar emulator (via VMWare) here at Behavior and laughed our asses at it for about 20 minutes. It really is a shining testament to the disastrous effects of theory-driven- and designer-driven-design. You’re not missing much, but if you need to satisfy you curiosity and/or have a few lulz you should try it.

There is an almost gleeful delight in trashing the XO.  I’ve shown my XO to dozens of people (most of them willing observers, though I have been a little aggressive with strangers in the subway) and a significant number are ready to pronounce it an arrogant failure after moments, or, to be fair, minutes of interaction.

Perhaps that’s the right litmus — aren’t interactive designers taught that we have mere seconds to win an audience? On the other hand, don’t games, handhelds, new OSes, and even the iPhone require several minutes, or even hours of acclimation?

Many of the criticisms above are against the concept itself, some of them are about the implementation. The Economist ran a review that seems pretty even-handed. The reviewer respects the concept:

Indeed, Mr Negroponte’s vision was brilliant. He planned to blanket the developing world with tens of millions of $100 laptops for kids. The low cost would come from a tripartite “perfect storm�. First, economies of scale: sales would be directly to governments, who could only buy quantities above 1m. Second, the machines would bypass Intel’s processors and Microsoft’s software in favour of open-source stuff. Third, commodity parts would keep the price low.

He also respects the innovations (the screen, flash drive, element-resistance, PC to tablet factor). His final reasons for criticizing it are that: 1) they were naive about dealing with educational ministries; 2) naive about distribution and how quickly they would achieve scale; 3) that the final implementation stinks (his machine crashed frequently and he found the OS impenetrable); 4) that the folks were arrogant to the point of not listening to any criticisms legitimate or otherwise.

OLPC folks respond on their wiki, in a tone that is as respectful as the Economist’s critique. Their main points can be summarized: 1) a few facts in the article are wrong; 2) yes, the flaws exist, but these are flaws (crashing, slow load times) that are true to all computers, so the question is ‘by what standards are these flaws too severe?; 3) some bigger picture factors that may be overlooked, such as the rhythms of educational software, the tradeoffs needed in such an ambitious design. The closing of the wiki post is great. It includes the last, praising, line of the Economist’s review, and the OLPC response:

“Ultimately the OLPC initiative will be remembered less for what it produced than the products it spawned. The initiative is like running the four-minute mile: no one could do it, until someone actually did it. Then many people did.�

Not a bad obituary, but our work is not yet finished.

I bought mine, partially cuz I love the idea, and partially out of professional interest in the design.  Initial thoughts from the first day of play in a Flickr photoset.