Here's a crazy idea. Let's teach Ruby in intro programming classes instead of C, so as not to frighten away 90% of the population with pointer dereferencing and segfaults. You should teach C eventually to make them appreciate high level languages, but just not right away.
We teach Shakespeare and Calculus in high school, though those are certainly not the easiest choices in their respective fields. For an introductory programming class, I agree that Python is a great choice. However for seniors in a computer science class, C should be a fairly important part of the curriculum. There are some fundamentals of computer science that are best grasped through experimentation with C (pointers, memory, etc).
Note that several years ago the AP Computer Science exam got switched from C++ to Java. I personally support this change because most high schoolers are not mature enough to fully grasp the complexities of C++. Java is a bit gentler (Python would be better). However I don't see why C can't be the Shakespeare of Computer Science in high school. Everyone hates studying it, but it's probably better for the student in the long run.
On a slightly related note, I don't hear of too many efforts to make Calculus more accessible to girls or Shakespeare more accessible to men. This whole women in programming thing is an interesting phenomena. Don't get me wrong, efforts like this are great, and I've even participated in a few myself. But let's be clear about the source of the 'problem'. Deeply embedded society pressures and bias in early education are two major factors working against women in programming. By the time they've gotten to high school we've already lost quite a few for good. Let's also be careful not to attribute every inequality in sight to socialization. That's a nice PC answer, but unlikely.
By the time students get to Shakespeare and calculus, they've been reading books critically and doing math in school for at least a half dozen years. If they have six years of CS/programming education, too, then I agree that they would be well-served by a course in C.
There are a whole hell of a lot of concepts in computer science that I think are both more useful and far more educational than understanding how to do pointer math and learning how to carefully allocate and free memory for your data.
Actually, they haven't. For the most part, they've never read anything critically (or even really read anything period), and the crank-turning and hoo-ha that goes in math class certainly isn't mathematics.
So by the time they hit Calculus and Shakespeare, they're in year 11 of their time in a velvet jail and have learned to very adequately mimic the bizarre squawking the guards insist is requisite for passage to The Island.
What they have not done is mathematics or reading in any real sense.
I had to read Romeo and Juliet my Freshman year in high school. My understanding is that R&J is a fairly common Freshman text. Your point about Calculus doesn't quite fit either because you don't need an extensive foundation to start programming. That's why we see a lot more 14 year olds programming than doing Calculus. You're a genius if you're doing Calculus at age 14, but I bet a lot of people who read this site started programming that early or earlier. I'm sure we've got our fair share of accelerated math people as well, but they jut are not as common.
My point is that I feel the things that C illustrates and the style of programming it enforces are not that important to learn if you are just starting to program. If you are approaching it from an academic math/CS standpoint, you'd be better off teaching something much simpler like Scheme, and if you're approaching it from a building-cool-stuff sort of hands-on standpoint, I suggest you would be better off teaching Python, Ruby, or one of many other languages with less drudge work (a lot of syntax cruft, manual memory allocation, not very abstract) to do than C.
There may be value to learning C, but I think it's not a good return on investment of your time until you understand many other foundational ideas about programming and CS.
Sure, let's count learning to read, write, and spell correctly.... oh wait, you use that for programming too. ;-)
Also, you should probably revise your thinking so "English" isn't just everything that has to do with books. It's critical theory, complex analysis, and a bunch of other things. The stuff most people do in 1st through 6th grade is to English as learning how to type is to programming.
I agree that you shouldn't make the classes too easy, but you have to get something out of the extra difficulty. The reason that we should teach Python instead of C is the reason why we teach Arabic numerals instead of Roman numerals. Roman numerals are more difficult, but you get no insight out of this time spent learning Roman numerals. It is extra human created tedious difficulty.
You get a much better understanding of pointers and memory from learning a machine language. In C you'd be spending your time on syntactic issues like & and * in expressions and in types. This is what makes pointers difficult. In assembly there's only STORE x INTO y and LOAD x FROM y.
Excellent observation about understanding pointers; I never really had any difficulty understanding pointers in C, but I had already programmed in assembly language.
Shakespeare on the other hand, is still opaque to me.
"We teach Shakespeare and Calculus in high school, though those are certainly not the easiest choices in their respective fields."
I'm not sure that's the best argument. In my sophomore year in high school, we covered Romeo and Juliet. At that point, it is fair to say that I had reading skills well above average. I could read original texts from the 19th, 18th, and even to a limited extent the 17th centuries and understand them just fine.
Shakespeare I could make the plot out, but little more. The footnotes assured us that all kinds of funny jokes and puns were being made, but we pretty much had to take that on faith. Seeing the plays helps a bit more because you get the additional channel of seeing real people, but even then we're taking the humor and puns on faith. I suspect that if someone from Shakespeare's time could see a modern production they would consider it atrocious (once they got past the superior delivery technology); places my footnotes assured me were hilarious were usually delivered straight in the videos I saw, as if the actors didn't realize they were making a pun and it was supposed to be funny. Most likely because, well, they didn't realize it.
I question whether we teach Shakespeare as anything other than "the traditional curriculum", and with a desperate-but-futile hope that this will somehow inculcate a desire to be exposed to more "good culture" in the students... fresh on the heels of teaching the students that "good culture" is written in an incomprehensible vaguely-English language which is completely inaccessible without footnotes which assure us that it's good, really, in plain contradiction to the evidence of their eyes. The shock of the literati at the failure of this plan is amusing, but still sort of odd.
I see a lot of parallels to this with teaching high school students C. We're going to teach them a nearly-40-year-old language (might as well be from the 1500s) written to run on computers the size of rooms with less power than the cell phones the students have, with copious verbiage telling the students no, really, this is computing! despite the fact their eyes will tell them their programs resemble none of the programs they actually use? Gosh, how could that ever fail?!
Save it for college.
Just to highlight one of my points, I think one of the worst things you can do is contradict what the students are seeing. They see an archaic, musty language that can barely be understood, that resembles nothing they see in current use, but you assure them that this is the height of Western culture and they'll be better for understanding it. All you do when you do that is blow your precious credibility. I'm sure one or two people will pop up here and insist that Shakespeare changed their lives in school. Maybe one of them is even actually telling the truth, and not just defending the status quo because it's the status quo. But nevertheless, statistically, it's a terrible approach; it can't convert anyone who isn't already very interested and it turns away the vast majority of any who may be interested. I'm not saying we should be feeding our students educational candy, but let's teach things in High School that, when we claim they are awesome, might actually be able to stand up to that point, within the student's actual reference frame and not the "desired" one, which only exists in the educator's head. Dickens, or Melville, Swift perhaps, or perhaps Python or Processing. Jumping to C or Shakespeare isn't done to help the students, it's done to gratify the teachers, and it fails in entirely predictable ways.
A world full of great literature and great computer languages, and we insist on homing in on the very things most likely to turn off the students forever. Then we wonder how schools manage to turn out so many students who hate "learning". Gosh, how could that be?
Edit: One final note. I often get a lot of shock when I question Shakespeare's value. But I think "question dogma" is part of the Hacker's credos. Is Shakespeare holy because you've worked with exposing people to it and seen it change their lives (not just your own, either) and because you yourself love it, or is it holy to you because people have been telling you all your life how wonderful it is to the point that saying "Shakespeare is great!" has become a social signaling mechanism to indicate how cultured you are? I freely acknowledge that there are people who truly do love it... I just don't think they are that numerous.
> The footnotes assured us that all kinds of funny jokes and puns were being made,
The worst part about those high school Shakespeare texts was that for the humour that was blatantly sexual, they tried to find some contrived explanation to present instead.
For example take a passage: "She that’s a maid now, and laughs at my departure, / shall not be a maid long, unless things be cut shorter."
This is a reference to the "maids" being played by male actors. The text had some contrived explanation about if the play were shorter in length, when really the line is clearly a joke about CUTTING OFF A PENIS.
When you miss the fact that Shakespeare is just full of eighth grade schoolboy humour, then you actually miss a lot of the point.
I agree with your sentiment, however, I'd like to pick an nit:
"I see a lot of parallels to this with teaching high school students C. We're going to teach them a nearly-40-year-old language (might as well be from the 1500s)"
The age of the language isn't so much the issue; Lisp is quite a few years older than C and ML came on the scene the same year C did. I think Smalltalk is just about as old as C. Yet teaching Lisp or Smalltalk (using Squeak) would be a much better idea than teaching C as a first language.
Teaching C as a first language would be akin to having only stick-shift cars in Driver's Ed. Now, I happen to think it's important to learn to drive a stick, but first you start the kiddies off with automatics so they get used to other, more important aspects of driving and then towards the end you teach them how to drive a stick.
You are in effect making the argument that "tough learning" has no place in the curriculum. It's absolutely crucial that students are confronted with learning curves that are discontinuous compared to what they have previously faced. The other road leads to common denominator mediocrity. The real problem is that you cannot set an individual curve for students because of the need to teach them en masse.
The most significant experiences in my life have been about overcoming learning curves that seemed insurmountable. Two experiences stand out for me, one with mathematics, the other with literature. Even one experience like that can last you a lifetime. It can lead to a career and a lifelong passion.
My epiphany in mathematics came at age 16, the one in literature much later at age 22. The reason both came so relatively late was that my arrival was by circuitous paths of self-study. Unfortunately I never had a truly inspiring teacher.
I'll say a little (or a lot) about the epiphany I had with literature since you seem to have a particular problem with canonical literature. Until my teens, I read more than most kids but not very much and virtually nothing but pulp fantasy and science fiction. When I became interested in programming and later mathematics I read a lot of technical books in those areas. But anything vaguely literary with a challenge accompanied by no immediate reward was a complete turn-off.
Poetry in particular was incomprehensible to me. I could not understand any of the poetry we covered in school, and I could not understand how anyone could understand it, much less appreciate it. In short, my attitude was the same as yours: All these people who pretend to love it must be pretentious pricks all hoodwinked by the Emperor's New Clothes.
I eventually stopped reading almost any fantasy and science fiction and stuck to nonfiction, mostly technical material. But when I was age 22 I picked up a copy of Pullman's The Golden Compass and read it through in one sitting. I thought (and still think) it was the best work of fantasy I had read.
I looked the author up online and found that elements of the books had been inspired by Milton's Paradise Lost. I had heard of this work but never read it. I swifly found a free copy on Gutenberg and discovered to my dismay that it was a narrative poem. A fucking poem! But this writer guy says it's the best poem in the English language, so there must be something to it. I tried to read the first page but was defeated by Milton's tortuous syntax and mythological allusions again and again.
I gave up and told myself that poetry was simply not for me. But at least I had now convinced myself that this kind of old and difficult poetry was not an elaborate inside joke or some sort of hoax.
A few months later, after rereading The Golden Compass, I decided to give it another shot and this time ordered a handsome printed copy that incidentally came with an introduction and a chapter by chapter commentary by Pullman. This time I forced myself to read everything aloud and that was the trick that let me crack it.
Wow! Poetry should be read out loud!
For some stupid reason this had never really occurred to me. Soon I had gone through and understood the first several pages. Suddenly I could read and understand the poem with surprising ease. It was the most sudden zero-to-hero experience I'd had in my life. I was covered in goosebumps and tears were welling up in my eyes. I was forced to stand up and pace around the room for several minutes to calm myself down. Finally I sat down and spent several hours memorizing the first couple of thousand lines, despite my generally awful memory for such things. After that experience I found I could read through a book of the poem at the rate of one per night. Not all the allusions made perfect sense on a first pass, but now that I had learned to sound out the words the syntax was no longer an obstacle at all.
The poem possessed me so completely that I eventually memorized the first two books of the poem almost effortlessly, possessing the poem in turn. And when I say memorize I don't mean by rote but by the kind of compulsive rereading and rethinking that only passion can motivate.
Anyway, that was my epiphany. I decided to try Shakespeare next and bought a copy of Midsummer Night's Dream. That was a lot harder! The word to word, sentence to sentence writing of Shakespeare is so far beyond the likes of Marlowe, Milton or Melville in terms of subtlety and modulation. All of those guys are at their best when they write majestic rhetoric blasting on all cylinders. That's not to say they lack subtlety in other areas; both Milton and Melville are incredibly deep.
Anyway, I slogged my way through Midsummer Night's Dream. It was fun but it didn't deliver Milton's spine-tingling bombast in the same way. It was subtle the way the world is subtle. If you read it superficially you are likely to leave with little else but the plot and a general but vague sense of the characters. Shakespeare's plots are expendable and the strength of his characters are in their variety, distinction and depth, so this way of reading him does him poor justice.
I put aside Shakespeare for another year or so. My next attempt was with Hamlet and that was what cracked Shakespeare for me. It had all the gratifying high rhetoric I so loved about Milton but it also had complexity in subtlety and depth to last you an infinite number of rereadings. From there I eventually went on to read all of his works. Unlike Paradise Lost, several of them took me several rereadings to love. They rarely give away their secrets on the first stroll through.
Holy crap was that ever long. Anyway, I present this as antidote to your criticism. What this and a few similar experiences has taught me is to never assume that something is a worthless sham simply because you don't get it right away.
"You are in effect making the argument that "tough learning" has no place in the curriculum.... you seem to have a particular problem with canonical literature."
No, which is why I explicitly mentioned Swift, Dickens, and Melville. We read both Gulliver's Travels and A Modest Proposal in school (nowhere near a complete list, of course), and even if it wasn't everybody's "favorite literature ever", at least it was comprehensible to people. You could send students home with a reading assignment, and they might actually do it, and then when they come to class the next day, they can discuss the text, and find out the deeper meanings, rather than having to come to class to find out just what the hell it was they just read. This is not a valuable use of literature time. Also, as I said, I read quite a bit of the "canon" of my own free will and could read Ye Olde Texte far better than most people of my age. More than once I was given a school assignment to read a book I had already read.
Your story is not a counter to my point; it proves my point. You had to be interested and bring a lot of effort to the task of reading those things. This makes them grotesquely unsuitable to be using them in high school where those things are not generally available. That you found it rewarding is great. I too have done that, both in the computer domain and the literature domain. But an awareness of what it took is precisely why I say it's a terrible approach to general education.
I never said it was a worthless sham. I said it wasn't worth teaching in high school, and that it should be left for college. Based on the age you had your experiences, you hardly provide a counterexample for that. It would be way better to simply allude to Shakespeare's genius and let motivated people find it, than spend valuable time jamming it down students throats only for them to see an incomprehensible mass that they are never again motivated to look into. There's no way school could every have given you that experience. (Maybe a really motivated teacher, but even then it would have been a personal interaction, not a schooling thing.)
Basically, my position is that you have to educate the students you have, not the students you wish you had. And, if we are going to build our education system around teaching masses of children, building our core curricula around something that will maybe appeal to one or two kids per year, and then only several years later, is such a waste when we could be teaching people things that might actually encourage rather than discourage the pursuit of further learning.
So often the choice is presented as "Shakespeare or nothing", or "only incomphrehensible literature or nothing". You sort of have those memes in your response too, when you read into my post that I'm against literature. But when I spell that out, it's obvious that there are tons of other choices. There's thousands of other school-worthy candidates that won't have the side effect of firmly squashing all future interest in "literature".
(There's also a frequently unrecognized assumption that we have infinite time for students to learn in, and therefore it is sufficient to establish that something is "good", and therefore we then put it in the curriculum. That's not even close to true. Our time as students is terribly limited compared to what there is to learn, and it's getting worse every year. The question is not "Is Shakespeare worth learning?". Duh. Of course it is. But then, there's enough "stuff worth learning" to fill your lifetime and beyond. The question is, is it the best thing to learn? That's what I'm challenging. Given the enormous liabilities that teaching Shakespeare... oh, and by the way, C, since the parallel continues to hold(!)... brings to the table, including the very-real risk of completely destroying any interest in literature/programming that may have been present, why shouldn't we teach the things that don't carry that risk?)
I often get a lot of shock when I question Shakespeare's value.
I'm not shocked at all; your viewpoint is actually fairly common. However, your points about Shakespeare are evidence that you really never fully understood why we study Shakespeare. I'm not saying it's your fault. Indeed, more than likely it's because of the sorry state of English education in high school. I'll do my best to explain here.
I'll start by acknowledging that enlightenment/revelation is a significant piece to it. Explaining this to someone who hasn't studied English in depth is like trying to explain why Lisp is so beautiful to a Blub programmer. It's possible, but it requires a writer of Graham's caliber and I fear I'm not yet skilled enough for the task. For now you'll just have to take it on my word that such enlightenment exists. From personal experience, I know it is possible to achieve it on the high school level and the impact is long lasting. Without Powers, Faulkner, Shakespeare, Achebe, Hemingway, there's no way I could have achieved the level of success I have in Computer Science. If I hadn't gotten the hard stuff in high school, I wouldn't have bothered to continue to study it on the side in college.
My personal anecdote is not my main point, however. It's a single data point with no evidence, and I present it in the hope that it be useful in some way. The real reason why we study Shakespeare in high school is not for the enlightenment that results from his words. It's because his influence on the past and present of literature is overwhelming. Without Shakespeare you wouldn't have Dickens, Coleridge, Melville, etc. All of the works of these people contain Shakespeare's influence. Hell, he helped change the English language itself. Indeed, there's a great parallel to C here. Do you really think we teach C so that people will program in it? No! We teach C because the concepts contained in the language underly so many different topic areas in Computer Science. Who cares about the enlightenment from understanding Hamlet, pointer arithmetic, memory management, and so on! What's really important is understanding thees things so you can understand the modern, cutting-edge stuff.
It's fine if you think the culture stuff is bullshit. But if you like Dickens and Melville then you should understand Shakespeare. And damn, if more people understood C, the most popular open source browser wouldn't be memory leak hell. ;-)
I'll close with a quote from Spolsky:
Today, to work on CityDesk, I need to know Visual Basic, COM, ATL, C++, InnoSetup, Internet Explorer internals, regular expressions, DOM, HTML, CSS, and XML. All high level tools compared to the old K&R stuff, but I still have to know the K&R stuff or I'm toast.
Ten years ago, we might have imagined that new programming paradigms would have made programming easier by now. Indeed, the abstractions we've created over the years do allow us to deal with new orders of complexity in software development that we didn't have to deal with ten or fifteen years ago, like GUI programming and network programming. And while these great tools, like modern OO forms-based languages, let us get a lot of work done incredibly quickly, suddenly one day we need to figure out a problem where the abstraction leaked, and it takes 2 weeks. And when you need to hire a programmer to do mostly VB programming, it's not good enough to hire a VB programmer, because they will get completely stuck in tar every time the VB abstraction leaks.
And yes, I know you might advocate simply waiting until college to teach this. But want to know what's harder than teaching a 17 year old pointers? Teaching 21 year old hotshot Java programmer pointers.
"We teach Shakespeare and Calculus in high school, though those are certainly not the easiest choices in their respective fields."
I think a better analogy is Driver's Ed. We don't throw kids into stick-shift cars in Driver's Ed (at least not at first) we have them drive automatics so they can learn the more important aspects of driving. Then, later on towards the end of the course (well they used to anyway) we have them learn to drive a stick after they've mastered the basics of driving.
Now, personally, I think it's important that the students learn how to drive a stick. But not at first. Same with choice of programming language: I think it's important that the student have some exposure to pointers, dereferencing and even some memory management, but not at first. Let them learn the more important aspects of programming first and when they've become comfortable with them then transition to that other stuff that I'll call "mechanics". Oh, and not every highschool student needs to learn the mechanics, just those who think they want to go onto software engineering.
It's not a crazy idea at all, nor is it even a new idea (see BASIC, Pascal, etc.), but it is not appropriate in all cases.
I would never recommend trying to teach C to a group of high school kids that were otherwise indifferent to computers (eg not curious tech-nerds).
Teaching methods and knowing your audience are important, too like recognizing that most of your class is bilingual and making an analogy that would be useful and relevant to them. I once explained Computer Science in a nutshell to my sister --a pianist, one of those self-proclaimed not-good-at-math people, but someone I know and communicate with very well-- in about 45 minutes; starting with binary logic and circuits and putting it all together with and ending with high level languages and the internet. She didn't retain it all of course but I pointed her at the Head First HTML/CSS book and she picked it up no problem.
I think they should learn something like Z80 assembler. Get them used to the idea that computation on a von Neumann architecture occurs within a one-dimensional array of fixed-width, integer-addressed memory cells. When they are used to that, they can move onto things like C.
A perfect geek toy for kids would be, to me, something about the size and form factor of a cellphone with a QWERTY keyboard. There'd be a switch on the front such that in one position it functions just like a graphing calculator, and in the other position it behaves more like a KIM-1 computer, letting you craft machine language routines for it directly. There'd be a low-level monitor of sorts that would let you examine memory contents, assign breakpoints, etc. and maybe some simple I/O routines to load and store programs from the attached flash storage. Otherwise it's a 1970s era hobbyist computer, except it has loads of memory, a sweet color display, an ARM CPU and you can fit it in your pocket.
I think traditionally, it is considered better to learn it the hardway first. It teaches you more about the progression of thought through which innovation leads and provides a better understanding of the content.
That's why we learn Reimann Sums before we learn how to do the little tricks in calculus.
There has also been a lot of discussion lately about how higher level languages are "dumming down" programmers. Colleges who have switched to Java for example are producing graduates that are having more difficult times in the "real world."
Which is ironic because colleges switched to Java because Java is used in the "real world".
Of course, Java with all its class-and-interface scaffolding adds too much accidental complexity too. It's not a matter of higher-level, it's a matter of verbose. Scheme is a high-level language but it's still "the hard way" in a very important sense.
"it is considered better to learn it the hardway[sic] first"
Blatantly false. For a certain narrow class of people, this is the best way to learn. There are hundreds of other ways, and some of the people who have a different preferred style would make great programmers.
"Colleges who have switched to Java for example are producing graduates that are having more difficult times in the "real world.""
Also blatantly false. CS programs have always churned out a lot of mediocre programmers, a handful of good ones, and a few brilliant ones. Only the language has changed.
Absolutely. My high school taught C++ and it convinced me to not major in computer science. I was hacking on Perl and Visual Basic at home and couldn't imagine four years of pointers, segfaults, getters, and setters. I just wanted to have fun.
The key to this (lovely) story, IMHO, is the iterative nature of the lesson. Each step was very small, built on the previous one, and resulted in clearly visible changes to the program. For novices, this is absolutely vital- especially in a classroom environment like this. Now, re: the appropriateness of Ruby's somewhat-odd syntax for this lesson- I agree that there's definitely some weird arbitrary-seeming stuff in Ruby ("darkred" vs "dark red", etc.). But, let's face it- ALL languages have a few weird syntax things here and there, especially from the standpoint of a novice programmer. It's just part of the way that computers and programming languages work, and anybody who wants to write code is going to have to learn to deal with it. From my experience teaching novice programmers, however, I can tell you that a lot depends on HOW
The problem is that, with most languages and in most intro lessons that I've seen, the student is exposed to all of them all at once without much context. I feel like the lesson described in this article did a great job of introducing the students to these language "features" one or two at a time and in a relatively sensible order. Eg, the students first got used to the idea that the computer was picky about upper- vs. lower-case letters. Next, they got their heads around the idea that attributes often involve a weird sequence of characters ("=>"). Then, they were exposed to the "identifiers-can't-have-spaces" rule. Etc., etc., etc. By the time they were exposed to each new oddity of the syntax, they'd had a chance to absorb the one that had gone before.
I'm amazed as well that it went that well.
And its really great to hear that they weren't put off, lets hope the coming generation can help improve the genre ratio within programming workplaces.
I like you approach, and choosing shoes probably was a good choice. Instant gratification should be #1 goal in classes like this, its all about building excitement, which it seems you nailed.
"Another girl dug into the documentation she had seen me use, and figured out how to use the color picker to let the user select a color for the line!"
I'm impressed. Figuring out manuals can at times be hard even for us who use them on a day-to-day basis.
When it comes to pedagogy, the sigils and syntax of Ruby have always bothered me. The author notes the girls having trouble with dark red versus darkred, or getting the hash syntax right. I imagine a number of them wondered why "center" and "xx-large" went inside quotation marks, while saddlebrown did not. Furthermore, special string values like "center" and "xx-large" somewhat undermine the teacher's assertion that quotations demarcate "a sort of free zone for text."
Are there better languages or tools for this task? Has anyone here tried to teach programming to non-programmers using a visual language like Scratch or a simpler language like Scheme or Basic? How did it go? Shoes is an incredible platform, I just wish the underlying language had a more immediately comprehensible syntax. (PLT-Shoes, anyone?)
"Furthermore, special string values like "center" and "xx-large" somewhat undermine the teacher's assertion that quotations demarcate "a sort of free zone for text.""
I was delineating two modes of failure. The first, and most obvious, is when your program won't run at all and you get a syntax error. That's what happens when you misspell something outside the quotes.
But if you misspell something inside the quotes, you get the second mode, in which your program runs but you get incorrect behavior.
For us this isn't a huge difference; they're both bugs. But for the girls it was huge. If their program won't run, it's busted, but if it runs but isn't quite right, well, that's what Microsoft charges $$$$ for.
The average high-school girl might not be able to instantly see the elegance of the idea that functions, programs, data structures, numbers, etc. are all interchangeable and that it all can be represented recursively with lists all the way down. (High school girls who have read Hofstadter, maybe.)
They might be able to understand "later, rinse, repeat", or making a cake by making three layers, and then making the frosting, where each becomes a separate modular and reusable problem.
His point isn't that Scheme's syntax is elegant, but that the syntax is simple.
There's no need to worry about the precedence of operators, or indentation, or why you use one character to separate function parameters and variable declarations, but you use another character to separate the arguments of a "for" statement, which happens to be the same character you need to use at the end of statements (but as a terminator, not a separator), but don't forget that control structures are not separate statements even though they're on separate lines, etc.
Each of these points is an obstacle between the beginning programmer's intent and a running program. Scheme's almost total lack of syntax gets a beginning programmer started quicker.
That might seem simpler to you, but you have the kind of mind I was talking about:)
To the average person, to make a toy car, it's straightforward to take a box, put on two axles, then put on two wheels. You could do all that with just toothpicks, but it's not really simpler in practice.
Totally awesome. It's so cool to hear such great success stories about teaching programming, especially to those who are unlikely to take it up otherwise. I wish I could upvote this 100 times.
Instead of just having the program files on USB memory stick could you have a had a custom bootable linux environment instead - then getting the environment set up would just be a case of plugging in to a PC and booting up. Sometimes getting to the point where you start programming can be hard.
I'd say it's one of the biggest barriers to entry out there. Getting a dev environment set up for anything other than hacking a Wordpress template is non-trivial for most.
I use the stuff at PortableApps.com quite often with consulting clients who need Firefox/Audacity/Gimp in a hurry. Could there be a portable Ruby/Python/Django dev setup that could get a novice going quickly without having to install a bunch of files?
I work on Debian Live, a framework for building customised live systems (we also make "official" live images as a special case). Something like this is easy to make and maintain using it.
