August 26, 2003

Open Courseware

One reason that MIT is a global treasure:

Wired 11.09: MIT Everyware: Lam Vi Quoc negotiates his scooter through Ho Chi Minh City's relentless stream of pedal traffic and hangs a right down a crowded alley. He climbs the steep wooden stairs of the tiny house he shares with nine family members, passing by his mother, who is stooped on the floor of the second level preparing lunch. He ascends another set of even steeper steps to the third level and settles on a stool at a small desk, pushing aside the rolled-up mat he sleeps on with one of his brothers. To the smell of a chicken roasting on a grill in the alley and the clang of the next-door neighbor's metalworking operation, Lam turns on his Pentium 4 PC, and soon the screen displays Lecture 2 of Laboratory in Software Engineering, a course taught each semester on the campus of the Massachusetts Institute of Technology. "Here," he says, pointing at the screen. "This is where I got the idea to use decoupling as a way of integrating two programs."

In a huge brick house that Evan Hoff shares with three other guys in Nashville, the 20-year-old brings up the MIT Web site and connects to the same material Lam is studying halfway around the world. "This is the lecture on data abstraction," Hoff explains. "I went over this in community college, but that class only took it so far. This teaches you about the three different specification conditions, the things you put in documentation to let future programmers know how to use it. In community college we covered only two of them."

When MIT announced to the world in April 2001 that it would be posting the content of some 2,000 classes on the Web, it hoped the program - dubbed OpenCourseWare - would spur a worldwide movement among educators to share knowledge and improve teaching methods. No institution of higher learning had ever proposed anything as revolutionary, or as daunting. MIT would make everything, from video lectures and class notes to tests and course outlines, available to any joker with a browser. The academic world was shocked by MIT's audacity - and skeptical of the experiment. At a time when most enterprises were racing to profit from the Internet and universities were peddling every conceivable variant of distance learning, here was the pinnacle of technology and science education ready to give it away. Not the degrees, which now cost about $41,000 a year, but the content. No registration required.

"It's a profoundly simple idea that was not intuitive," recalls Anne Margulies, the former Harvard assistant provost and executive director of information systems who was hired to be OpenCourseWare's executive director. "At the time, the world was clamping down on information, limiting it to those who could pay for it." Soon foundation money was gushing in to support the initiative. MIT earned the distinction as the only university forward-thinking enough to open-source itself. To test the concept, the university posted 50 courses last year...

Posted by DeLong at August 26, 2003 03:02 PM | TrackBack


Is there a "homo economicus" rationale for the MIT program, perhaps in terms of prestige or global recognition? Does economics have anything useful to say about this particular model of education? Inquring minds want to know...

Posted by: Tom Slee on August 26, 2003 04:33 PM

I was thinking of sending the link to the introductory macro class to Donald Luskin and friends so that they might avoid embarrasing themselves in the future. Alas, the syllabus appears to be too advanced for that purpose. Perhaps there is a community college level course somewhere...

Posted by: L.K. on August 26, 2003 04:38 PM

So far OCW looks like it's stuff you'd get on course web pages, with the exception of Gilbert Strang's excellent linear algebra lectures. Berkeley has many video lectures on biology, chemistry, and computer science online, for free, even if their stated intent isn't to share with the world. (anyone really interested will have to browse around a bit to figure out where).

There's also some stuff at and elsewhere. I watched or listened to lectures from 3 different universities to study for my intro bio class.

Posted by: Shai on August 26, 2003 05:40 PM

Not to detract from the positive aspects of MIT making their course materials available to the world, especially when other universities (like the little liberal arts school two miles up the Charles river :-) are closing as much as possible (could IP lawsuits be far behind?), but I think that the MIT learning experience is not completely "onlineable."

Going to class gets you about 5% of the learning; reading the materials, another 5%. The remaining 90% come from solving the problem sets, doing the assignments, and cooperating/competing with the other nerds. The incentives for solving said problem sets/creative assignments are difficult to replicate over the web.

(There are also issues of certification, of learning to tie your sleep cycles to problem set deadlines, and building up the immune system by eating in Walker Memorial.)

Posted by: Jose Silva on August 26, 2003 07:44 PM

I'll just correct myself: It seems that a few more OCW courses do have video lectures now.

If anyone is interested in lecture video and audio on the web just email and ill copy and paste all the favorites I have. (mostly computer science, but some chem, bio, astronomy, philosophy, etc). Some are unofficial and I don't want to interfere with ongoing classes by directly inserting links into the blogosphere...

Posted by: Shai on August 26, 2003 07:45 PM

I agree with Jose, motivation is a serious problem. But I think the 5+5+90% breakdown is pretty arbitrary.

I think it depends on the type of course. In computer science you can get by pretty easily by just going to lecture and doing the reading. If I'm stuck on a problem set I just send a message over to the course newsgroup, but I could just as easily get an answer from sci.math or by google, etc. If I'm programming I can usually solve my problem by debugging and googling. If I think I'm done I submit the problem set and a program verifies the correct output. done.

However, I did take a course in "European History from 1300-1900" that wouldn't translate at all because 90% of the course is reading hundreds of pages of primary text and discussing it in tutorial.

But in intro psych I never once talked to the prof or any TA's. Tutorials were a message board but as far as I could tell only a minority posted out of the 1500 in the class. Studying lecture notes I made on powerpoint slides and reading Gleitman et al was enough to get a B+.

I enjoy the aesthetic experience of "being there", especially chatting with the profs after class, but I think you can get a lot out of reading a text, watching lectures, making a lot of notes, and doing problem sets. But it depends, I suppose.

Posted by: Shai on August 26, 2003 08:38 PM

I've been distinctly underwhelmed by the MIT OCW course I've looked at, mostly advanced physics. I hope this is simply teething problems, but what I have seen is pretty much nothing. There an intro to the course, a week by week list of material to be studied, the list of textbooks, and problem sets for each week (but clicking on the answer sets gives a broken link).
Now maybe the pool of people who want to study advanced physics on their own is pretty damn small, but this is pathetic --- better to do nothing than raise people's hopes and waste their time.
At the very least I'd like to see the answers to the problems so that those of us following at home know we are doing things correctly, and I really don't see why they can't also have the video lectures.

Posted by: Maynard Handley on August 26, 2003 09:13 PM

Jose Silva writes: " I think that the MIT learning experience is not completely "onlineable."

"Going to class gets you about 5% of the learning; reading the materials, another 5%. The remaining 90% come from solving the problem sets, doing the assignments, and cooperating/competing with the other nerds. The incentives for solving said problem sets/creative assignments are difficult to replicate over the web."

This is true, though I'd quibble over the numbers.

But for most people, the MIT experience isn't an option at all, so they make do with whatever school happens to fit their circumstance. Most people won't get anywhere near something approximating an "MIT experience".

MIT's open courseware can do quite a bit to help a motivated student bolster an education from the not-so-great institution he or she happens to be stuck attending.

Another consideration is that if you tackle the MIT-level homework assignments or projects, and complete them, that could be a significant boost of self confidence for someone attending a third rate school.

Posted by: Jon H on August 27, 2003 09:05 AM

It's a great idea. And I agree that the percentages vary by types of classes and types of learners. Practicing problem sets is important for the maths and harder sciences (and economics) but less important for other things. Most knowledge is available in books - then you need someone to discus your problems, thoughts and questions with - then you go back to the books - and so on.

The best thing I got at University was the library card. The second best thing I got was access to Profs. The least useful thing I was given (indeed often a complete waste of time) were the lectures. Lectures (as opposed to a good seminar) are easily substituted for by actually doing the reading, in general.

Posted by: Ian Welsh on August 27, 2003 06:50 PM
Post a comment