Pacific Connection(英語)

MIT's OpenCourseWare Launches a Movement

In 2000, a faculty committee at the Massachusetts Institute of Technology considered a for-profit "distance learning" service for the institution--a sort of "MIT online." According to an MIT PowerPoint presentation, the committee first produced a "10-inch thick" report that concluded that there was no business model-which was not the kind of conclusive insight then-president Charles Vest was looking for. So the committee went back to work and submitted a single-page memo, which suggested that MIT take the basic written material professors put together in preparation for teaching course-and put it online.

From that simple idea, a movement has formed-OpenCourseWare-which includes not only the original MIT site ( but more than 100 participating colleges and universities, linked by a consortium (, that stretches from Asia to North and South America to Europe, the Middle East and South Africa. In Japan, the Japan Opencoursware Consortium includes 17 universities. In the United States, MIT has been joined by Harvard Law School and the Berkman Center for Internet and Society, Michigan State, Tufts, and the University of California, Berkeley, among others.

Other schools, while not formally affiliated with the OCW Consortium, are making similar moves to put more of their material online and in the public domain. One of the most notable examples is Harvard University's Faculty of Arts and Sciences, which will make faculty member's scholarly articles available for free in an online repository. The move came as a response to what the university called an overly restricted scholarly publishing system that has limited the number of people able access the research.

The MIT OCW pilot site opened in 2002 with 50 courses. By November 2007, almost every MIT course was represented. Today, the stats are impressive. 1,800 courses with their associated syllabi and reading lists; notes from 15,000 lectures; 9,000 assignments; 900 exams, about 40% with the answers. Some courses have gone well beyond this, with video lectures, textbooks, animations and simulations. Most of the multimedia content is available on MIT's "iTunes University" site which has more than 1,000 tracks from almost 50 courses.

The resemblance to the open source movement is obvious. The OCW isn't a quite a collaborative model in the sense that you can "improve" upon a professor's work or add a description of your own course. But the material is published under a Creative Commons license, which resembles in intent the GNU General Public License for source code. The terms of use are straightforward: keep it non-commercial and give proper credit to MIT.

The "open sourcing" of university-quality material is also one more step in the Web's becoming a place for information that is, to some degree, vetted. The distinction is important because the Web is now the propagator of both facts and misinformation, headlines and hoaxes, carefully researched reporting and unfiltered blather. As it has turned out, the best way to separate legitimate news from illegitimate nonsense is to do online what people having been doing for years offline-consider the source of the material. Who is the writer? Who is the publisher? What are the reputations of both? A strong reputation doesn't guarantee absolute truth, of course, just as, conversely, not all anonymous rumors are false. But reputation is a good start. You can see this transition on the U.S. version of Wikipedia, where the anyone-can-edit philosophy remains true, but if you make a claim, you better have a "notable" source that you cite in a footnote. OpenCourseWare is significant precisely because the professors and their institutions have reputations to keep.

OpenCourseWare is also represents another step in the democratization of education-putting high-end educational materials in the hands not just of enrolled students, but of anyone with the curiosity and tenacity to understand it. MIT has identified almost half the visitors to its site as being "self-learners," most of whom are either adding to their personal knowledge or keeping up with developments in their field. When it comes to keeping up with technology, that pretty much describes all of us.

A conversation with MIT's Shigeru Miyagawa

MIT Linguistics Professor Shigeru Miyagawa was part of the original faculty committee that came up with the idea of OpenCourseWare, as well helping to start the Japan OCW Consortium. He was a good candidate: in 1994, he put MIT's entire Japanese language program online. (The link can still be found at Professor Miyagawa is spending his sabbatical as a visiting professor at Keio University. I interviewed him by phone from his home in Tokyo.

How did the idea for open courseware come up?

In spring 2000, the provost and president of MIT asked a few us to serve on an MIT-wide committee to come up with an e-learning strategy. The committee consisted of a member of each school; I was called in as a member of the school of humanities, arts and social science. Back then, the bubble was still bubbling, so the assumption was that this would be a for-profit model. To that end, we were joined by members of Booz Allen Hamilton, a top tier consulting firm that does big e-learning projects all over the world.

We began by studying a large number of existing for-profit models. We interviewed some 60 institutions and companies that were doing for-profit e-learning ventures. We contacted our alumni. We sent out a large number of questionnaires and heard from about 500 respondents. And we also talked to 60 MIT faculty members who were already putting up their teaching materials on the Internet. That turned out to be especially important: we wanted to understand why faculty members would take the initiative on their own.

And what did you conclude?

Something important: that these faculty members were motivated not by compensation-because they weren't getting any-but because they wanted to improve the teaching of their courses. They were spending the time putting things up on the Web strictly out of the spirit of education.

And we also realized that a for-profit model would be difficult to do, at least for us, because of some constraints. From the beginning, we had decided that we would not offer credit or a degree. We felt very strongly that MIT education was something that we offer only on campus with a select group of students. It was not something we could replicate online. That brought home the point to us that we shouldn't be doing ""

Around the same time, we realized that those 60 faculty members who were digitizing their own teaching material strictly in the spirit of education represented the direction forward. They became our model-the exact opposite of a for-profit model. What if we took the teaching materials for almost 2,000 courses and simply gave them away for free with virtually no conditions, except that it has to be for non-profit use and give MIT proper attribution?

So this the way OCW began?

Yes. Individual faculty members, not just in the U.S. but around the world, had discovered the Web and were putting up some teaching materials. But MIT was the first in doing this in an organized, institute-wide effort. There is no registration. We have cleared copyright restrictions under Creative Commons. The material is there for the world to use.

Was OpenCourseWare influenced by open source software?

Yes. MIT's culture is very much open source. We have some open source gurus here, and MIT is very much an open institution. You can walk onto campus. You can walk into the buildings. OpenCourseWare reflects that culture.

Have you been surprised by how much the idea has grown?

I've been shocked and astounded, because we began so modestly. Our original idea was simply to have enough content so that people will find it useful. We proposed the idea in 2000, President Charles Vest announced it in April 2001, and we started building open courseware in 2002. Initially the courses were what we imagined-just syllabus and lecture notes. But some professors also put up video of their lectures, which was quite remarkable. More recently, some people have put up the actual assignments and the answers to those assignments-even the final examinations and the answers. That is remarkable. It means that these faculty members are actually changing the content of their class every time they offer it. To me, that means we made the right decision, because we captured the educational spirit of the MIT faculty members. And there are other schools now that are starting to do this, as well.

That's good for MIT students. What about the rest of the world?

OpenCourseWare has made available on the Internet very high quality educational materials that are vetted by the educational institutions involved. If you look at an MIT physics course, for example, the quality of the lecture notes, simulations, and videos are vetted by the professor first and foremost, and by MIT. And they are all available for anyone to use.

We get about two million "visits" per month, a visit being a 30-minute interval during which time a user may do multiple page views. The OCW people tell me that translates into roughly 1.6 million real human beings per month. You can see why those of us who tentatively proposed this open courseware in the first place are shocked. We had no idea it would come this far.

These visits are from around the world?

Yes. The largest number are from North America, followed by East Asia-Japan and China-followed then by Western Europe, India, Russia and Eastern Europe, South America, and the northern part of Africa. The demand seems to follow the quality of the Internet connections. We get the fewest visits from sub-Saharan Africa. Translation matters here. There's a lot of interest in Asia because of the Chinese translation. Likewise, we have good numbers in South America partly because of Spanish and Portuguese translations.

We've done surveys, as well. About 49 percent of the visitors are "self-learners": people who are interested in a subject and have the self-drive to take a look. About 32 percent are students, typically high school or university. The remaining 17 percent are teachers in high schools and colleges.

You were involved in founding the Japan Opencourseware Consortium. What's that?

It is a Japanese consortium of universities that follows the OpenCourseWare model. I happened to be in Japan in the summer 2003 and spoke to the presidents of six universities- the University of Tokyo, Tokyo Institute of Technology, Kyoto University, Osaka University, Waseda University, and here at Keio University. They all knew about open courseware and all were very excited about the model. So all six said yes, they would start open courseware. Now they are up to 19 universities. Right now, there is a global open courseware consortium with about 180 universities around the world from every continent. But the Japanese open courseware consortium is the most impressive in the sense that the very best universities in Japan are collaborating. I don't know of any other single effort in Japan where such well-known universities have come together in such large numbers, yet a lot of Japanese still don't know about it. You can get to it at

You mentioned that some professors are putting up videos of their lectures. How typical is that?

Video lectures are still unusual, but some MIT professors were videotaping themselves even before open courseware started. For example, Professor Walter Lewin for Physics 802 was written up in the New York Times and has become an Internet guru because his video site is so remarkable and entertaining. Another popular online teacher is Gil Strang, who has posted videos for every class of his linear algebra course.

And there are great things happening here in Japan. For example, the University of Tokyo carries some video lectures by the Nobel laureate physicist Professor [Masatoshi] Koshiba and Todai President [Hiroshi] Komiyama. They jointly hold a course for freshmen and they made it available as video lectures, even podcasting. It's been massively popular, so much so that I think that it put a huge burden on their server. Another example comes from the medical school at Kyoto, which is famous for its kidney transplants. One of the professors who is a specialist in this area is so excited about OpenCourseWare that he has made available the actual video of a kidney transplant operation. It is something! And recently, Kyoto University made available material from Dr. Hideki Yukawa, Japan's first Nobel Laureate. It has amazing stuff, like a hand-written draft of his Nobel Prize article.

The videos sound like another of those "shocking" repercussions you weren't anticipating at the beginning.

That's right. But as it turns out, if you produce a video, then people are going to come by and look. I just discovered that someone had put up one of my open courseware video lectures from open courseware on You Tube. Why not?

There are other repercussions, as well. It used to be that teaching materials remained protected within the confines of the classroom. With OpenCourseWare, there's a fundamental shift: we are exposing to the world what we do in our classrooms. And to be frank, that does seem like something that we as a university ought to do-because this sort of transparency is already expected in other parts of our life. When I recently bought a webcam, for example, I could go to the Sony website and get a lot of information about the product. But that hasn't been true in higher education. A high school student looking at the coursework on, say, electromagnetism would get only a couple of paragraphs of description from the catalog. By contrast, open courseware allows you to see what we are actually teaching. That's a remarkable change. Last year in a survey we did with our freshmen, 17 percent said that OpenCourseWare was a significant factor in choosing MIT.

Beyond MIT's reputation?

Yes, the fact that they can get a concrete sense of what we do at MIT.

What about the professors themselves? Has OpenCourseWare changed how they teach?

Inevitably. When you know that not only your students, but the whole world is going to look at your teaching materials, you work very hard to come up with the very best. People tell us that's the case. And we are also seeing more sharing of teaching materials, not only within the institution, but between institutions. I'm starting to hear from people all over the world that they are using some of the material I put on open courseware.

What about the democratization of education? Is that a long term benefit as well?

Yes. One thing we are seeing with the global OpenCourseWare consortium is that we have huge numbers of universities that are joining with the same goal and spirit of sharing their materials with the world. There's a sense of a community of people in higher education around the world that are coming together and sharing their knowledge.

An MIT OpenCourseWare "Speciman": 6.821 Programming Languages

Among the 1800 or so courses described on the MIT OpenCourseWare site is Professor Michael Ernst's 6.821 Programming Languages, a graduate-level course that "teaches the principles of functional, imperative, and logic programming languages." Virtually all of the associated materials (but none of the lectures) for the course as it was taught in fall 2002 are online. These include:

  • A compiler and documentation for Scheme+, the LISP-dialect language used in the course (MIT has licensed both under GNU);
  • A course description, including prerequisites, problem sets ("Be forewarned that many of the problems, especially the programming assignments, are challenging"), the two exams, grading criteria and more;
  • A calendar of the lecture topics;
  • The assigned problem sets;
  • Nine selected exams from 1998-2002, three of which include the solutions;
  • A ZIP file containing all the courseware in a single file.

A set of FAQs on the site explains both the benefits and restrictions of OCW. There is no college credit. Beyond its own faculty, MIT doesn't certify any schools or instructors to teach the material, and the MIT instructor can be emailed, he or she is probably too busy to reply. The exams samples aren't necessarily complete-instructors publish only as much as they are comfortable with. On the other hand, with proper credit to MIT, you can re-publish and even translate the material.