1. Are MOOCs an idea that were floating around the halls of universities for some time now, or was the first one in 2008 really a watershed moment?
Many of the ideas that go into a MOOC were around before CCK08 but that course marks the first time the format came together. In particular, we would point to David Wiley's Introduction to Open Education course, which was offered as an open wiki (later called the Wiley Wiki - see https://sites.google.com/site/themoocguide/cck08---mooc-basics ) and Alec Couros's open course ECI831 - Social Media and Open Education (see https://sites.google.com/site/themoocguide/social-media-and-open-education ). These two courses were of course influenced by other work in the field - the concept of open education, in which Wiley was a pioneer, with a license preceeding the Creative Commons licenses, the open wiki, which of course was made famous by Wikipedia, and more.
What makes the MOOC offered by George Siemens and myself different was that it was a distributed course. This is what enabled the 'massive' part of 'Massive Open Online Course'. The software developed to support the course - called gRSShopper, written by myself - was designed to enable the use of open educational resources (OERs) and to aggregate student contributions nwritten using their own weblog environment (and later, discussion boards, Twitter, Facebook, Delicious, and more). I've been working with aggregators since the beginning of RSS and of course have been influenced here by the work of people like Dave Winer and Aaron Swartz, among many others. The OER movement itself has roots in the open access movement, which created the Open Archives Initiative, and eventually the UNESCO OER program.
What made CCK08 a watershed moment was the realization that the use of distributed open resources would support - with ease - an attendance in the thousands. We weren't expecting 2200 people in CCK08; George Siemens has quipped that we were expecting about 24 people, if we were lucky. After all, the course was devoted to a pretty obscure topic - the theory of Connectivism, a pedagogical theory articulated by George and myself. And the software and course design were the first to explictly invoke the theory, and to focus on connections rather than content, which suggested the distributed and connected approach.
2. Did you expect to see MOOCs explode in popularity so quickly, giving rise to these new online academies - like Khan, Udacity, Coursera - that hope to provide Ivy League level courses to anyone with an internet connection?
I'm not sure whether I expected the format to take off so quickly, but I was not surprised at all that once it proved successful it would be adopted by the Ivy Leagues (who would receive credit for its 'discovery') because this follows a well-established pattern in our field.
The idea of open licensing existed well before being made famous by such things as GPL and Creative Commons - I've mentioned David Wiley's open license, and I have discussed in other work (such as the MUDLib open licenses by Lars Pensjo and George Reese - http://www.downes.ca/post/55283). The Open Archives Initiative software and specification was developed in Europe, but became MIT's more widely used and well-known DSpace. Open educational content was also around before being made famous by MIT's OpenCourseWare. Someone from Yale has been annoited the expert on Edupunk. Learning Objects, the Learning Management System, Learning Design specifications - there were all developed elsewhere.
I never had any doubt that the model itself would be successful. Though we hear a great deal about the quality of learning resources and the need for credentials, the demand from people without access to any university resources has been consistent and strong. There is a large following throughout the world for all this work in open online education, because it eliminates one of the great advantages the wealthy have always enjoyed over the poor. And with open access, we can work on things like quality, assessment and credentials on an ongoing basis.
3. Of these recent start-ups, Codecademy and Udacity both specialize in teaching students to program computers and write code (either Python or Java). Their founders say that they'll branch out into other areas, such as the humanities, hard sciences, and social sciences, with time. (Coursera lists a few humanities courses on its website, but none are currently offered.) How successfully do you think professors can teach, say, Shakespeare or Heidegger, via a MOOC? Computer programming lends itself to an online module, with assignments than can be auto-graded. How optimistic (or pessimistic) are you that poetry, art, and physics can be similarly taught?
The Codecademy and Udacity haven't faced some of the issues in massive open online courses that we've already faced. My background is in philosophy, my first open online educational resource (which has also been widely popular over the years) was "Stephen's Guide to the Logical Fallacies," launched in 1995. http://www.fallacies.ca and really due for an upgrading (I am considering offering a 'Logic and Critical Thinking MOOC' in the fall to do just this). We offered a MOOC in 2010 called 'Critical Literacies', which was another humanities course. And of course we've seen a lot of activity in open online learning in Philosophy, from the Dreyfus course in Existentialism offered through iTunesU to the series of podcasts being created by Peter Adamson at King's College London http://www.historyofphilosophy.net/
It's actually very easy to create self-marking quizzes - I remember the software, Hot Potatoes developed by Martin Holmes at the University of Victoria. There is a full set of IMS test question and interoperability specifications. There is even essay-marking software - Wired covered it in 2001 http://www.wired.com/culture/education/news/2001/08/45806 (my coverage http://www.downes.ca/archive/01/08_23_news_OLDaily.htm ). So if people want to go that way there's plenty of oppportunity. But it won't happen.
One of the characteristics of our MOOCs that is not a part of the Codecademy and Udacity model is the understanding that the evaluation of learning is not about testing for content acquisition. We say explicitly that the content is the "McGuffin" - it is the thing that gets people together, gets them talking, gets them thinking in new ways. The content could be very different - the AI course, for example, could have opted for very different content, or it can get away (as it did) with quickly-created hand-drawn videos and lessons. And people can learn and remember this content - even without understanding the material. So while you can test for content acquisition, you need to think of assessment as something quite different.
In the MOOCs we've offered, we have said very clearly that you (as a student) define what counts as success. There is no single metric, because people go into the course for many different purposes. That's why we see many different levels of activity (as we also saw in the AI course).
With respect to actual assessment and credentialing, there are two basic approaches (or three, if you count badges (see the Mozilla badge program), but I don't really). The first is the Big Data approach - instead of using a few dozen data points, which is what the testing regimen does, you track a student's activities and construct a profile from the full spectrum of his interactions with the material and other learners. This is the work of a field called 'Learning Analytics' (which should be 'discovered' by the Stanford-MIT nexus any time now). The second, which is my own approach, is a network clustering approach - the idea is that in a network of interactions in a community, expertise constitutes a 'cluster' of activity, and a person's learning can be assessed as a form of proximity to that cluster. The Learning Analytics and Network Analysis approaches are not mutually exclusive.
What does this mean in practice? Let's take the study of Heidegger. There is a worldwide community of people who are interested in Heidegger, centered around some of the experts (such as Dreyfus, for example). The people most expert in Heidegger tend to communicate with each other, and to be followed or read by the rest of the community. Other Heidegger clusters also exist and are followed to a lesser degree by the experts. Heidegger novices begin by following a MOOC in Heidegger and gradually contributing their own thoughts (the process we've described is 'Aggregate, Remix, Repurpose, Feed Forward'). As they contribute the work they offer is read by, and talked about by, other people. They are, in essence, 'recognized' as having mastered Heidegger by other people who have already mastered Heidegger.
This is what actually happens in the pre-internet world; you can graduate with a PhD in Heidegger but nobody will hire you unless you have made a contribution to the field that is considered remarkable (literally: worth remarking on) by others in the field. You can study and pass all the tests you want in order to be a brain surgeon, but you will never be one until you have been recognized as such by someone who is already a brain surgeon, and this through a long (and arduous) internship process. Exams and creddentials are shorthand used to create a screening process, but this is no longer needed when the entirety of network data is available to employers and other experts.
As time goes by, the people at Codecademy and Udacity will understand that the community that forms around the courses or subjects are a lot more important than the content. They will discover (if they look) for example that people return to courses they've already offered because these people know they can establish their credentials by participating in the community and helping other people. Whatever their own assessment methods are, they will be superceded by the actual assessments made by their community as a whole. After all, when 248 people get a perfect score in the course http://blogs.reuters.com/felix-salmon/2012/01/23/udacity-and-the-future-of-online-universities/ how else do you establish which of these really understands, and which of these just remembers well?
4. The founder of Udacity, Sebastian Thrun (also a Google fellow and Stanford professor), talks about his motivation to launch the site in terms traditional academia being too privilege- and class-based, and saying that we have a duty to use technology to bring education to the millions of folks worldwide who don't have the opportunity or money to spend four years at Harvard or Yale. Do you similarly feel that academia is too rigid and privileged, and would be improved be being deconstructed and "flipped"?
Yes. I've spent a lifetime pursuing this objective.
But let's be clear about exactly what this objective is. It isn't about (as the OECD report was titled) "Giving Knowledge for Free". http://www.oecd.org/document/41/0,3746,en_21571361_49995565_38659497_1_1_1_1,00.html That is, it isn't about the wonderful rich people engaging in charitable work as some sort of civic duty (as though that somehow made they wealth OK). It's about actually empowering people to develop and create their own learning, their own education. So not only do they not depend on us for learning, but also, their learning is not subject to our value-judgements and prejudices. We (those of working in MOOCs) have also been clear about the influences of people like Ivan Illich and Paulo Freire. And it's not just about 'flipping' courses. It's about reducing and eventually eliminating the learned dependence on the expert and the elite - not as a celebration of anti-intellectualism, but as a result of widespread and equitable access to expertise.
None of this happens by magic. There isn't some 'invisible hand' creating a fair and equitable education marketplace. The system needs to be built with an understanding that personal empowerment and community networks are the goal and objective.