What sets Coursera apart from my other experiences with distance learning (MIT's OCW, Stanford's video courses, Khan Academy)? All previous venues shared the essential qualities of being free, providing rich video resources, being available on demand, and not being restricted by prerequisites or tied to a syllabus (although Khan does supply a context graph of recommended relationships). Coursera courses provide, in addition to all of this, a focus on interaction among all participants that pervades the structure and experience of the course.
User Forums: These are the primary mechanism through which students interact with each other, the community TAs, and the professor(s). They provide a chance to ask questions and engage interactively with the answers. In addition, forum users frequently volunteer their expertise by recommending resources and providing insights that can be both surprising and very helpful. This is the key difference between the Coursera student experience and the experience of watching videos in other, non-interactive online courses: MIT-OCW and Stanford Video lectures provide no interaction, no community, while Khan Academy, with its very different focus and granularity, has some minor interactive responses to specific videos. Coursera's user forums are relevant to the entire course and comprise many thousands of highly engaged participants.
In the beginning I was interested just in the videos and tended to ignore the forums, feeling I didn't want to waste my time with others as ignorant as I. However, I found that I was very, very mistaken: the Coursera forums have become a resource that is unique in my experience and has provided both guidance and enriching ongoing dialogue/debate I've never before experienced. Some forum members have become friends with whom I continue to expand my understanding.
Study Groups: These provide a second interaction mechanism, and may be online and distributed or co-located. At the beginning of each course students are strongly encouraged to form study groups based on whatever criteria they find congenial. The study groups become in effect small cohesive communities where ideas are explored in a safe space and people get to know each other. Again, initially I scorned the study groups, thinking I preferred to work things out on my own, and again I was wrong. This time around, in Keith Devlin's Introduction to Mathematical Thinking, the study group I've formed with a friend who is also taking the course is turning out to be enormously helpful: he and I debate our differing reasons for assignment answers, egg each other on to support our positions, and uncover new resources, which we then post to the forum.
Peer Review: This is a third form of interaction, one which — justifiably in my opinion —is very controversial. My experience is that while doing a peer review is quite valuable in the same way that attempting to teach is a very powerful way to learn, peer review responses are not so useful. My opinion was unfortunately reinforced early on by a disastrous experience with idiotic peer reviews, or non-reviews, of an essay I'd spent a week researching and writing. However, engaging in rebuttal and the subsequent interactive dialogue is quite useful.
From my perspective as a student, the major problems involve structural inadequacies in search, forums, and resources.
Searches: The most critical defect in Coursera is the brain-dead search facility, which is a simple string-only search over the titles and text of the forum. You cannot search on the names of posters — e.g., you cannot find all posts by a particular person — you cannot search the rest of the course site, and you cannot do simple Booleans such as "find this but not that", much less take advantage of regular expression patterns. Many subject-specific user forums use Google search, which while not perfect is much more useful than the current Coursera search; Coursera should do the same.
Searches should be faceted, e.g., search on post author, date, ..., the scope should be the full course website, and they should be able to be saved and then used for search refinements. The same automatic visualization tools that should illustrate the evolving forum graph structure (see below) could be used to visualize the results of searches and sub-searches. Structure/relationship visualization is a key tool in gaining deep understanding.
Forums Structure: Issues and possible solutions include the following points.
1. It is currently impossible to track all threads. Forum software needs to automatically assign author-editable tags to entries, and from that develop an emergent substructure among the threads. Threads should be sortable by tag, creation and modification date, author, and title.
2. The current structure is like a rigid class hierarchy and needs cross-cutting views; it should be a graph structure to reflect the emerging multiple POV (point-of-view)s and LOD (level-of-detail)s.
3. The community TAs need a tool for traversing the forums effectively and adding intermediate structure as needed, beyond the automatic evolution suggested in Point 1.
4. There should be a topics forum that is independent of lecture and assignment and could have automatic links into relevant lectures and other forum threads. Obviously, the topics forum needs to evolve deep structure as the course proceeds.
5. An evolving linked visualization of the interacting threads graph would be extremely valuable. The NSDL Science Literacy Maps (http://strandmaps.nsdl.org/) illustrate one possibility. StrandMaps would be a great addition to the courses.
In sum, what is needed is a combination of full-faceted search plus an evolving forum structure with multiple points of view.
Resources: In general, the resources are a fairly traditional set of lectures and recommended readings. The videos I've seen tend to be straightforward, high-quality lectures; the exception to this pattern is a modern poetry course with videos of hour-long close reading discussions by the professor and several graduate students sitting around a conference table. However, in the courses I've taken so far there is no metalevel visualization of context, no use of 2D or 3D visualization of the dynamics of the material, much less the forum threads, no set of relationship graphs among themes, no real integration or connections with the larger domain. In short, there is no reference to or exploration of the ecology of which the subject is part. It is as if hypertext had never been invented. Finally, while forums can be a rich source of recommendations for books, people, and websites, they too lack this awareness of any larger frame of reference.
Why people take the courses: Reasons for taking the courses, which are especially diverse with Coursera due to its heterogeneity and interactivity, include: testing the waters, curiosity, need for community, opportunity to get questions answered, and gaining perspective, as well as a serious intent to complete all the material.Further, as the Coursera courses have progressed, professors are realizing that their target audience is primarily adults, often adults with many other obligations. Thus, the current tendency is to close a course to new enrollments at the end of the course but to keep it accessible to those who did enroll at least until the next time the course is given.Prof. Devlin, for example, has decided to keep the fall 2012 site of his mathematical thinking course open. It would be nice if Coursera established a policy of keeping the course materials on a persistent basis, like the MIT OCW, Stanford video, and Khan Academy materials.
Working with the forums to counteract rigidity: As mentioned above, the forum structures are rigid, like a rigid class structure, and badly need cross-cutting and refactoring capabilities. In the absence of facilities for doing this, I've developed workaround strategies that help compensate for and manage the sometimes overwhelming chaos of thousands of unstructured threads.
First of all, from the beginning of a course in which I intend to be seriously involved, I take advantage of the forums' latest posts list on the forum home page. This lets me track new threads of interest, as well as interesting people and community TAs (remember that it is not possible to search on names). I then subscribe to threads that seem promising and capture content I want to save and work with on my local system.
In addition, in the General Discussion forum I've established threads for topics, experts, and resources I think are important and keep these threads foregrounded by periodically posting to them and providing links to related forum posts I've discovered during my daily prowls of the forum.
Search: Unfortunately, there is little that can be done with the brain-dead search facility. A further frustration is that when you subscribe to a thread and an email arrives with a new post or comment, clicking on the link takes you not to the post but to the top of the thread, and because you can't search on the name of the poster, you are reduced to attempting to discover where the comment is coming from by either scrolling down the thread or trying to enter a string from the comment into the search engine.
 Coursera is one of three major vendors of MOOC (Massive Open Online Course) courseware that have come to prominence in the last year. Since I have direct experience with just Coursera, I have only referenced it in this article. For a brief overview and comparison of three vendors - Coursera, EdX, and Udacity - see the article "MOOC vendors: A Comparison Overview"