Traditionally, scholars distinguish between three main categories of learning theories: behaviorism, cognitivism andconstructivism. Stephen Downes and others would add a fourth one: connectivism, but this is disputed.

The central application of connectivism to date is as a theory of what happens in Massive Open Online Courses.

The connectivist theory describes learning as a process of creating connections and developing networks. It is based on the premise that knowledge exists out in the world, rather than inside an individual's mind. Connectivism sees the network as a central metaphor for learning, with a node in the network being a concept (data, feelings, images, etc.) that can be meaningfully related to other nodes. Not all connections are of equal strength in this metaphor; in fact, many connections may be quite weak.

On a practical level, this approach recommends that learning should focus on where to find information (streams), and how to evaluate and mash up those streams, rather than trying to enter lots of (perishable) information into one's skull. Knowing the pipes is more important than knowing what exactly each pipe contains at a given moment. This is the theory. The practice takes place in Connectivist MOOCs (cMOOCs), like Change11. Here, people are free to participate at will. Each week a subject is discussed during synchronous sessions, which are recorded and uploaded for reference on the Change11 website. The site also includes an archive of daily newsletters and RSS-feeds of blog posts and tweets from participants.

cMOOCs tend to be learner-centered. People are encouraged to pursue their own interests and link up with others who might help them. But the distributed and free nature of the projects also leads to complaints; participants often find it confusing when they attempt to follow up on all the discussions (the facilitators say one should not try to follow up on all the content).

Stephen Downes: This implies a pedagogy that (a) seeks to describe 'successful' networks (as identified by their properties, which I have characterized as diversity, autonomy, openness, and connectivity); and (b) seeks to describe the practices that lead to such networks, both in the individual and in society (which I have characterized as modeling and demonstration (on the part of a teacher) and practice and reflection (on the part of a learner).

One example of a MOOC that claims to embody the connectivist theory is change.mooc.ca. The "how it works" section of the site explains what connectivism means in practice. The MOOC organizers developed a number of ways to combine the distributed nature of the discussions with the need for a constantly updated overview and for a federated structure. So, if your team wants to organize an open online course, these are five points to take into consideration:

There is no body of content the participants have to memorize, but the learning results from activities they undertake. The activities are different for each person. A course schedule with suggested reading, assignments for synchronous or asynchronous sessions is provided (e.g. using Google Docs spreadsheets internally, Google Calendar externally; one could also use a wiki), but participants are free to pick and choose what they work on. Normally there is a topic, activities, reading resources and often a guest speaker for each week. One should even reflect upon the question whether a start and end date are actually needed. It is crucial to explain the particular philosophy of this kind of MOOC, and this right from the outset, because chances are learners will come with expectations informed by their more traditional learning experiences.

1. It is important to discuss the "internal" aspects, such as self-motivation: what do the participants want to achieve, what is their larger goal? And what are their intentions when they select certain activities (rather than other possibilities)? Everyone has her own intended outcome. Suggest that participants meditate on all this and jot down their objectives. And how can they avoid becoming stressed out and getting depressed because they feel they cannot "keep up with all this?" The facilitators should have a good look at these motivations, even if it's impossible to assist every participant individually (for large-scale MOOCs).

2. Ideally, participants should prepare for this course by acquiring the necessary digital skills. Which skills are "necessary" can be decided by the group itself in advance. It's all about selecting, choosing, remixing - also called "curating". There are lots of tools which you can use for this: blogs, social bookmarks, wikis, mindmaps, forums, social dashboards, networks such as Twitter with their possibilities such as hashtags and lists. Maybe these tools are self-evident for some, but not necessarily for all the participants.

3. The course is not located in one place but is distributed across the web: on various blogs and blogging platforms, on various groups and online networks, on photo- and video-sharing platforms, on mindmaps and other visualization platforms, on various tools for synchronous sessions. This wide variety is in itself an important learning element.

4. There are weekly synchronous sessions (using Blackboard collaborate, or similar group chatting tool). During these sessions, experts and participants give presentations and enter into discussions. Groups of participants also have synchronous meetings at other venues (such as Second Life). Try to plan this well in advance!

5. Many participants highly appreciate efforts to give an overview of the proceedings. Specifically, the Daily Newsletter is a kind of hub, a community newspaper. In that Daily there is also a list of the blog posts mentioning the course-specific tag (e.g. "Change11"), also the tweets with hashtag #change11 are listed in the Daily. Of course, the MOOC has a site where sessions, newsletters and other resources are archived and discussion threads can be read.

From the very beginning of the course, it's necessary to explain the importance of tagging the various contributions, to suggest a hashtag.

For harvesting all this distributed content, Stephen Downes advocates the use of gRSShopper, which is a personal web environment that combines resource aggregation, a personal dataspace, and personal publishing (Downes developed it and would like to build a hosted version - eventually financed via Kickstarter). The gRSShopper can be found on the registration page, which is useful primarily for sending the newsletter. It allows you to organize your online content any way you want, to import content - your own or others' - from remote sites, to remix and repurpose it, and to distribute it as RSS, web pages, JSON data, or RSS feeds.

Stephen Downes: For example, the gRSShopper harvester will harvest a link from a given feed. A person, if he or she has admin privileges, can transform this link into a post, adding his or her own comments. The post will contain information about the original link's author and journal. Content in gRSShopper is created and manipulated through the use of system code that allows administrators to harvest, map, and display data, as well as to link to and create their own content. gRSShopper is also intended to act as a fully-fledged publishing tool.

Alternatives for registrations: Google Groups for instance. But specific rules about privacy should be dealt with: what will be the status of the contributions? In this MOOC the status is public and open by default, for Downes this is an important element of the course.

Some MOOCs use Moodle, but Downes dislikes the centralization aspect and it's not as open as it could be, saying "people feel better writing in their own space." Other possibilities: Google Groups, Wordpress, Diigo, Twitter, Facebook page, Second Life; but each course uses different mixtures of the many tools out there. People choose their environment - whether it is WoW or Minecraft. Students use Blogger, WordPress, Tumblr, Posterous as blogging tools.

Give participants a means to contribute their blogfeed. In "Add a New Feed," Downes explains how to get this structure and additional explanations (via videos) in order to contribute their blog feed. The administrator in this case uses gRSShopper to process the content and put it in a database, process it and send it to other people. Alternatively one can use Google Reader (the list of feeds is available as an OPML file - which can be imported to other platforms). There is also a plug-in for Wordpress that lets you use a Google Doc spreadsheet for the feeds, then Wordpress for the aggregation). Many other content management systems have RSS harvesting features.

Each individual could run her own aggregator, but Downes offers it as a service. But aggregators are needed, whether individual, centralized or both.

• Change MOOC: How this Course Works

• What is a MOOC (video)

• Success in a MOOC (video)

• Knowledge in a MOOC (video)

• Introduction and invitation (video)