Homebrew MLE with a difference
There are plenty of home-grown VLEs. There are even a few MLEs built in-house. But not quite so many with their own XML e-portfolio format, and extending into blogging and/or peer-to-peer software. Why do it that way? Because "In a school, the creation of meaningful data is a social process."
Tom Hoffman's Tuttle MLE was presented at O'Reilly's Emerging Technology Conference on the 28th of April, and is still a work in progress. Not a problem really, as it mostly doesn't attempt to build everything from scratch, but judiciously arranges and tweaks existing low cost / open source software.
Currently, the heart of the system is a fairly complex hierarchy centred around standards- academic standards, that is, not technical ones. Much like the UK, these drive the academic agenda in US high schools, and present a bit of a challenge for the progressive, project based, individually oriented high school were Tuttle was developed.
The challenge, then, was to map standards up to the appropriate subjects, and down to individual units of work, preferably in such a way that different kinds of assessment could be grouped into a type, as well as similar standards into a cluster. Then some way needed to be found to allow units to be set by teachers, and associated with the right standards. Likewise, from the students' perspective, a clear path of units-to-be-done needs to be generated, and some way of getting work towards the units back to teachers for marking. And all of this arranged by class and cohort.
At the moment, most of that work is managed by some extensions to Zope, the open source web application server, and the hierarchy itself is represented in XML, using a XMLSchema Definition designed for the purpose.
One central server is not how its going to stay, though. Once all the data can be expressed in a well defined format, it becomes possible to have the essential data created and managed much closer to those who own it, and to whom it matters most: the teachers and the students.
Specifically, unit assignment writing and unit task planning and correction are core parts of the teaching and learning process, and you might well prefer to manage these tasks your own tools, rather than on "their" server. Because "they" decide what's going on the central server, while you decide what happens on your own machine, and, crucially, what comes out.
The Colloquia peer-to-peer VLE addressed quite similar concerns by doing entirely without a dedicated central server, and using email for communication between the Colloquia peers.
Tuttle's ideas for the future are subtly different: either harness existing blogging software, or use the open source Chandler Personal Information Manager (PIM) that is currently in the first betas.
The advantage of using Blogging software like Radio Userland or the open source Python Desktop and Community servers is that they combine the personal ownership element of a desktop application with the universality of the web. This is quite literal: the Python Desktop server is a simple 'web desktop' that runs on your own machine, both on-line and off-line, and can be customised to do what you want, they way you want it. Once you're done blogging, your content can be synchronised to a community server that makes sure the stuff remains accessible to other people even if you're offline. Yet because it is all web-based, it is easy to integrate it with central content management servers using technologies like RSS with traceback or XML-RPC.
The Chandler option is the more radical peer-to-peer solution. As Tom has already pointed out; Chandler is similar to Colloquia, but more of a PIM than groupware. In a sense, it is even more personal, since it provides access to your email, your contacts, your calendar and lots of other things that are significant to your whole life, and not just formal learning. But you can also reverse that: Chandler maybe a very good way of integrating learning into all other aspects of your on-line life.
Another point where the Tuttle work might have wider significance lies in the FHSStandardDocument XMLSchema. It's first role is to support a library of school-wide academic standards. This means that it could function as a kind of structured representation of what an educational institution has to offer. Something like that is exactly what's needed for the kind of curriculum negotiation pattern that Scott Wilson and Bill Olivier propose.
In a nutshell, if a student has a Personal Development Profile (expressed in IMS LIP) that sums up both their achievements and aspirations, a representation of the school's course offerings (expressed in something like FHSStandardDocument) could be used to negotiate what courses a student could most benefit from. Not as a way to facilitate automagic course planning by an MLE, but more as a way of facilitating the discussion between a student and a tutor. Helping to suggest opportunities, and narrowing down course options to what's possible or desirable.
What Tuttle demonstrates above all, is how much can be achieved by close observation of the social dynamics of an organisation, and then judiciously tweaking and extending the technology that is supposed to support those people. Not by re-inventing the wheel, but by making maximal use of the increasingly sophisticated open source tools that are becoming available.
There's a list of resources related to Tuttle at Ben Ryan's blog. Tuttle's creator isTom Hoffman. There's an abstract of his talk at the E-tech conference atO'Reilly's.