Hmmm... let me tell you my story with Moodle.
I'm just a lowly English teacher in a public school in Chicago. I took an interest in Educational Technology and got into Moodle a while back and started looking for ways to build out a server at my school.
I first started with shared hosting. I had to deal with all of the user registration, which was a pain, but I had two class periods using it, sometimes a class at a time. Then I moved to set up a LAMP server in an old Dell box we had laying around. Getting the permissions from the district to set up the server was an ordeal, but I was able to get it running and authenticating to their directory services. That solution worked well enough that I could use it as a 'proof of concept' for my administrators and we were able to purchase a legitimate, RAID-equipped, multiple core production-grade server.
My point here is that in some cases, teachers might be looking for some kind of tiny solution to prove that a Moodle implementation can work before a whole school dedicates the resources to suit the 300 to 3,000 clients that a school might need to serve.
In that case, it would be worthwhile to give the RPi a shot as an experimental Moodle server. How about this?
You had a Model B Raspberry Pi with an external drive for the database and moodledocs directory.
Moodle does support alternatives to Apache2 and MySQL. They aren't all that well tested, but it is possible to scale down some of Moodle's system requirements by using more resource-economical components. Nginx and SQLite?
Throw in some kind of DyDNS client and someone could have (with their school's permission, of course) a server up and going that could probably serve a single teacher's course load.
It might not be very fast, but neither was my shared hosting solution a year ago. It was enough though to get me used to Moodle and, more importantly, used to advocating for it.