I was trying out http://ankisrs.net/ and thought that it would be nicer, if one could learn stuff from within emacs.
So I threw together https://github.com/warsus/org-lrn, which basically implements the supermemo-2-algorithm (http://www.supermemo.com/english/ol/sm2.htm) and stores everything in org-file format, so it should be really easy to share and collaborate on decks.
Also you can use the nice org-capture to really quickly add new entries (see the example configfile).
It's probably not much more than a prototype and i'm a pretty busy, but thought the concept was useful enough to post it.