My 2 cents

I have give the subject of academic wiki some thought in the last 6 months. I mean a wiki that could replace the paper journal and its publisher. But it would need a different collaboration model, In any sense that affects thier income the Academia is quite conservative - almost backwards.

A researcher would not want to see his data sets vandalized or corrected anonymously. Such a wiki would require   both public and private zones private - with complex setting to allow group based collaboration. The data would be useless to anyone wanting to replicate experiment results without documentation and access to the original analysis code. It would be neccessary to have anonymous peer review. It would require a new bunch of licenses - since many universities are interested in having patents for in quickly and transparently sharing thier discoveries.

There are also social matters such as a prestidge and fashion and a great resistance to change of publishing in your community. To get it going would require an large community of influential academics who are not afraid to chalange the large publishers and thier editors who are the current stake holders

I would imagine developing such an acdemiwiki could be to developed quite quickly once MW matures. I would need cloud based storage, a source control system, advanced means of authentications. P2P capabilities. But almost all of these things now are avaiable.