The Internet is already made of "rat's nests" like these. DNS records, BGP route tables, cross-domain embedding, all of these are pulling together information from lots of inconsistent sources to render a page in your browser, and it works pretty well.
The problem Wikipedia is having at the moment is that some people have taken over sections of it as their own "turf". And there will always be a holy war between the completionists and the deletionists, etc. With a federated wiki, you get to pick what kind of info you want to see, and who gets to curate it.
