What's important about this isn't the server or client, but the protocol.
Like all peer to peer networks, all you need is a clear documenting of the protocol and anyone can create their own servers or clients. This way people concerned with SEO can expose the wiki's contents easier, and users can design a more attractive user interface, all while addressing the same content and sharing it the same way.
If this works as a protocol, it should allow anyone to collaborate on documents that synchronize independent of a central source. This is really useful for people who need to be able to share some information in a multitude of ways and on different networks, online or offline. The only downside is it seems the addressable content has a limited scope. And of course, there's probably (i'm guessing?) not much in the way of access controls to prevent people from getting or modifying information they shouldn't have.
This is precisely my take. I wasn't familiar with SFW until last night (this post), but I've been playing around with this concept for awhile myself.
The only real drawback that I'm seeing so far is some link fragility. There are two countering forces at work here. You can take a link to someone else's content and mirror it back to your own SFW. If the source content changes and you don't take the "pull," your content is no longer relevant. This might mean for popular sources, lots of replicated copies of the source, but not always current. There is also a problem with the always available aspect. If someone is using content from another source, what do you do when the source goes down or just changes locations. At least with a non-federated wiki, the entire set of content goes offline together. That model is less prone to rot since it prunes entire branches and not just bits and pieces.
Like all peer to peer networks, all you need is a clear documenting of the protocol and anyone can create their own servers or clients. This way people concerned with SEO can expose the wiki's contents easier, and users can design a more attractive user interface, all while addressing the same content and sharing it the same way.
If this works as a protocol, it should allow anyone to collaborate on documents that synchronize independent of a central source. This is really useful for people who need to be able to share some information in a multitude of ways and on different networks, online or offline. The only downside is it seems the addressable content has a limited scope. And of course, there's probably (i'm guessing?) not much in the way of access controls to prevent people from getting or modifying information they shouldn't have.