Why Wiki Forks

The forking thing is a consequence of two principles: collaborative links and everything is editable.

Say you start on your own site, browse a bit and then make an edit. Your change goes to your site, the origin, where you logged in.

Say you start on your own site, browse a bit and then make an edit to someone else page. Your change goes to your site, the origin, where you logged in.

Say you start on your own site, browse a bit and then find a page you want to keep in your site. You click fork. The unchanged page goes to your site, the origin, where you logged in.

Notice that the journal remembers where the page came from so that it can follow links back there even when viewing the forked page on your own site. This is the “collaborative link”.

Things are different if you don’t login to your site or start at some other site where you can’t login. But we find a way to make these cases similar.

Many of our more interesting workflows come from exploiting the way pages move around the federation. Its unusual because most web site operators don’t want you to leave their site. Their business model is based on you not leaving.

Another interesting case is pulling up an older revision from history and forking that. The early version becomes the version and subsequent history is lost. Something to remember, the Journal tells how a page got to be what it is now, not everything that ever happened to it.

One thing I've learned in the last six years is that people don't fork enough pages to keep a small federation alive. So now I'm thinking, What if every time I forked a page my server quietly make a clone of the whole site. If I followed a link from the forked page wiki would still go to the remote site for it unless that site was no longer there. Then the server would consult the automatic whole-site archives.

A farm would of course optimize out duplications so as to not save popular sites over and over. A farm might even offer to share pages from an archive, or even whole archive files when asked.

I'm finding this rather hard to think through. What would be good rules for preserving or culling the collections of whole sites? Hard to say. And their presents would probably change how people used wiki too. Maybe even for the better.