Tomer Chachamu wrote:
An Apache mod rewrite rule defers all request to the same html page. This page is just a stub that kicks off the Javascript that fetches, parses and presents the content. Wikiwyg can run on top of any existing mediawiki without changes. All content is fetched using XMLHTTP and a mix of action=raw and Special:Export. Links are coloured red or blue a few seconds after they are first displayed, because it takes a little time between parsing the link and getting back a HEAD request to check if the article exists.
Does this mean that we can use it to browse Wikipedia, with a few modifications? That would decrease server load, right?
I've changed the main page from something silly to something that explains this. Yes, it could be used on top of any mediawiki, alongside the existing interface. Running two interfaces has issues though, we'd have to try to replicate any bugs in the PHP parser to keep the thing consistent.
It may not have template inclusion support and it displays masses of XML per page, but it's worth it. ;)
Actually, I'm working on templates now. Templates are actually quite easy, just another http request. The XML dump is just so I can see that the parser is working fine, it won't be there when I'm done.
Tables also aren't supported, nor editing, which is my next sub-project.
By the way, I see <link avaliable="unknown"> a lot in the XML. It should be "available", not "avaliable". ;)
Feel free to correct any other spellings - never my strong point.