The criticism around AFTv5 in terms of product design (nevermind the code) is largely echoed in the comments, yet we seem rather sure that we're giving editors a tool of importance. My daily sampling of what's flowing into the enwiki db from the feature appears to be 99% garbage, with the onus being on volunteers to sort the wheat from the chaff. If we had a dead simple, highly function, and well designed discussion system (see LiquidThreads), wouldn't that be the ideal route for "high value" feedback from knowledgeable non-editors instead of an anonymous one-way text box at the bottom of the articles that's guaranteed to be a garbage collector?
I've been roundly critical of AFTv5; but there are good things to draw from the process, if not the outcome.
It was nice, for example, to see Oliver assigned to bridging the developer-editor gap. It hasn't been 100% successful but it has been pleasant to see the feedback from developers -> wiki.
That said there were downsides; like, the tool seemed to have conflicting aims (is it for editors? For recruitment?) and seemed to lack feedback from wiki -> developers (the tool itself has a number of obvious "flaws" for anyone used to dealing with the wiki eco system).
So; steps forward.
Tom