Yesterday Thralas notified me that the AUR2-BR ( http://aur2.archlinux-br.org/ ) repo is available at http://projetos.archlinux-br.org/thaur/ . They have done a lot of work, and I mean a LOT... It may be very close to finishing. Having skimmed over the code it looks like it's mostly just an AUR port to Django, which is good, but it could be better. Before we started developing AUR2 we were designing it for a couple of days, and discussing the best options for X, Y and Z. I think the current ERD ( http://ius.student.utwente.nl/aur/aur.png ) is very good and suits most, if not all, of our needs. I'm still critically thinking about that ERD and general design of AUR2 to make it flexible and generally better than AUR. We have a few ideas that would make AUR really awesome, like using an SCM backend, being able to edit PKGBUILDs on the fly from AUR, etc. I think we would achieve better results if we continued our work on AUR2 and simply ported AUR2-BR code, while critically thinking about each piece of code. This would result in a better design, more flexibility, better/more efficient code and possibly less redundant if we use/expand on namcap for PKGBUILD checking. Now if I were Lincoln or Douglas I'd be pissed off for someone suggesting this :P. So I'd like some discussion, especially from them.