My main disagreement was providing a pacman -So interface that queried the server directly.
For three main reasons: 1. The load on the server from having pacman -So run on end users systems. This is also not something that could be mirrored. It would *all* point to the arch webserver (of which we only have one right now). 2. This would *tightly* couple pacman to the arch website. It is hard to express how much I disagree with such a tight coupling. 3. Why would someone need to search to see what owns a file that they don't have on their system, with pacman?
