> What is the list opinions and arguments on this matter > regarding what have been said? The way I've been partitioning disks has a long legacy, back before laptops were common, and most UNIX devices were servers. Here's how we partition a "standard" Linux desktop or laptop: / (root) 16GB /boot 100MB /var 3GB the rest of the disk goes into "/disk0", a hold over from pre-volume management days when server disks were /disk1, /disk2, ... /diskN. We create a /disk0/scratch area (drwxrwxrwt) for users to write into, most users don't have root privs. On laptops, users $HOME directories are underneath /disk0/home. The goal is to keep the OS partitions small, giving the most free space to the user, and prevent them from filling the disk and impacting system performance. For the most part this works, but we occasionally run into problems with upgrades if the new version does not fit in the footprint of the prior version. In this case, we backup /disk0, re-partition and re-install, then restore the contents of /disk0 To summarize... separtate user data from system data, and separate dynamic data from static data.
