Just an idea - what about inserting "checkpoints" in the handbook that describe how to recover to that point after a reboot? Many people seem to restart from scratch when they experience a problem during installation or just run out of time, although it's most often just a waste of time. Some detailed descriptions on how to recover (mount partitions, chroot etc.) would make the install process less painfull.
Imho it would be far easier to have a small section about this in the FAQ (kinda like http://www.gentoo.org/doc/en/faq.xml#bootrescue) or in the Gentoo/x86 Tips 'n Tricks document (which is also useful for other architectures).
Hrm, I'd appreciate a listing on what "checkpoints" you feel are a good choice to be included. I think most users trip over the kernel configuration and the bootloader configuration. Those two parts are easily covered by the FAQ (with a little extension on bootloader handling).
I think failing installations occur due to a) a finished installation that doesn't work after reboot and b) an interrupted installation due to a corrupted CD, packages that fail to compile, users getting too confused, users that didn't drink enough coffee or don't like their computer running when they sleep :) Thinking about how "checkpoints" could be implemented, I had the idea of marking sections in the handbook as "redo" (like mounting partitions, chrooting) and "skip" (like partitioning, unpacking the stage tarballs), but it would get too ugly (or even confusing) I'm afraid. Another approach would be to extend the handbook with another chapter "Troubleshooting" at the end. Many people seem to restart the installation before reading the FAQ or asking in the forums for help. The troubleshooting chapter would include a collection of the FAQ entries mentioned above (how to get back to kernel configuration) plus some more common pitfalls like these: for a) - grub.conf is misconfigured (error 15 File not found and other) -> instructions on how to invoke grub shell, tab completion etc. - system immediately reboots -> recheck kernel configuration - kernel hangs with kernel panic VFS unable to mount... -> forgetting to specify root= or real_root= in grub.conf or wrong configured kernel (filesystem, IDE support) - booting hangs with "unable to check /dev/BOOT"... -> fstab not configured. for b) I think a summary on what has to be done to reach a certain stage (install missing packages, edit a config file, configure kernel) would be sufficient. This could also be used by a) as reference.
I've written a generic explanation on how to recover from a broken installation. Going more into detail would clutter up any document (just check the forums for all possible issues you can come across - it's impossible to document all those) and I think that this small explanation is still sufficient for all users to comprehend how to fix their system. See http://www.gentoo.org/doc/en/gentoo-x86-tipsntricks.xml#recover
