> 1) Is there a reason to prefer XHTML 1.0 over XHTML 1.1?
None that I know of - I guess I just had the 1.0 header to hand & for this particular use, I don't think there was any difference between 1.0 & 1.1.
> 3) gwynevans -- did you whip this out as an example, or was it something you have? I ask because it has no title, for instance.
At the time I'd not considered custom metadata & pre-processing, so just had a 'build_ePub.bat' in the folder which set some of the metadata via the Caliber command-line, e.g. 'html2epub --margin-right=10 --level1-toc="//h2" --chapter="//h2" --cover="Konrath, J.A - Jack Daniels 01 - Whiskey Sour.png" -t "Whiskey Sour" -a "Konrath, J.A" "Konrath, J.A - Jack Daniels 01 - Whiskey Sour.html"'
> 4) I'm thinking to move all CSS to a separate file, any reason I shouldn't?
If you've come up with a standard set of styles that you want to reuse, then it's worth considering, although the main reason to do so in the web-site case is to allow global changes by editing the one file, which may be less of an issue in this particular usage.
> (BTW, the page-break-before: always;
part -- is that specific to ebooks, or part of XHTML? 'Cause I've been wondering about how to hard-code that.
Standard, but it's less well known as it's focussing on the print side of things - http://www.w3schools.com/Css/pr_print_pagebb.asp