I'm starting to love the typography of Edward Tufte's "The Visual Display of Quantitative Information." I'm loving it because unlike a lot of other texts I've read its margins provide room for very detailed notes. In a normal book if I wanted to sketch out an amazing proof, I wouldn't have the space, but his book has room for that and more.

You know what would suck? If people didn't have any idea that choosing one page design over another actually changed how there product would be used. Not that long ago there was an interview with Joel Spolsky on here where he pointed out that small differences could really change a community. This is just as true in books as it is in an internet application.

So there is no perfect page. There is no secret law. Your better off doing A/B testing than treating anything but the gospel as gospel.