As Andy points out there are times when you do things differently for specific reasons and that looks to be the case with the Macdev situation (it also was when I was doing a version of the Freestyle valve system with the tip o-ring on the center shaft that always wanted to come off). Sometimes you are constrained in the space/location and it causes different decisions. I haven't found a perfect science to it, but have found my method is usually a good starting point for most of the things I design.
I pretty much always rely on extrusion to seal the o-rings with the lower durometer o-rings, and it seems to work well for the low forces and low frictions we find in many paintball situations.
Can't say I have looked at the book for a very long time. Much of the standard industrial methods they specify just simply doesn't work for paintball. Much of what I do comes from years of doing it and some amazing conversations with someone I consider to be a pneumatics Guru, Mike Woods (of Nova, SuperNova, Assault etc. fame)
I've had many o-rings that ended up with a little too large a groove diameter and too thin a groove cross sections such that they leaked, and simply putting the parts back on the lathe and opening them up got everything to seal up perfectly.
That's called indexing. You need to specify the parts that need to align and give an angular tolerance. Some machine shops will hate you for it, others won't care. It can be a HUGE PITA to get right after you take into account tolerances and finishes. Try not to make it overly critical if you can.