I have the FB  glued onto the neck of the '37 Hauser model I've been working on, my first classical guitar, and it's planed dead flat along its length.  The Courtnall book suggests that it will be fine to fret it this way, but I've also read that I should be leveling the neck with a little downwards force applied (a bar clamp attached to the headstock with a 2 1/2lb weight on it, so it's a little back-bowed) in order to give the neck .006" - .010" of relief at the 7th fret when the clamp is removed. This is supposed to bring the neck back to level after the frets are applied.

The fret tangs are nearly a perfect fit for the fret slots, perhaps two thousandths thinner.

I'm not sure which approach to follow at this point.

.002" sounds pretty close to me. The bead on the fret wire may compress the board a bit more than you anticipated. The fret board that I just did had a slot size of .023" and I used wire at .0185" fretting while the neck is still square to compress the board to the max and avoid any back bow. It came out nice and flat.............I usually rely on string tension to pull it into a little relief. 

As opposed to relief, tapering the board has to do with maintaining a high enough action without having a too high bridge. This can also be done by putting a lift in the neck or by doing both. Tapering away the bass side additionally obviates the need to have the saddle significantly higher on the bass side.

I'm building the guitar using the solera outlined in the Courtnall book - 25' radius carved into the solera and a 2-3mm tilt at the neck end. The neck was glued onto the top in the solera. I've been following a method outlined by David Schramm, and so far when I check the action using his method, it's within spec. I'm also getting the measurements outlined in the Courtnall book, in the paragraph that outlines a method for the preliminary checking of the action.

My FB is a bit thicker than the Hauser plan at the soundhole. I could taper the entire lower end of the FB at this point, but this will effectively give the FB a downward slope toward the bridge, so I'll need to be careful not to go too low and check the action height often.  If I understand all the information here correctly, I should get the taper correct first, and then get the relief into the board.

Man, I thought I was confused before...

The Romanillos method has the solera creating a 5mm drop at the end of the guitar starting from just beyond the waist. This combined with about a 25' radius at the bridge location and extending (to a gradually lessening degree...) forward through the soundhole area gives the top not only a side to side arch but a pronounced arch from front to back (i.e."domed"). The neck lift is 2mm.

This creates a seemingly complex geometry to the body/neck but because the top where it joins the sides is curved downward while the center is arched upward, the 2mm creates enough lift to keep a normal bridge height. The fretboard has a slight taper.