The key words here are Observation and Attention to detail. 
Pay Attention so you don't break anything, and Observe so that you know how it all goes back together. Simple! 

Splitting the cases is no problem, but pulling the pinion gear can before hand can be. If you insist on fabricating your own puller instead of ordering one, be certain to make a collar for the end of the press-screw that will go over the delicate oil snout on the pinion shaft and not press on it directly. Once the gear is removed, the case studs and bolts can be removed, and the cases can be dislodged from each other with a light,soft mallet. 

After you pick all the rollers back off the floor, we are ready to split the wheels: Undo the locks and crankpin nuts, and then wait until no one is looking, and place a large chisle for a wedge between the wheels at about 2 o'clock from the pin, and smack it. I find this method superior to the smack on the edge of the wheel at 3 o'clock, since it takes less abuse, and does not leave hammer dents, which even a copper mallet will leave on a Z wheel. 

The shafts also have locks and nuts, and after they are removed, I place a copper or brass slug on the butt of it before it is tapped out. (It seems that the swing of even a soft hammer tends to damage threads at this step. Save the shafts even if they are spalled, for you may wish to use them to lapp the flywheel tapers later. 

The wrist pin bushings can be pressed out. Many people prefer a simple hand held screwpress, which is especially handy for an emergency bushing repair without splitting the cases. Similarly, the rod races are pressed in, however you won't want to remove them unless they are spalled or shot out too big for oversize rollers. With proper pucks and sleeves, This can even be done in a large vise. 

Let's all fill in the huge gaps I have left with your own techniques before we get on to the Prep and Inspection steps. 

From: indianjohn <johnmarg@pilot.infi.net> 
The only input I have is to be sure to remove the washer under the pinion gear nut (also, the nut is left hand thread). This is sometimes hard to see and I know from experience that it will screw up a perfectly good puller if left in place. When removing the main shafts, I back off the nuts about 1/2-1 turn and then use a brass drift and hammer to break the tapers loose. This keeps the shafts from going on a journey when they come loose. You seem to have the diaassembly phase pretty well covered in my opinion. 

From: "Marc Gunderson" <indian@beachaccess.com.au> 
I agree with the problem with all the rollers falling on the floor. This is the most painful part. So don't worry and use new ones. 

From: Keith <Packardv8@aol.com> 
There is no reason for this to occur from the drive shaft. Cut an 8 inch length of 1" PVC pipe. rotate crankcase assembly so it is resting on the right side.  Slide pvc tube over drive shaft against the bearings.  hold down on tube while removing left crankcase half and the rollers will stay on the shaft. Removal of flywheel assembly from right case is a little trickier. Slowly lift flywheel assembly from right case while kind'a wobbling the flywheel assembly back and forth. Doesnt always work for the right side '47 and earlier. 
  
Cotten wrote: 
> smack it. I find this method superior to the smack on the edge of the >wheel at 3 o'clock, since it takes less abuse, and does not leave hammer >dents, which even a copper mallet will leave on a Z wheel.>> 

USE ONLY A LEAD hammer. Grasp con-rods at the wrist pin joints together in one hand and lift to the 12o'clock position. Now slightly tilt the flywheel assembly back onto the edge of the right flywheel so that the left wheel is not touching the bench. Using a LEAD hammer strike the left wheel at 3 o'clock or 9 o'clock.  But do like Cotten says, always leave the retaining nuts loosened BUT ON, so that when it comes apart it wont go flying apart.  If no lead hammer is available you can use a piece of hard maple and a regular hammer, but this usually requires 2 people.  

From: "Marc Gunderson" <indian@beachaccess.com.au> 
Well Cotten, with regards to splitting the flywheels your method seems a bit toooooo rough. Undo the nut on the respective shaft and use a Copper hammer and hit the end of the shaft. Hit it squarely and woompa, it will release from the taper. And all the tools are shown in the back of the Army Manual. So anyone can make the required tools to do the job. 

From: "Cotten" <Liberty@npoint.net> 
I'm glad to get so many responses! My replies are not to be defensive, but perhaps to clarify: 

Marc wrote: 
> With regards to splitting the flywheels your method seems a bit toooooo >rough. 

I firmly believe a demonstration of my wedge method would show it uses much less hammer force (yet often it releases both ends of the crankpin from its tapers). The side-smack method as per the factory (and Keith's description) often requires a brute blow.  

>Undo the nut on the respective shaft and use a Copper hammer and hit the >end of the shaft. Hit it squarely and woompa, it will release from the >taper. 

Yes! But to hit squarely has become a rare skill in itself. (I found this out teaching mechanical engineering grad students.) This is why I suggest to 'drift' them out as per the manual, except I use only a short puck to free up my other hand (to catch the shaft before it joins the rollers on the floor). Certainly we expect that individuals who eclectically choose from this discussion will pay close attention to such details. 

From: "Marc Gunderson" <indian@beachaccess.com.au> 
With regards to the conrod races, you will need a special press tool to ensure correct removal and replacement, which is shown (in the manual). I made one and it was cheap and easy. 

From: Keith <Packardv8@aol.com> 
This is true.   BUT, it can be done in a vice; a press makes it easier, but a vice will do the job. 

From: "Cotten" <Liberty@npoint.net> 
The "pucks and sleeves" I referred to are the 'special press tools' (easily lathe-cut), and they only require a little more patience to be used in a vice. On female rods it is important to remember to support the individual web, and not squeeze the webs together. For this reason it is prudent to always press the races outward, and not toward the centerline. The alignment of the webs are delicate, and often will splay and skew if dropped, hammered on, or improperly supported when straightening the beam.