Well here we go again.

I will try to get back to what I was trying to say previously .
Rob, I cannot say too much about John design, I had one but I never try it in real world application.
If I remember right John had the toroid and 4 diodes . That design will work and is all in how you tune it.
Have to get the right ratio in between the primary, secondary and fly back."
On the secondary you want to produce as much voltage possible to be able to charge the capacitor at 10-12...etc K RPM's . The Fly back needs to get 1/2 voltage from the data sheet of the transistor you are using on the base of the transistor if you are at ...6v system. On the capacitor on "idle" no trigger signal you want to tune the system from bias resistor to stay 20% below the voltage rating... This will prevent the overcharging, internal discharge to ground and possible failure of the capacitor.when U first turn on and the engine is not running. I look on better choices for capacitors and they are several out there starting from $8 up to $32 that are at 500,600V but then you have to replace SCR's etc. ... you have to pick a voltage I will suggest use 6v as base and get the best performance then you can go fairly easy down to 4.8 or up to 6.6v without too many changes.
John's design will be capable to deliver the 10-12k RPM's ...but keep in mind that most airplanes engines
are in 8500-9500RPM range and to have that engine to scream at 10k with a smaller prop just to spin the prop 10k...my personal opinion is not quite beneficial. Go for the most tork , load the engine with a bigger prop/pitch combination to give you the best static trust . This way you prevent the wear and tear on the engine and get better fuel economy.
You want to hold the capacitor charged at a constant voltage at the target RPM. Run the system for at list 5 min and see what it does, does voltage drop, increase...Increase will be most wanted that way you will know that the capacitor is getting charged at target RPM and is room for more if you go Pass that as log that the charge voltage is enough to deliver enough energy(Rob's mJ) . Goant try it on the best engine you have and better to be on a factory airplane engine, not a chainsaw ,weedy...etc. An engine with High compression to make sure you have enough "Rob" to overcomethe compression and produce enough to ignite the cylinder. If at top RPM you do not get any sputter,cough...LOL, then u good. If you do...you in trouble....LOL...But go ahead and put a fresh plug and start again if that is still present then go ahead and bump the voltage bit more u and feel your transistor. If it gets hot...then may have to add a radiator or different transistor. The MJE521 is labeled NRD on mouser so someday will be discontinued.So try to find a better choice , Hint ...TO220,use it the way it is or cut the tab off and then send them back at mouser or digikey for a partial return....LOL.

Rob design I think could work better due to the oscillating design and then again...more components so
I see a constant battle on what design to use.
Build one each, get a one of the heli engines or a RC car, boat do a conversion to CDI and It will let you know what will work or not and then stick with that.

Try to keep it simple as possible , less components...the better. More sophisticated ....more chances for failure and remember the size. We are space and weight limited. I get numerous questions why cannotbe smaller....why it has to be the same for size as a 30-100cc when the engine is smaller. Why a 14CC engine cannot use a smaller CDI....and why it has to be the same as a 50cc. Go figure that and try to explain to someone that asks you that question ...why....I hope no one will ask me here that...GRRRRRRRRR....LOL. But in case someone is curious why... well the amount of energy necessary to ignite the fuel is the same on a small engine or big engine/compression. In order to produce that energy (mJ) that Rob talked about you will need the same components. Why cannot be a chip that will do that?
Yes they are some and I look in to that ...but the voltage that we useis not going to be enough to produce the energy we need. The dedicated ignition IC they also cannot be used in our case...so we use what we use and try to be as efficient as possible to make from nothing something that will get your engine running reliable on the air.
After several hundred units that I solder and play with I got to original Bill Carpenter words. People wants to fly to be recreational not to stay at the field or home programming the unit with this and that curve. 95% of the people they just want to flip the prop and go in the air. Couple years back I was discussing all those options from the current soft with MSD, Kill SW and couple others functions but all those will add to space, weight and $. When today you go on hobbyking andbuy a blue unit with few bucks....how are going to compete with that ...and then again when that goes out you do not have with whom to talk with so you buy again and then u get to the price where u can talk with someone and help you out one way or another.

But ....final words....get something , test it for a season on several types of engines, users and then u will find it is a POC or not (Proof Of Concept) like a good friend of mine like to use. If you test it in your back yard only....my not know and since is your design may not be that exigent with it.

Thanks and sorry for boring you out.

Adrian