If you could not get it done in a "reasonable time frame"...Say under 10 minutes...how are we going to teach / train anyone to get there without losing them in the minute details required to produce your images?
I think the answer to that is to provide sample images that need just one correction (or at least instruct the participants to only concentrate on one correction). If the brief is "make the plants in this image look healthier without disturbing any other part of the image" then that might be a 10 minute job. If you present half a dozen such challenges, then introduce an image that requires each of the 6 steps in order as a final challenge, people would probably stay with it.
I am often asked, "How did you even think of that?"
I think that relates to the rustiness I was talking about earlier in the thread. Using CM comes down to assessment of what needs fixing, followed by applying techniques that fix the problems. Although you can teach techniques, they can be hard to remember (Chris isn't the only one who struggles there!). A Curvemeister crib sheet might be a good idea. Ultimately you need to instill in people the willingness to experiment, see it all go wrong, then back out and try something else.
I think that was the most important point I gleaned from the interactive Skype session we did some months ago: on the boards users only see the solution, and it gives the impression that the guide knew which magic bullet to use. In practice the guide will try lots of techniques until he finds the one that works. Once there, he will fine tune the technique towards the end goal. Getting people familiar with an arsenal of techniques is step one. Step two is getting people comfortable with the idea of doing it wrong over and over until something comes right. Step three is getting people to home in on the right technique as quickly as possible. I think experience is the only guaranteed answer to that last one.
So push, push and then push again to get people to do the interactive session! Use new tricky images, or, if necessary, "forget" how to do ones you've seen before and demonstrate the iterative nature of the challenge.