I've read, replied to, and even initiated enough discussions on this and the original forum over a number of years that I really feel anxiety about even sticking my toe into this one. But so, just one possibly useful comment.
Forum readers, as has just been noted, come from many backgrounds. One of mine is as a consulting psychologist, and as I was reading these comments something from long ago was tugging at my mind. I finally recalled what it was and how it seems to me to fit here.
There is a very (to me) useful concept that has been around for fifty+ years, originally applied to worker-employer relationships, called the "psychological contract." In a nutshell, this refers to the implicit, not spelled out or even easily articulated, "contract" that exists between people. It is based on the normal but sometimes erroneous assumptions that develop in most relationships, along the lines of "if I do X, you will do Y." If we work hard, we will get paid bonuses. If I cook him supper, he will love me. If I buy a piece of software, it will be as easy to use as other tools I've bought before and it will always do X or Y.
Often, the big blowups in relationships of all sorts occur when one's sense of what is "supposed to happen" -- what one is "owed" perhaps (since people are mentioning the cost of Tinderbox), seem to have to been violated.
In terms of software, much of what we might consider a psychological contract is of course based on our experiences with other softwares, including the degree of complexity of the learning curve, what the thing should do and so on. Often, for me and I'm sure others, the expectation is for instance is that to do X or Y in Tinderbox, there "must be" some simple, direct way to do it... or some way based on stuff I've already learned from practice. Or at least, one single place where I can find the instructions.
Perhaps the typical "ratio" of expectation (based on using other tools) to easy success is unusually low with Tinderbox, at least for beyond very basic tasks. And therein, often, lies the problem - that's where people get frustrated.
There may be no easy solution. For many of us, we stick around because the payoff is still pretty high (eventually, at least.) But I do think that it can be a difficult process for many potential users to find that what they assumed was the "deal" they were making -- that I'll invest some time in this tool and so it will do something for me, was a "contract that seemed not to be honored". The human nervous system responds powerfully to the feeling of having a "contract" violated.
Finally, it seems to me that this cuts both ways in relationships. As many people, developers and forum members who put much effort into helping, must feel every bit as "cheated" ("violated" seemed too strong maybe) when their efforts to help seem not to be sufficient or appreciated. (Much the way people with severe depressions, often proving to be very hard to help, tend to frustrate their would be helpers.)
Don't know if this is at all useful, except that sometimes, in my squishy, mushy business, just finding another way to describe or articulate the problem proves helpful. Dominique Raynaud's nice example, for me (as usual!) rings true: it helps me personally to compare Tinderbox mastery to learning something as complex as a new language. You start with an easy phrasebook and learn to order coffee and find the bathrooms, but to really get around will take practice. The trick is deciding if you can accept, and have time for, that level of challenge.