I think part of the problem with transistors is that there are so damn many of them that have been created since the early 1960's. This leads a novice to conclude that each one must have had a purpose and ideal application... so how on earth are they suppose to be able to pick the right one for their project. Picking the right one is about 1) knowing what your circuit is doing. 2) having a clear idea what you need and creating set of search parameters from those items.
In my mind there are only a few kinds of electronics personalities. (yeah, Ok, there are others, but lets stick with 3... and yes, this somewhat tongue in cheek)
- The hobbyist
They know enough to apply some basic concepts and turn them into reality with maybe some trial and error involved. Basic Skills, can sort of read a datasheet, understands most circuits and most components and can whip up a simple circuit without too much trouble or maybe using a reference like a book, magazine or datasheet. Probably has a soldering iron, some basic tools and a multi-meter. Most probably only follows design work other people have documented. How a transistor or FET works is eluding them but they will gladly take a transistor circuit design from someone and build it.
- The technician
Won't start a project without all the data sheets. Can create a circuit that is a variant of what is in the data sheet. Can repair a lot of electronic things without too much trouble or be able to know when to toss an item in the trash. Knows what transistor "beta" is and understands the concept that "MAX" data sheet values are not where you push your design towards. Understands basic math related to circuit design. Probably has a good soldering iron, tons of tools, multiple meters, home made test gear, a scope and also wants things like a spectrum analyzer.
- The engineer
Knows enough calculus to make the technicians head spin. Will gladly spend 4 hours figuring out how a 2 transistor oscillator works and can tell you, given random values, what frequency is that it operates at. Believes that spectrum analyzers are for sissies since you should be able to whip up a FFT computation and be done with it. Believes that if you don't do the math, you don't know Jack. Overkill.
So really, not everyone is going to be a #3 and a lot of people are already #1's. I think the goal here should be to shoot for the middle where you have the basics down and understood without the need to talk a whole lot "atomic structure" beyond some of the very basics.
Knowing WHEN to use a PNP or NPN or Darlington or FET and how to read data sheets are key goals to getting beyond step #1. In my opinion, most people messing around with Arduino should aiming for Step 2. At least that's where I think the answers should be aimed at.