I guess that people are getting bored with this topic but there is something I’d like to say.
Thanks for Leigh and others from illuminating discussions. Physical aspects have been thoroughly illuminated. But I’d like to address some didactic views. There is quite a lot of evidence that students do not reach conceptual understanding on electric circuits even though they might be able to solve problems using Kirchhoff’s laws. For instance, Professor Eric Mazur gives one convincing example from Harvard in his Peer Instruction Manual. Professor Arnold B. Arons discusses in depth difficulties and remedies in his marvelous book Teaching Introductory Physics. Just for the record: these professors are ”real” physicists.
My point is that without explicit teaching of conceptual analysis students simply will not learn it (of course there are always some exceptions). If you don’t believe this try out some exercises from Peer Instruction Manual! I have found out, like some research papers suggest, that use of these physically ”naive” assumptions gives the students an opportunity to practice physical concepts. After relationships between electric field, potential and current are understood I complete the physical aspects that are initially neglected. I always make the students to determine how resistance depends on current in case of a resistor and a light bulb. In the end (hopefully) the students can do reasoning recognizing how different assumptions affect the conclusion. Final exams involve reasoning and calculations. Nobody gets an A just by knowing how to solve equations.
I agree with Leigh that we must try to teach physics as a description of the real world. But I have noticed that to achieve this ”naive” assumptions may be needed in the beginning. Well, I'm willing to consider using resistors as an example instead of ligh bulbs; it would make the assumptions less naive.