I too taught in a B-school, and I heard the same sorts of comments, from both undergraduates and MBAs. There's an old saying here that "it's every soldier's right to complain" (in fact, I think that's in the Uniform Code of Military Justice). The analogy is that every student has the right to complain (and I have the right to ignore them). :-)
That said, I think the depth and rigor of quant courses should depend on the audience. I taught introductory LP (and in some cases a wee bit of IP) models, for instance. For a course aimed at students majoring in management science, I would go into a fair bit of depth (though not to the point of proving the Theorem of the Alternative). For generic business students, my goal was (a) to stimulate some quantitative thinking, (b) make them intelligent consumers of models (cooked up by consultants or their organization's OR group), including (c) teaching them something about the limitations of models, and (d) hopefully enable them to build some simple models for problems that they might reasonably be expected to solve by themselves (and that might plausibly enounter). That's pretty much in descending order of importance.
For the general business majors, I treated LP solvers, simulation programs etc. as black boxes. Insert the model, turn the crank, get the output, figure out what it's telling you. They don't need to know how to solve an LP; they just need to understand what unbounded and infeasible mean, how to interpret output, how to spot that alternative optima may exist (and why all optimal solutions are not equally good in practice).
Unfortunately, the least important of my goals (being able to write and solve LPs, simulation models, Markov chain or queueing models, ...) proves to be the easiest to test.
-------------------------------------------
Paul Rubin
Professor Emeritus
Michigan State University
East Lansing MI
-------------------------------------------