Everybody has their own reality, and can realize others' only in approximation :). I agree with all your points, but not that they're universal. The scenario you have IS more likely to apply, but there's edge cases.

For the optimization project I mentioned, the chip cost was probably around $10K per 8051 (plain-vanilla no-flash UT69RH051 Radiation-Hardened MicroController: http://ams.aeroflex.com/pagesprod...rh051.pdf). No change in hardware allowed by the project, and external PROM fixed at 32K. Targeted for $250M-per-unit satellites (and launched on at least a few), so they didn't care about optimization costs, just that it was done in spec and on schedule.

For the OP's project, maybe they also aren't flexible on hardware. The controller might not be the only thing they have to squeeze into the PROM. Or they might want to use this as an addition to a lab course, and the dev boards are already in the lab. I think it may be better to teach at least some part of digital systems labs in assembly, there's less distance between the hex code you're stepping through and the source language.

Re the challenge, I wouldn't want to get into it, but I have heard of a solar array controller board being produced on contract by a team of 5 in China for $10K in a month, including prototype and full documentation set, so I wouldn't want to say this conversion couldn't be outsourced overseas for $200. Maybe someone could even be found to do it in the USA, on sologig, etc., for that price. Not saying I'd consider it fair payment (in the USA), but some might accept it.