Why make these parts reprogrammable if the mfg doesn't want them to be reused? If he wants them to be reused, he the MFG, and not the user, should bear the responsibility for keeping the doc's updated, in light of what's been learned. They, after all, are the ones who, first and foremost, should have learned it.

If you have "lab-stock" as many organizations do, it should be useable FOREVER, and supported until the end of time. If they "end-of-life" their parts, they should make that intention clear when the part hits the market.

One huge argument for IAP methods was field-fixes and calibration. If the part is no longer supported, how can that be done. Out here in the "real" world, we're required to keep machines and instruments that use MCU's calibrated and operational for 20-30 years. The development of a new MCU is no reason to replace it if it still works. Try convincing your client for whom you built 5 boards a decade ago that you need him to spend another $20K (inflation!) just to keep his machines fashionably updated. You'll find out why he keeps that piece of pipe next to his desk.

RE