Although I think what is most fortunate is that people listen to you when you suggest that the HW be fixed.
thanx
VP types tend to think that expensive HW fixes (turning the board) are more easily covered up with firmware band-aids. This is not so much a problem at my current position, but I have live through it...

So have I. The problem is that a band-aid (HW fixed in SW) typically will reduce the problem rather than fix it. I do recall one such where delaying the read if a pin to the max the rest of the project could stand reduced a "frequent" problem to "rarely". I was not happy, but you do (most of the time) listen to the one whose signature is on the paycheck. The problem in a case like this is that when the white collar types see it reduced, they always say "reduce it more" and "that's impossible" is the equivalent to silence to these guys ears.

Another occasion, while developing a small market project with 48 identical boards (all required for test) I kept hammering "I can not make it work with these boards" (not My design) and kept getting the reply "we can not afoord 48 new boards". Well after 2 months of wasted time mgmt gave in and 3 days after the new boards arrived, the product prototype worked.

Now, I have a story where hardware was actually "fixed" by software. This, however, was not per se a fix, but due to the original developer not understanding the hardware and stating that the good hardware was bad. When software was developed that used the HW as intended all started working.

Thus I say with conviction hardware should be fixed with hardware, software with software. Example 1 above show that doing anything else may alleviate the problem, but never fix it.

Erik