Originally Posted by
Water pilot
It is quite possible that MCAS had nothing to do with either accident, but the existence of this poorly designed feature does point to a flawed "system design and roll-out strategy." I have debugged complex systems (both software and electrical) and quite frequently the close review of the system reveals design faults that unfortunately turn out not to have anything much to do with the actual problem that you were tasked to fix. MCAS is glaringly bad, the question that pilots and passengers should ask is "what else don't we know about?" That something else might be the real culprit.
For all we know, something else is going on -- I do not like the reports of data corruption that threw fault codes, attempted to be fixed by "cleaning the connectors." Random data presented to software can expose all sorts of bugs, including in the recording systems that we are using to try to analyze this situation. I worked with a former Boeing programmer (not on avionics) and his code was not exactly fault tolerant with respect to bad network packets; hardware/software engineers seem to have a mindset that precludes the possibility of bad data being presented to their code.
This is spoken like someone who understands the interaction between hardware and software. It's exactly what I wondered, as well. Sure, maybe it's as simple as altering the sensor inputs, but the input itself, wherever it comes from, has to be trusted all the time. If input is suspect, then output is too, and that's a cause for concern.