Readit News[1] https://www.codeofhonor.com/blog/tough-times-on-the-road-to-...
While the tech described probably has merits, anyone who has been near DB lately would instinctively go: why are you doing this when you cannot get the basics right?
Ok, random question - looking at the pictures of a P-8A, I don't understand how it dropped anything while in flight, it's not like the pilot could chuck the packs out of their window. How does does that work? Are they equipped with a cargo door that can be opened in-flight?
Airbus also functions very much like a quasi governmental institution in many parts, so there's less interest in squeezing everything to death to save money.
Finally, Airbus generally has a KISS mindset, and are very conservative w.r.t change in engineering practice and tooling. When I was there we spent way, way, way, way, way more time testing than writing software - and the software was written in a way that any software engineer could walk off the street and understand it.
Oh, and quite low levels of outsourcing in critical software - they save that for things that don't have people's lives on the line.
I do understand that IP, for example using an actor's likeness via CGI+AI, is a very real issue that needs to be addressed.
The voice recorder overwrites itself on a two-hour loop. Two hours of voice data takes about a gigabyte of space at most. There is no technical barrier to right sizing this, and there is nothing special about the aerospace use case that prevents it.
Why would anyone think a two-hour buffer for something so critical would be appropriate? And why would it continue to overwrite itself after it’s grounded? Why is there no backup? Has it never been thought relevant to gather, say, an entire flights worth of data instead?
This highlights a complete failure on multiple levels and an inability to critically think about the problem space. How much time was spent implementing a system that under most circumstances where it would be needed would render itself entirely useless?