Posted on 04/02/2020 9:41:53 AM PDT by dayglored
No I can’t. It has been to long. Some of the places I read have news that never hits the MSM. MY guess now is that maybe it was an internal leak. Obviously a problem like this does not go from being discovered to being in the news the next day. And once a probable problem has been identified then some time would be taken to verify it before issuing any directive.
If it’s Boeing, I ain’t going.
Not all the computers in your car power down. I know, wrote automotive S/W for transfer case, differential control... Knew a guy that did the same on another platform. He wrote his code to perform a detailed bit on power cycles. No one told him the module was always powered. Long story short, module never failed until owner replaced the battery. This was typically after the 3 year warranty was expired. Lol. Working with idiots led by other idiots with no time for requirements or system knowledge. Module did have a very low warranty rate so that was a plus to management.
What if you cross the International Date Line?..................
wind river for the real time OS.
imaginably, boeing for the cockpit instrumentation applications, since the cockpit instruments are boeing specific.
both imaginably would have to run 51+ day stress tests to find and fix the problems before the systems are released into the field. 51+ day stress tests might be very expensive in money and time costs.
Your thoughts on this?
“wind river for the real time OS.”
Ok it is Linux. I will have to install a copy in Hyper-v and take a look. You can run on PowerPC and ARM.
At wind river site now. Looks interesting. Systems have a came a long way since I last work on Avionics.
> Thats not a very good operating system. An op system must be robust and not do stupid things.
I am no vxworks fan but vxworks is used in dozens if not hundreds of similar products, and apparently the 51 day failures are boeing 787 specific. if your conjecture is correct then would the problem be noticable in other products in which vxworks is used.
> I was making a software presentation to Boeing engineering once and one of their uh, multicultural software people asked me, “What’s an opcode?”
OMG
Thats the best time to do it. :)
> Ok it is Linux.
I am not claiming that. I have no inside knowledge of current vxworks, but my impression received over time has been that it is a proprietary RTOS which is not linux based and contains differences from linux, internal and external. According to wind river, it implements a POSIX PSE52 interface.
I drive a bus so I wouldn’t know - however, if a Boeing plane needs updates and the electronics isn’t turned on every so often then there’d be a problem; just like those upgrades that automatically download to your PC.
Every time battery gets replaced or disconnected, have to reprogram all the radio stations!
There you go again. Thinking outside the box. Going to get you in trouble.
There you go again. Thinking outside the box. Going to get you in trouble. Make the bean counters mad.
https://en.wikipedia.org/wiki/FIFO_(computing_and_electronics)
https://en.wikipedia.org/wiki/Garbage_collection_(computer_science)
> I drive a bus so I wouldnt know - however, if a Boeing plane needs updates and the electronics isnt turned on every so often then thered be a problem; just like those upgrades that automatically download to your PC.
Actually it would probably be IMHO a security violation to attach such a system to the open internet, since it would allow access to the system from outside the intended means of communication. More likely, a software upgrade would be a controlled manual process with explicit checklist style instructions that must be manually checked off in sequence. Automation in this situation might mean less control over the system since at any given moment in time, the system is in an unknown (either updated, out of date, or in between) state for each software subsystem in the system. There might be dozens, hundreds or thousands of software subsystems, each verified to work with each other (or not), but always accompanied by version designations of the associated software and hardware environment in which it is certified to operate.
Within each hardware and software subsystem, there normally is a process that governs how changes are made to a subsystem to produce a future version. Usually that process is some form of ISO 9000. It has a test component which regulates how subsystem tests are created and changed (generally, a procedure analagous to the manner in which the subsystems themselves are changed).
All of this is usually very time consuming and expensive. It becomes very tempting to take short cuts. Perhaps the problems with the boeing 737 max are the result of some short cuts that were taken to avoid massive testing requirements. These short cuts were apparently taken during a recent period at which boeing stock was at (then) all time high prices, which possibly makes it so much the more embarrassing for boeing, especially at a time when they are looking for government handouts to prevent massive layoffs due to covid-19 problems and this problem. The presence of so many avionics software related problems on so many different products may indicate a high level organizational failure to recognize and properly manage avionics system software. Software management is IMHO often not unlike dodge citXXXXXXXXX an art form.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.