This was posted on a US site
That was on an SR22 Cirrus.
Now, there is nothing wrong with running “Windows” as such provided it is stripped down of just about everything that interfaces to the outside world with incoming events – like network connections and USB. Windows is terrible at handling hardware error conditions and it is easy e.g. to make a computer run at 0.1% of its normal speed just by having a faulty USB device plugged in.
There is an edition of Windows NT Embedded which is thus stripped down, but then there isn’t much to gain from using Windows in the first place. You do get a nice GUI development environment, but you can’t use the standard Windows GUI objects in the finished product because customers would just laugh.
From what I have seen, a lot of “glass” interconnects with RJ45 cabling. I would assume that might be ethernet but I suppose they might be using RJ45 connectors for ARINC 429.
I doubt it.
We design instrument clusters for cars, an example being the Maclaren P1 recently shown on Top Gear.
We would not use Windows for any critical product, and things like speedometers and other vehicle instruments are classed as critical. They also need to boot up quickly.
Aircraft instruments are even more critical.
IPL is the Intel Image Processing Library.
It seems strange to use Windows for such a project. Too much complexity. However, the NT kernel is solid technology. Unless this is a fake (which I think it may well be), it means certification requirements are a joke because if you had to demonstrate anything about system design at all, the complexity of NT (esp. with the GDI subsystem) would make that an impossible task.
I doubt it.
Doubt which part?
The pic was posted by the man who claims to have taken it.
Airbus’ glass cockpit is Windows and 486 based, and so are most glass cockpits and MFDs
Airbus’ glass cockpit is Windows and 486 based, and so are most glass cockpits and MFDs
Do you have any reference for this? I can’t quite believe it.
I seriously doubt you could certify Windows for anything with Design Assurance Level C or higher if DO-178 was to be taken seriously.
A lot of this stuff used to be based on Integrity-178B
What I was trying to say was that I doubt it’s possible to produce windows based primary flight instruments to a certifiable standard.
On reflection, it may be possible, but we found that other solutions were much more reliable and less costly to produce. We have a small but talented team of between 5 and 10 software engineers working on driver display systems. They do not favour windows. The chaps on our infotainment team (which is the satnav/radio/music bit) have different constraints and look at all sorts of things, including Windows Automotive, Android, Linux, etc
You won’t get “windows” certified, as you can’t prove it will work ok in all situations for any critical tasks as Tomjnx described. It just doesn’t meet DO-178b or the most recent DO-178c
I am sure the pic is not fake.
Also I cannot understand how anybody can logically prove that any program with more than about (say) 100 instructions will work in all situations.
i know it A) because i worked for Airbus and b) because i have two friends flying A320 to A340, and everybody in that scene knows the typical Windows “Blue Screen” that occasionally happens in Airbus cockpits
