A320 family and A330/A340 have mechanical backup to the rudder and trimmable horizontal stabilizer.
A380 and A350 don’t have mechanical backups but have more redundant electro-hydraulical actuators.
Peter wrote:
…Pascal, which (I did it at univ, and even one of my firm’s comms products is a protocol converter with a Pascal compiler) by definition cannot contain any bugs, because it takes an infinite amount of time to write code which does anything
You must have had taken very strange Comp. Sci. courses at the university.
Airborne_Again wrote:
You must have had taken very strange Comp. Sci. courses at the university.
I also learned Pascal at the university. Turbo Pascal in 1987-88, it was the standard at that time. I have never used it since. Today it’s mostly LabVIEW for me, with some C and even FORTAN now and then 
I dunno, I studied a Pascal derivative (Oberon-2) in compilers course at one of the UKs better universities around 2007. Nothing wrong with it as a language, it’s just fallen out of fashion and doesn’t have the ecosystem that the more popular systems do.
Come to think of it. The flight control system in the Gripen fighter was also programmed in Pascal, at least originally.
Powerful aircraft !
LeSving wrote:
I also learned Pascal at the university.
I didn’t write that Pascal was strange. It is the concept that Pascal programs “by definition cannot contain any bugs” which seems exceedingly strange to me.
Lots of bugs when writing Pascal code as I remember. I think the point with Pascal was (at the time), it forces you to write structured code. It probably still do for all I know.
I was obviously pulling your leg when I said they cannot contain bugs! There is a smiley there!
Pascal is just a hard language to generate lots of code (especially efficient code) in, due to strong typing and a general lack of useful constructs. It was a favourite teaching language for these reasons. The lack of “ecosystem” is an understatement, and the tools all became a dead end. In the embedded world it was probably IAR (with spectacularly inefficient implementations of sscanf and sprintf etc etc) with the last one, and Borland (Delphi) on the IBM PC / 80×86 CPU.
But as always somebody who is brilliant in any particular (maybe unsuitable) language will do a better job than an amateur in a “more suitable” language.
The problem with teaching Pascal is that it produced loads of people who could write the Erastothenes Sieve (and other useless code examples) in Pascal but could not develop a working microprocessor-based product which, like so much taught at univ, was pretty useless in real life. They should have been teaching how to use assembler (for the startup code) and C – then and today, 40 years later, equally.
However I think the language is far from the hardest thing in getting GA avionics right. Most designers struggle to do even basic analog stuff right. The good ones have long retired and few new ones are coming in. Look at the QA / reliability issues which e.g. Aspen and Avidyne had, with really very basic electronics. In the certified world that just gets you a bad reputation; in the uncertified world it seems acceptable but that doesn’t make it good enough.
