The problem is not what you can and what you can't do with any specific language. The problem is how in hell do you prove to the authorities that you haven't done anything that is forbidden. In that manner, there is no language better to do that than ADA. Personally, I would never fly in a plane which had its fly-by-wire system written in Java, or any managed language. When lifes are at stake, one should stick with what is secure and proven.

Besides that, doing "Modified Condition/Decision Coverage(MC/DC)" is already a pain in the ass with structured programming, imagine doing it in OOP. Stack overflow is also hard to guarantee in OOP.

Paul Sanders (AlpineSoft) wrote:

Yes, I can see the sense of that, although I would have thought that embedded systems built on modern hardware (and I know that not all are) would have ample memory. Seems to me actually that there's an argument here for a language like C# (or Java), with garbage collection and no 'free' function, as the chances of a potentially fatal memory leak are reduced. Hard to provide realtime guarantees with such a language though since GC can happen any old time. Paul Sanders http://www.alpinesoft.co.uk

There is a norm called DO 178-B that says so. OOP is not safe enough to work unless you be extremely cautious about it, and very extremely cautious. So to prevent errors and, therefore, fatal accidents, OOP is forbidden to this date. A new review of the norm, the DO 178-C, is said to permit OOP under extreme controlled conditions. So, when you need to prove that your SW is bug-free, you just avoid at all costs to use OOP. Besides not using OOP, it is becoming more and more recommended the use of formal methods to help the proof of bug-free SW.

Of course I do. Work with really critical systems, such as avionics, does not allow you to use object oriented programming. So C and ADA are a must.