When will we use a new programming language for aircrafts?

ivanotter

From South Africa with love
Joined
30 March 2012
Messages
4
Reaction score
0
Hello all,

I am new to this forum. I found it through alternatehistory.com.

I have followed this site for some time and found it very very interesting, also because of the people with a deep knowledge about things.

I have a little topic close to my heart, somehow:

It seems that ADA is used as the programming language for systems integration and also weapons integration.

Now, ADA is something that came with ark. Even I programmed in it very many moons ago.

Has the world not moved on? Isn't there something better? (I have read up on the ADA compilers, but even so!).

Maybe Microsoft is not the way forward in weapons integration, etc.

Is the usage of something not particular efficient causing the very very long development cycles and integration cycles?

Just a question.

Yours,

Ivan
 
1. ADA was designed specifically for mission-critical tasks. Building a new language is one thing, proving that it is suitable as a replacement for ADA is a lot more difficult. Replacing ADA will be expensive.

2. The type of application ADA is used for typically has a long life. So you'll need to keep ADA around anyway to support 'legacy' systems, and a company will have to support 2 languages (ADA and its successor) for a long time.

3. is ADA inefficient? Just because it is old, doesn't mean it is obsolete.
 
Yes, true. If you have to build a new language (I did compilers, many years ago and I am not in the IT industry anymore).

But do we not have anything around with better features and as suited to the task?

Support of legacy systems is a "b***ch", true. The South African payroll system is still on Adabas/Natural as far as I know. Maintenance contract was awarded some 3 years back at $35 million over 5 years (!).

That said, is there nothing else? And how much does the (old fashioned) ADA cost in time of integration? if any?

Ivan
 
I assume you already know of C++ usage for JSF. They of course have standards for it, I think even many games developers think of using them.

http://www.jsf.mil/downloads/down_documentation.htm
 
mz said:
I assume you already know of C++ usage for JSF. They of course have standards for it, I think even many games developers think of using them.

http://www.jsf.mil/downloads/down_documentation.htm

Given the mess that is the JSF software, that's not exactly a ringing endorsement.

Also, a while back, the US Navy started an effort to try and replace various ADA based systems with applications written primarily in C++, IIRC. It has not been a success, to put it mildly.

And then we come to the USN's by now infamous outsourced E-mail system, which was specifically intended to be non-ADA based. Enough said. [shudders]

The main complaint that I've heard about ADA seems to boil down to the fact that it won't let programmers take easy (but often dangerous) programming shortcuts in order to meet LOC quotas, deadlines, and the like, unlike C++, for example. If that is some people's definition of a 'not user friendly' and 'too complicated' programming language, well, that is their affair.
 
Bit late, but I haven't checked in here in a while


First point, it's Ada, not ADA. ADA is Air Defence Artillery. Ada, OTOH, is named for Ada Lovelace, Byron's daughter. Babbage's collaborator and arguably the first programmer.


Second point. People talking about using C in any variation for safety critical work doesn't give me the warm-fuzzies. I can't help thinking of the MISRA C guidelines for using C in the auto industry which IIRC start out something like: "We'd really prefer people not use C for safety critical work, but if you must..."


The aviation industry doesn't work to the same drivers as the rest of the software industry. What we need from software isn't the newest gimmick or the faster execution time, but absolute predictability and testability and the ability to show that the programme and the design are mathematically identical. Ada was specifically designed with those kinds of aim in mind, and Ada combined with design/programming by contract capabilities (such as SPARK) provides built in protection against common forms of error in other languages.


The long development cycles referred to by the OP don't have anything to do with Ada, the original 757RT flight control system technology demonstrator that preceded the 777 PFCS was developed as three parallel lanes of software/hardware implemented in three different languages: Ada, PL/M, C (IIRC); there was no difference in implementation times between the three. Long development times in aerospace and defence work result from the extreme safety critical nature of the work, that it tends to be at a level of technological risk far greater than would be tolerated in other industries, the sheer complexity of the requirements and the necessary imposition of a level of testing that far surpasses anything undertaken anywhere else in the more general software industry. (A difference which anyone with money saved with Nat West will have had cause to rue in the last week or so).


If your life depends on it, which would you rather use: something designed for the task, and designed to make most common errors impossible, or a generalist tool that goes out of its way to make anything possible....
 
When will we use a new programming language for aircrafts?

When will non-English-speaking members learn that the plural of "aircraft" is AIRCRAFT ?!?
 

Similar threads

Back
Top Bottom