ada is not dead , it even gained in popularity

A prof of mine at U participated at Ada's development. To me its real time capabilities are very interesting. But at the moment to me this was more kind of a hobby. Because at the moment I have no project to have any use for real time. Any real time I did in the past (and I'm talking real real time, not "it's very quick, so come on, some tiny lags don't matter") I did so far on 8- & 16bit µCs - not really capable to crosscompile Ada on to it (I think).
However, AFAIK where Ada is used is at ESA, and (I think) in nuclear power plants. So it could be a job opportunity if some one knows Ada.
 
Well, defense spending is up around the world, and much of the gear needs lots of software. No wonder that Ada gets more popular (if it indeed is).
 
I've used Ada a long time ago. Like every other language, pros and cons; good when used correctly, but bad if you do things wrong. Not sure if it's been fixed but you could run into similar circular dependencies if you don't pay attention to "interface" vs "implementation"
If you have good design, the language almost doesn't matter.
When dealing with Government contracts, sometimes your choice of languages is limited to "one".
 
"Real-Time Executive for Multiprocessor Systems, formerly Real-Time Executive for Missile Systems, and then Real-Time Executive for Military Systems" (Wikipedia) - you are allowed to talk about it? 😅
"Initial release 1993 - Latest release 6.1 / January 22, 2025" It's for sure not dead. And if it was for missiles, it's for sure fast enough to fulfil some hardcore real time demands. :cool:
WOW! 🤩 It comes on a gitlab-repo, supports a buttload of platforms, seem to have some FreeBSD relations, and is POSIX!! very cool thing! *BOOKMARKED*
Thanks for the tip! Very interesting indeed, even if you don't want to build a missile.😁
I will definitely have a closer look at it.
(As I always say: This forum's a gold mine.)👍
 
you are allowed to talk about it?
A lot of times tech people look down on DoD work, but those who have worked in that space will think "Cool app bro, but have you written software for a tank, a submarine, a fighter plane? How about processing radar pulses to make a guess as to what that is?"

Hardware constraints , redundancy (fly by wire planes), documenting every little thing.
 
Hardware constraints , redundancy (fly by wire planes), documenting every little thing.
And don't forget all the approval procedures, strict regulations, and QM ain't not just a paper tiger, but done hardcore.

Automotive is whining for years why they are not allowed to have stear by wire like airplanes, which have it for decades. For the very reasons those monkeys would whine even louder, when they see what's needed to fulfill the same regulations to gain the same reliability. Their development costs would explode, they not only can dump all their way too short but clipped anyway underdo project schedules, and release tests after market release QM outsmarting, and this "hey look: I found a digipoti for half the price. I think that may fit. Let's replace the one in the BOM with this one!" *CRASH* "We are so very sorry. This rare exception case will be investigated very accurate, to find and punish those responsible bastards, while we already know, that's not the executives' fault. And we promise we will produce even more quality regulations, which is completely pointless, because all this shit could have been prevented if we simply sticked to our own regulations we already have in the first place."
 
Automotive is whining for years why they are not allowed to have stear by wire like airplanes, which have it for decades. For the very reasons those monkeys would whine even louder, when they see what's needed to fulfill the same regulations to gain the same reliability.
Exactly.
On one hand steer by wire for cars should be easier because cars only move on X and Y axis, planes add Z, velocity of cars are orders of magnitude lower than planes so decisions have more time.
But the flip side is planes don't fly in formation the way autos do on the highways, so autos need to look around a lot more.

QA: very true. Even when looking at code and the fix is "change i++ to i--" review is not trivial.
Subject automotive to the same processes as military and even commercial airplanes and all the people will start complaining about "losing velocity/synergy/whatever the heck buzzword the disruptors are using today" :)
 
And don't forget all the approval procedures, strict regulations, and QM ain't not just a paper tiger, but done hardcore.
Sure. You can't fly to the curb and wait for the tow plane, so to speak. The RTEMS part was for some hardware you definitely don't want to malfunction or to come back asking for directions. Autonomous swarm navigation was a thing a long time ago.
QA: very true. Even when looking at code and the fix is "change i++ to i--" review is not trivial.
Been there, done that, spent a day running around doing test cases for bad outcome, good outcome, have the code reviewed, the tests reviewed, all added to CI, verified the results, have that reviewed, ...

When a friend mentioned the relevant norms for drive-by-wire to people at the BMW booth at a large fair, they looked like they wanted to sick security on him. That stuff will take a long way to come to a road near you (at least here). And certification will be a complete CF for the first few attempts.
 
Sure. You can't fly to the curb and wait for the tow plane, so to speak. The RTEMS part was for some hardware you definitely don't want to malfunction or to come back asking for directions. Autonomous swarm navigation was a thing a long time ago.

Been there, done that, spent a day running around doing test cases for bad outcome, good outcome, have the code reviewed, the tests reviewed, all added to CI, verified the results, have that reviewed, ...

When a friend mentioned the relevant norms for drive-by-wire to people at the BMW booth at a large fair, they looked like they wanted to sick security on him. That stuff will take a long way to come to a road near you (at least here). And certification will be a complete CF for the first few attempts.
Imagine that. Old dogs may actually have valuable data/experience :)
 
I did a build of current RTEMS these days to evaluate it for $JOB, and it is interesting what happened in the last years. Thanks for making me dig it out of my swapped out memory areas 😎
 
Back
Top