Hmmm. Smells like avionics or process control.
I was around when "software reuse" became the big buzzword and Ada was introduced as a measure (in part) to reduce software development cost in increasingly computerized military systems. It didn't really work at the time, but it felt good to a lot of manager types.
The problem is identifying what is the difference between hard real time requirements (and I only define "hard" as in if you miss taking action and someone will get killed or injured or something is damaged or destroyed, but that's not the standard definition) and just keeping the system usable.
Technology cannot eliminate the difference in requirements for different applications.
I have to agree, much as I am in love with the idea of software reuse. However, I would like to s/cannot eliminate/has not yet eliminated/. I'm an optimist. I may not live to solve it, but perhaps one of my sons will. One thing that my reading in history has convinced me of is "never misunderestimate the power of the human mind".
I’m in aerospace software.
I have to admit to being an Ada-phile. I like the language a lot more than C++, and find it very, very powerful, without carrying the inherent danger of blowing one’s rhetorical foot off! ;-P
Oh well...
Y’all have a great evening.
I have no doubt about the power of the human mind, but my comment has more to do with the plain fact that hard real-time embedded systems are driven by their requirements, which are expression of the functions performed by the system. If the systems didn't necessarily perform different functions, then one would likely avoid the cost of developing a separate, dissimilar system unless there were some other compelling reason to do so. Most businessmen don't purposefully waste money, so...
The power of reuse is embodied by the essence of object-oriented programming, where hierarchical construction allows for overlays on top of basic functionality, lending the ability to specialize a generic object. Were it not for the inherent dead and deactivated code (not to mention the compiler-invented subroutines outside the view of mortal men), this would be a perfect solution to most applications - once embedded computers get fast enough to absorb the bloat.
Of course, in aero the life span of a given system is measured in decades, meaning some of the processors still out there (and being upgraded) may be pre-1980 technology! It kind of makes the "processing power" tsunami into a bathtub wave - not much impact! LOL