2 ms·
Excellent response. My default recommendation is Ada + SPARK but I can easily recommend a C subset + tooling for those reasons. Especially on the compiler issue
by nickpsecurity 11y ago
Excellent response. My default recommendation is Ada + SPARK but I can easily recommend a C subset + tooling for those reasons. Especially on the compiler issue as it's known that Ada compilers aren't what they advertise themselves to be. The static analysis observation I'm not as sure about. A combo of SPARK, safe concurrency, the monitor concept w/ restarts or failovers, and one analysis tool should cover about anything they want. There's a quality vs quantity effect in Ada vs C static analysis where I can't be sure which wins here.
The ecosystem is really weak, though. There's no doubt about that. I don't see that changing regardless of whether Ada stays around. It's why I'm excited about Rust's mainstream adoption where it might get where Ada is or exceed it as tooling improves. It's one of few times mainstream goes for The Right Thing over Worse is Better. Eiffel, esp SCOOP model, being another good example that has significant momentum. The CompSci people are doing amazing things with SCOOP extensions right now that could be implemented in an Ada successor.
So, still potential at least. Plus, at least our tooling for C is now good enough to emulate Ada's safety with the right people doing the work. Plus, we only have to buy 5 tools now. Much improvement over old days. ;)
- notalaser 11y ago> It's why I'm excited about Rust's mainstream adoption where it might get where Ada is or exceed it as tooling improves. Me too! We had a great chance with Ada and we kindda wasted it. Rust is literally a second chance. This is pretty rare in tech.