Re:Toyota's firmware: Bad design and its consequences
by arink on Tuesday October 29, 2013 @11:50PM
(#6639364)
Primarily size. When you have 256k or less to work with, you don't have much choice.
Specific to Erlang: http://www.erlang.org/faq/implementations.html (see 8.9)
Re:Toyota's firmware: Bad design and its consequences
by Fr0styMatt on Wednesday October 30, 2013 @12:15AM
(#6639460)
Ah, thanks for the link.
It's easy to forget how small the RAM budgets still are in some parts of the embedded world.
Why does the industry use something like C? Even with the MISRA guidelines, it still seems like the wrong tool for the job when you have things out there like Erlang, which are designed for fault-tolerance and concurrency from the get go. I can imagine the requirement for hard real-time operation might be an issue (not sure enough about Erlang specifically to say if it does or doesn't address or hinder this, so I'll leave that to someone else).
What's that industry moving towards, software wise, to address this? Is Ada an alternative?
C (even MISRA) just seems like a horribly risky thing to use in this scenario.