We develop embedded systems almost exclusively in C. This is particularly true for 8 bit processors like the Atmel AVR and also 16 bit processors like the MSP430.
We also work with a range 32bit processors such as various ARM derivatives and these are generally also done in C.
Our specialty is design for low cost electronics manufacture where we do both the hardware and embedded software side and we can significantly reduce cost or improve performance by selecting to do certain functions in hardware or software. In general, moving functions into software saves manufacturing cost but only so long as it doesn't require excessive CPU grunt to handle.
An exception to coding everything in C is some Java Embedded work with the Micro Kernel achitecture such as you find with mobile IP devices like the Cinterion modules, http://www.cinterion.com , where they are set up to execute Java applications direct from FLASH.
As to how I got into embedded C, I began my engineering life as an analogue hardware engineer and got my start in embedded C because I joined a company that needed a C programmer. The deal was, learn C between now and when you start and the job is yours. So I bought a couple of books and learnt C in 5 weeks. Of course there was a lot of subtleties I didn't understand but I was able to start cutting code on day one.
So I am grateful that that was forced upon me because I have been coding in C now for 21 years and it is a very useful skill to have.
It is interesting that it continues to grow in the market place.
Much like the Moore's law, the number of computing units sold per year grows about an order of magnitude in 5 years time. And interestingly, the size of the system reduces by some constant (linear), I guess. Applications that go with all these small compute-engines are not large database applications but small well defined and targeted applications. These applications may not need support of OS like uCLinux or Win-CE but a simple development tool like that from Keil. Implication of this observation is that C will continue to be one of the popular choices in the immediate future.
Chrome is an interesting case, expecting the desktop to move to cloud, giving further boost to small compute-engine market. This will require new language strategies. I am not sure of Py core but if it is small like C (or Java), you have a possible winner candidate. Give me ten years for that.
