@Jan,

my practical experience was exactly opposite.

I like C, because reusing code on another CPU was many times easier.

I use lot of code almost unchanged on different architectures, e.g. key debouncing, UART-buffering, software RTC+calendar, 1-wire, I2C, CAN and many others on 8051, AVR, ARM.

Also I test often new algorithm with Borland-C before I use it on the 8051.

If you have difficulties to reuse code, I suspect your writing style was not modular.

Its not uncommon for me, that also small projects (AT89C4051) consist of 10 or more object files.

I see also no difference to translate big or small projects, both are many times easier under C.


I agree only in one point, translating code from big machines with gulping big memory down to 8-bit machines may be hard. But the reverse direction on originally written 8-bit code was easy.


Peter