> By the same account, I expect your typical developer of a regular company, myself included, seniors included, etc, etc, to be completely lost if they had to optimize code for a given hardware and needed to write some assembly, or whatever, disassemble some JVM class files, etc.
I've done both. I've gone from high level C# to counting clock cycles for embedded, and then back again.
There is a transition period, but it is perfectly do-able. It is helpful if you have someone to show you the ropes, but after that it isn't too bad.
Heck I've met some teenagers who are into disassembling JVM/CLR stuff. (Which is actually, IMHO, much easier than raw assembly).
Have you ever done scaling calculations for AWS services? Same sort of logic applies to writing code for embedded, there is just a different number of mhz and instead of gigabytes of memory you are talking kilobytes of memory, but other than order of magnitudes, the reasoning is actually not that dissimilar.
I've done both. I've gone from high level C# to counting clock cycles for embedded, and then back again.
There is a transition period, but it is perfectly do-able. It is helpful if you have someone to show you the ropes, but after that it isn't too bad.
Heck I've met some teenagers who are into disassembling JVM/CLR stuff. (Which is actually, IMHO, much easier than raw assembly).
Have you ever done scaling calculations for AWS services? Same sort of logic applies to writing code for embedded, there is just a different number of mhz and instead of gigabytes of memory you are talking kilobytes of memory, but other than order of magnitudes, the reasoning is actually not that dissimilar.