"Quick - name the most important personal computer of the late 1970s and early 1980s. Those of you who mentioned the legendary Apple II - that's fine. I respect your decision. Forced to think objectively in 2012, I may even agree. But if you just named Radio Shack's TRS-80, you made me smile. Your choice is entirely defensible. And back in the TRS-80's heyday, I not only would have agreed with it but would have vehemently opposed any other candidate."
To view parent comment, click here.
To read all comments associated with this story, please click here.
Member since:2006-10-18
Assembler can be optimized better than most HLL's, but it should also be said that modern optimizing compilers produce better machine code than all but the most elite of programmers could, with a fraction of the effort. At some point, you have to determine how much it will cost to achieve perfect optimization.
Member since:2005-07-08
With most modern processors (x86, x64, Itanium, ARM v8), this is no longer true.
The set of out-of-order execution conditions, parallel pipelines, types of SIMD instructions, multithreading, multiplecores, cache levels, NUMA memory is so big, that only genius would be able to do it for a specific processor/motherboard.
Assembly programming is only manageable by humans for the simple PIC processors, or some of the RISC models available in the market.
Member since:2005-07-06
Looking at what programmers are mostly out there, who we can afford, how much we are willing to pay them, asm is infeasible. Usually, better for the code to be high level (that's also the case with C++ vs .NET)
Remember, 80+% of drivers think they are above average, and 95% of programmers think they are in the top 5%.
Member since:
2011-05-12
I stopped coding in the BASIC/Assembler period. BASIC was slow, Assembler was fast. So each time anything but Assembler is used I get the feeling things aren't as effective/fast as they could be.
But I promisse I will start learning a (modern) language.