I was curious, are more skilled programmers better able to see "the matrix" when they code? In other words, are they able too see their code at a lower and more comprehensive manner than less skilled programmers? Since all code derives itself from assembler language, I was curious if more skilled programmers are better at seeing the 1's and 0's in their code.
More experienced programmers should have better analytical skills, which allows them to better understand the side effects of every line of code and to maintain a larger portion of the data flow in their minds as they code.
A deeper knowledge of the effects of the code on the processor is only occasionally desirable. High level languages rob of the developer any chance he has of controlling how their code will be processed. Compilers will make most of the choices concerning how the developer code will be translated to machine code. And many factors can contribute for the same piece of code in two different programs or two different places of the same program to be rendered to the processor differently. None more obvious than trying to compile the same program on two different machines. Compilers do operate entirely in a predictable manner, but the complexity is too high for a programmer to know how their code will compile.
So no. Hardly there will be a programmer that can see his code at the assembly level. The reason being he would be wrong most of the time.
That said, experienced and knowledgeable programmers do have a better generic understanding of the impact of their code. They cannot be precise, but can be more accurate in their predictions. An experienced and knowledgeable programmer will make better use of such things as the register keyword in C. But will not necessarily get it right all the time, or even 50% of the time.
In programming, the "matrix", as you call it, is not about a deeper understanding of the machine. The "matrix" is about a deeper understanding of the programming language. It's semantics and his capacity to maintain in his brain a larger portion of the code, which helps him mentally trace the side effects of a line of code through a larger portion of the source.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
