Thu, 20 May 2010 10:06:17 -0700, Walter Bright wrote: > retard wrote: >> It's a rookie mistake to believe that languages have some kind of >> differences performance wise. > > Well, they do. It's also true that these performance differences can be > swamped by the quality of the implementation, and the ability of the > programmer. But that doesn't mean there are not inherent performance > differences due to the semantics the language requires. > > It's like car racing. The performance is a combination of 3 factors: > > 1. the 'formula' for the particular class you're racing in 2. the > quality of the construction of the car to that formula 3. the ability of > the driver > > It's simply wrong to measure the performance and then naively attribute > it to one of those three, pretending the other two are constant.
Of course. The language/implementation comparisons are all faulty. You also need to model the performance of the programmer by building some kind of developer skill profiles and measure how the languages & implementations compete against each other in all these skill classes. For example the language shooutout site favors experienced programmers; bad programmers generate code with 2-3 orders of magnitude worse performance.
