In industry, I feel the triangle would come down to programmer time vs computation time. For quickly prototyping, we would go with Python. If we want to get the best performance from the hardware, we would pick C/C++.
Split_Personality_Computer
It seems like the easiest way to deal with productivity is to have more and more high level libraries added to your code. I'd say MATLAB is a good example of this, where pretty much most of what you'd like to do is already present as a built-in function.
Seeing as Python is written in C, why doesn't it allow for more parallelism? I don't see why it couldn't have a vector, thread, and GPU library to function as an APIs. Python will always be my favorite language and I don't see why people aren't trying to optimize its runtime. My belief is still that the ideal language would be C that looks like Python.
bob_sacamano
Can Go be considered as a language that lies between productivity and performance? Since Go provides concurrency primitives out-of-the-box like producer-consumer buffers and dynamic race detection, it's relatively simple to write scalable, bug-free distributed systems.
In industry, I feel the triangle would come down to programmer time vs computation time. For quickly prototyping, we would go with Python. If we want to get the best performance from the hardware, we would pick C/C++.
It seems like the easiest way to deal with productivity is to have more and more high level libraries added to your code. I'd say MATLAB is a good example of this, where pretty much most of what you'd like to do is already present as a built-in function.
Seeing as Python is written in C, why doesn't it allow for more parallelism? I don't see why it couldn't have a vector, thread, and GPU library to function as an APIs. Python will always be my favorite language and I don't see why people aren't trying to optimize its runtime. My belief is still that the ideal language would be C that looks like Python.
Can Go be considered as a language that lies between productivity and performance? Since Go provides concurrency primitives out-of-the-box like producer-consumer buffers and dynamic race detection, it's relatively simple to write scalable, bug-free distributed systems.