Python or C++

Python is more suitable for one-off scripts and rapid development. It's very useful as "glue" to connect programs together, parsing inputs and the like. But for performance and scalability, C++ is the way to go. A compiled C++ program will often run 10x faster than a comparable Python program.
 
Both are general purpose languages, but whichever of them has the best-fit libraries, your personal preference and wether you anticipate requiring optimal execution (then C might be even more applicable). Have landed on Golang myself, but use float64 as I don't want slowdowns and can live with the inevitable rounding and inequality-issues. Not CPAN-scale libraries either, but my decision is to delay structuring too much as long as I can stand it.

All else being equal, whatever makes you most happy, will make you more productive in the longer run.
 
C++ (less Java) for HFT and performance applications, Python/Matlab/R for data science.

"So, while its true that at the moment most trading firms will probably conduct any serious research in Matlab (or R) and then implement trading algos in C++ or Java (bear in mind these are usually two different jobs in firms of any significant size), I would recommend a good working Python knowledge to any aspiring quant trader as a useful skill." - source https://www.quantopian.com/posts/programming-language

It's all on Google, but you need to use the keyword quant to get the correct information.
 
Python for faster deployment but slower execution. With Pandas, Numpy and Numba (or any other JIT C compiler), you can catch up to C++ speed wise. For GUI development I use PyQt, pySide being a good almost identical alternative for licensing issues. I find the Python modules list excellent, if you can imagine it, it's usually out there already.
I have extremely limited experienced with C+ though.
 
You must log in or register to reply here.
Back
Top