The same could be said about CPAN and NPM. Yet Perl is basically dead and JavaScript isn't used for any machine learning tasks as far as I'm aware. WebAssembly did help bring a niche array of audio and video codecs to the ecosystem[1][2], something I'm yet to see from Python.
I don't use Python, but with what little exposure I've had to it at work, its overall sluggish performance and need to set up a dozen virtualenvs -- only to dockerize everything in cursed ways when deploying -- makes me wonder how or why people bother with it at all beyond some 5-line script. Then again, Perl used to be THE glue language in the past and mod_perl was as big as FastAPI, and Perl users would also point out how CPAN was unparalleled in breadth and depth. I wonder if Python will follow a similar fate as Perl. One can hope :-)
Python already has all those: the ctypes module is just as hard to use as WebAssembly, with a much lower barrier-to-entry.
* PythonCall.jl - https://github.com/cjdoris/PythonCall.jl
* NodeCall.jl - https://github.com/sunoru/NodeCall.j
* RCall.jl - https://github.com/JuliaInterop/RCall.jl
I tend to use Julia for most things and then just dip into another language’s ecosystem if I can’t find something to do the job and it’s too complex to build myself
// just fixed missing 'l' in link
During my PhD I was running some simulations using poorly written python code. initially it would take several hours. In that time i could go to the lab, run some wetlab experiments and the results of my simulations would be there when i got back to the office. It was only taking python "home" and building some of my own projects that i learned how to 1. write more pythonic code and 2. write more performant code. Now i work for a software company.
If i'd have stayed in in academia I would probably still be writing quick and dirty code and not worrying about the runtime because as a researcher there is always something else you can be doing.
Also, few scientific programmers have any notion of what C or Fortran is under the hood. Most are happy to stand on the shoulders of giants and do work with their specialized datasets. Which for the vast majority of researchers are not big data. If the one-time calculation takes 12 seconds instead of 0.1 seconds is not a problem worth optimizing.