> it's faster than other scripting languages

That certainly depends on your use case. For instance, the launch time is ridiculously slow, so that you cannot realistically run a small matrix computation in julia from within a shell loop. It is better to use octave for that, where the startup time is almost negligible (just a bit slower than starting a subshell).

I didn't find it slow, but:

    $ time julia -e 'print(1)'
    1
    real    0m0.438s
    user    0m0.300s
    sys     0m0.118s
    $ time python -c 'print(1)'
    1

    real    0m0.040s
    user    0m0.036s
    sys     0m0.003s

That said, instanciating julia every step of a bash loop.. I think it requires a jvm mindset, warmup once and iterate inside rather than outside.

> I think it requires a jvm mindset, warmup once and iterate inside rather than outside.

I do not have this mindset then. I prefer tools who are mindset oblivious. They are really useful!

For example, imagine I have a collection of a few hundred images with their projection matrices (in text files). I want to crop them and apply a simple imagemagick operation (which is not available from inside julia). The elementary solution is to run a shell loop to apply the crop, and call julia to perform a simple adaptation of each projection matrix. This is impossible today: most of the running time of such loop is spent on julia initialization. Half a second to do nothing is simply unacceptable in a serious scripting language.

I guessed that it wasn't your habits but if you step outside that box a bit you'll realize that lifting up or down things inside loops is the most natural thing to do. And for you example (which may not be your real use or workflow) you could have two loops, independent or coupled through a queue so that the julia process is only started once.

Of course! Once I realize that after a few seconds only a dozen images have been processed I cut the loop, remember that the julia repl is dog slow, and then, rewrite the task in a different way. But I would prefer not to have to do that. Moreover, for more complicated examples, there may be data dependencies that make the loop commutation non trivial.

The slow startup time may be a minor inconvenience, I agree. But nonetheless it seems to be a case of sloppy engineering, as other scripting languages do not have this egregious problem. It sets a bad tone, and makes me wonder if there are other hidden monsters inside the interpreter, that may be solved by "you are holding it wrong" like this one.

You have to use it inside a REPL, like R or Matlab. Julia using LLVM and compiling on-the-fly is not sloppy design, it's a design tradeoff (and a great one, if you really need speed).

Julia is not a scripting language, it's a language for mathematical analysis.

Also going back to your example the solution would be to use ImageMagick.jl from Julia, and extend the library (4 extra lines with ccall) if it's not available, and send a pull request to help the community.

> Julia is not a scripting language, it's a language for mathematical analysis.

Ok, this clarifies the matter a lot. So julia is not intended to be a general-purpose programming language. Are you involved in julia development?

(Besides, I strongly dislike the design tradeoff of not being a good unix citizen.)

https://docs.julialang.org/en/latest/manual/running-external...

The only trade-off here is the restricted resources the Julia devs can afford. Focusing on the scientific computation niche, the Julia compiler was built to compile code just ahead of time, since this way it can create code as fast as compiled languages while still being dynamic like any interpreted language (such as Matlab and Python). But there is nothing in the language that prevents it from being fully interpreted (and therefore having minimal compile time overhead, but generating poorly optimized code for long running processes), or alternatively pre-compiling and caching to have both (which is being worked on: https://github.com/JuliaLang/PackageCompiler.jl )