3 ms·
There is definitely an aspect of the 'the appropriate tool for the job'. We use Python extensively for 'glue' scripts on my project (build, deployment etc.). It
by rkangel 6y ago
There is definitely an aspect of the 'the appropriate tool for the job'. We use Python extensively for 'glue' scripts on my project (build, deployment etc.). It's wonderful and convenient for programs that are 10 to 1000 lines long. The result is clear and predictable and maintainable (particularly when compared to a bash script alternative).
I would never write a large production app in Python though (having done it a couple of times). The flexibility becomes unwieldy at that scale and you have to have tests that exercise every corner or you'll have problems hours or days into running.
- offtop5 6y agoThat's a very good way to look at it . With Rust do you need less Unit testing ?
- rkangel 6y agoYes. How much varies though. There are certain classes of test that take time in a dynamic language that you just don't have to write. They're unimportant 'just check I'm actually passing a list to this function and not a class' type tests rather than interesting algorithmic or behavioural tests.