We want to produce some content which motivates thinking "When should I write a standalone tool vs use existing tools".
e.g. if you can write a small sed script you may want to do that rather than write code.
But if you're processing a bunch of entries in a JSON file which requires keeping some state as you go, you maybe want to write a program rather than try to write a very complicated awk script.
I suggest we probably want to do this by coming up with a few examples of small data analysis exercises (e.g. doing something using the GitHub Issues API) and have them try to solve the exercises both ways, and come up with some reflections on what was easier/harder about each.
Revisiting some Piscine projects could be interesting here (e.g. Song Data).
We should be aware that the trainees will still be new to shell tools, and so things that are conceptually "easier" to do with shell tools may still feel very complicated because of lack of instinct/experience.
We want to produce some content which motivates thinking "When should I write a standalone tool vs use existing tools".
e.g. if you can write a small
sedscript you may want to do that rather than write code.But if you're processing a bunch of entries in a JSON file which requires keeping some state as you go, you maybe want to write a program rather than try to write a very complicated awk script.
I suggest we probably want to do this by coming up with a few examples of small data analysis exercises (e.g. doing something using the GitHub Issues API) and have them try to solve the exercises both ways, and come up with some reflections on what was easier/harder about each.
Revisiting some Piscine projects could be interesting here (e.g. Song Data).
We should be aware that the trainees will still be new to shell tools, and so things that are conceptually "easier" to do with shell tools may still feel very complicated because of lack of instinct/experience.