Learner Profiles/"Why" from R/python programmer perspective

[katrinabrock]

The current Episode 1 give a justification of why someone who has only or primarily used GUI programs in the past would want to start using shell for simple programming tasks.

In my experience, I've encountered folks who are quite confident in R, but hesitant to learn shell. A lot of the benefits listed in the current lesson can also be achieved by using R (or python or perl if you want to go there..). Would it be worthwhile to add a sentence or two specifically addressing why those who know or plan to learn other programming languages can benefit from learning shell as well?

Not sure if that kind of audience is part of your target. The "learner profiles" section of this lesson is a placeholder.

[bkmgit]

This is helpful. Perhaps the lesson could use more examples specific to the shell, for example running a task on a remote machine or using cron to automate a periodic task.

[omsai]

The shell introduces concepts that the R/python programming lessons do not:

• (Episode 2) Relative and absolute directories and tilde for the home directory.
• (Episodes 3 and 4) Creatively using commands beyond their prescribed names (e.g. mv for renaming, cat for printing a file).
• (Episode 4) The concept of pipes.

Many scientific software programs are CLI-only in languages not written in R or python. Shell skills enable one to use those programs. Using python's subprocess module and R's system / system2 functions for calling such programs can be tricky, especially if those programs expect pipes.

If people are using containers for reproducibility, those require shell commands. So does continuous integration.

running a task on a remote machine or using cron to automate a periodic task.

☝️ +1

I've said before I would not really recommend using the shell heavily, but it definitely has a place in the research computing toolbox. Once people have basic familiarity with the shell, they become better at recognizing when they need to fire up a shell to more efficiently do their work.

Besides the introduction episode, I think the instructor notes would be another place to motivate the lesson. The challenge with trying to a lot about the things that @bkmgit and I mention here is one would have to be careful about cognitive load. For example, I would not recommend talking about containers. One can talk about cron with an appropriate example or analogy. But accessing remote machines and effectively using programs not written in R/python may be some broadly useful motivations.

[katrinabrock]

Hmm...I interpreted the "Why use the shell?" section as reasons why someone might chose to learn shell at all, not reasons why someone would take the carpentries shell lesson compared to other carpentries lessons. So it's less about contrasting the content of the lesson with the content of other lessons and more about highlighting the tasks where shell excels.