Closed g4brielvs closed 4 months ago
Not entirely the same, but related to #19.
@bpstewar @BPJandree @kbjarkefur @randrescastaneda @tonyfujs If you have any ideas/suggestions, I would be very eager to brainstorm.
I have not used R this way so do not know what I would have to contribute. I know @arthur-shaw have tinkered with many different R set-ups.
A related feedback but not exactly on topic. I'm interested to hear what you think about having one one-size-fits-all template, or several templates with more targeted scopes.
The perspective I am coming from (which I know is only one of many perspectives) is that a template is most useful to the teams I am supporting if a template does not try to do everything. As that runs the risk of introducing confusion by adding a lot of files that for most projects never becomes relevant. A template can add a lot of value by pointing a team to what needs to be done/set-up. That aspect is lost a bit in the one-size-fits-all approach.
At the same time, for many teams it is not confusing to make as many tools as possible available in a single template. And then the one-size-fits-all is a great convivence.
The topic of this feedback is highly subjective. So I do not think there is a correct answer. Just interested to hear if you have thought of this and what your thinking is. Happy to move this discussion to a new issue if that is a better fit.
Out of the box, Jupyter Book supports notebooks running on IRkernel, however for the template to fully support R it'd be necessary to add the
r-base
dependency. Or exclude the R notebooks from the execution (as is by default today).Not sure if this should come as default or if it should be a decision on a project basis, especially because R projects may opt for a different documentation framework. For example, worldbank/pipr.