Feedback on Chapter 5: Fundamental development workflows

jthomasmock commented 2 years ago

Generally, I think mentioning the RStudio Packages global options is worthwhile.

Also should the ? Developing packages - linked support article instead link out to this book? The support article is quite old and references R 3.1 😬

5.1.3:

I very much think that screenshots throughout are worth it, but if avoiding almost everything is possible the Command Palette (Cmd + Shift + P) - eg Command Palette > "Create new project" in this example, that can help with "stickiness" of ideas vs the GUI.

5.1.4:

Possibly add a note for directory structures like arrow or quarto, where the package is part of a larger project in other languages.

5.2.1

To see the most useful keyboard shortcuts, press Alt + Shift + K or use Help > Keyboard Shortcuts Help.

Beating the Command Palette dead horse again here as an option/callout 🐴

5.4.1

Repeat of the previous comment in another chapter, worth noting that RStudio/devtools? will throw warnings with namespace conflicts.

devtools::load_all()
ℹ Loading regexcite
Warning message:

── Conflicts ────────────────────── regexcite conflicts ──
✖ hello() masks regexcite::hello()

Did you accidentally source a file rather than using `load_all()`?
Run `rm(list = c("hello"))` to remove the conflicts.

jennybc commented 2 years ago

Perhaps there's scope for a general section on: pick an IDE that knows about R and learn how it facilitates package development workflows. In this book, we recommend and will describe RStudio.

This quote from WTF feels relevant (https://rstats.wtf/project-oriented-workflow.html#ide-support-for-projects), i.e. once you decide to start developing packages, you will really benefit from using a proper IDE for R and learning it well:

Projects are a common and very attractive feature of many IDEs (1.2). Again, the practice of organizing work in projects is not prevalent among long-time coders because they use an IDE. It’s the other way around: one of the attractions of an IDE is that it makes it easier to exploit development practices that have proven to be useful across many languages and domains

jennybc commented 1 year ago

I think through various revisions, especially those contributed by @ateucher, we've addressed this as much as we can.

hadley / r-pkgs