[Question]: Standardize WORDLIST sorting using `radix`

[Bai-Li-NOAA]

What is your question?

I noticed that FIMS is using radix sorting for the WORDLIST in dev. This is a strict ASCII/C-style sort and doesn't mix cases. Do we want to standardize on the radix style? If so, we should update the instructions in .github/workflows/spell-check.yml and ask users to sort words using
sorted_words <- sort(unique(all_words), method = "radix")

Example:

words <- c("apple", "Banana", "banana", "Apple")

# Default sort (locale-dependent)
sort(words)
#> "apple" "Apple" "banana" "Banana"

# Radix sort (strict ASCII / C-style, case-sensitive)
sort(words, method = "radix")
#> "Apple" "Banana" "apple" "banana"

Additional context

No response

[kellijohnson-NOAA]

Honestly, I find the radix sorting annoying because it doesn't group like words together, which makes it harder when you are adding words.

[k-doering-NOAA]

Thanks for the suggestion @Bai-Li-NOAA ! Is the thought behind using radix that it doesn't depend on your locale, and thus the sorting will be the same regardless of where you are working and who you are working with? Or is there another problem it solves that I am missing?

I agree with @kellijohnson-NOAA , though, that the fact that the same words aren't grouped together makes it difficult for humans to parse. I wonder if there is some other option that isn't locale dependent and is more human-readable?

[Bai-Li-NOAA]

I don’t have a preference on which method we use, but I think it would be good to stay consistent with the {spelling} package and update the WORDLIST using R code rather than manually, especially for a large number of words. The default method for spelling::update_worklist() is radix, so I hope to use the same method in our workflow to avoid conflicts when merging.

If we want to sort words differently, maybe we can open an issue on {spelling} and address it in spelling::update_worklist(). Then we can update our GHA suggestions to follow the same method.

[Bai-Li-NOAA]

There are two {spelling} issues related to this discussion: ropensci/spelling#48 and ropensci/spelling#87

[iantaylor-NOAA]

After reading the issues that @Bai-Li-NOAA linked, and seeing that {spelling} has radix sorting hardwired at the moment, I like the idea of this package following that approach. I also find radix annoying, but even more annoying would be something that doesn't match the commonly-used standard package.

[k-doering-NOAA]

related to #226

[k-doering-NOAA]

Like @iantaylor-NOAA , I also think it probably does make sense to standardize on the same sorting that {spelling} is using.