clarify docs about NA, NaN, and numeric

[avehtari]

Originally posted to Mastodon my learning experience about NaN and NA

TIL

doc forNaN and is.nan() says: ‘NaN’ means ‘Not a Number’ and NA, ‘_Not Available_’ which is not a number as well... ->

1. so neither is "a number".

> is.nan(c(NaN, NA))
[1]  TRUE FALSE

2. So NaN is not "a number" and NA is not "not a number"

doc for is.numeric() doesn't mention NA nor NaN, but says ‘is.numeric’ is a more general test of an object being interpretable as numbers.

> c(is.numeric(NaN), is.numeric(NA))
[1]  TRUE FALSE

3. So NaN is "a number", and NA is not "a number"

doc for is.na() says indicates which elements are missing.

> is.na(c(NaN, NA))
[1] TRUE TRUE

4. So both are missing.

I feel like the documentation could be a bit more clear on these, and here is a proposal.

[bbolker]

once you've settled on a version you like, the best thing to do is to create a patch against the current SVN version and post it on https://bugs.r-project.org/ (I'm happy to do this if you like)

[avehtari]

@bbolker thanks for the review and fixes! I'll try submitting a patch

[avehtari]

@bbolker ah, I see I would need to apply for an account to be able to submit. I assume you already have an account, so please go ahead and submit it

[bbolker]

Will work on this when I get a chance.

[bbolker]

https://bugs.r-project.org/show_bug.cgi?id=18691

[avehtari]

Thanks!