Proposal: is.natural and is.whole functions to disambiguate is.count

[tripartio]

I recently discovered assertthat and I'm loving it. However, I ran into a little snag.

I find the name of the function is.count somewhat confusing. In my usage, a "count" should include the possibility that the count of x is zero, yet is.count(x) returns FALSE if x is zero. Perhaps I'm not thinking broadly enough, but I cannot readily think of any use case where it is impossible for a count to be zero. If such cases exist, I would expect them to be the exception rather than the rule.

Rather than quibbling on what properly or intuitively constitutes a "count", I propose the creation of two new functions that use well-defined mathematical terms to unambiguously specify the difference:

# TRUE if x is a natural number (positive integer, zero excluded)
is.natural <- function(x) {
  assertthat::is.count(x)
}

# TRUE if x is a whole number (non-negative integer, zero included)
is.whole <- function(x) {
  assertthat::is.count(x) || x == 0
}

Since assertthat::is.count is probably extensively used in existing packages, this proposal would not affect existing code yet it would probably help new users of assertthat avoid all the bugs I ran into because of my different understanding of what constitutes a "count"!

If this proposal is approved, I could write the code, test it with testthat, and then create a pull request.