I apologize as I know that this answer is likely in the manuals for writing an R package but in my reading and looking at other package\'s Collate fields I can\'t figure out 100
I think it comes from a time and situation when the order in which the files in the R/
directory were sourced (as opposed to the lexicographic default) still mattered. A similar reason for why the .onLoad()
function is often in a file named zzz.R
.
For what it is worth, not one of my packages uses the "Collate:" field in DESCRIPTION.
Edit: And as we're fond of empirical measures, here is a quick grep
on a machine with all CRAN sources:
$ grep ^Collate */DESCRIPTION | wc -l
292
$ ls -1 */DESCRIPTION | wc -l
3779
So that is 7.7% of all packages. I really would not worry about this unless you really must.