Minimum requirements for R package submission

Question

I'm a newbie to R package submission. I've written a program using basic functions inside R to manipulate phylogeny tree data. I've relied on the APE package. After almost year of working on it, it's time for me to submit the package and I have very little time to rewrite it as S3/4 style unless it is required.

It's just very basic at the moment with 30+ functions and there's a driver class. There's a lot of jargon in package submission and so it's difficult to understand the Google results. I would appreciate any help.

My functions are very basic. For example, getRoot gets the root of the current tree (APE phylo object) and getAncestor gets the ancestor to the current node:

getRoot <- function(cur_Tree){
    return(length(cur_Tree$tip.label)+1)
}

getAncestor <- function(cur_Node, cur_Tree){
    ...
return(ancestor)
}

1. Is this okay or do I have to do anything else to submit the package? Later on (within the next few months) I would have time to convert these functions to S3/4 but at the moment the most important thing it to get it out there on CRAN.

2. Does the vignette need to be written in latex or can I get away with writing all of the requirements in word? (I believe I've seen a vignette written in word -> pdf)

3. Any other suggestions/links?

Also, I think the R development team has done a phenomenal job with R and maintaining the package libraries. My intention is not to cut corners...it's just that I have a program written in R that is complete and I'd like to submit it. Furthermore, although github is a great resource for hosting code, my primary objective is to submit the package to CRAN.

Thanks!

Answer 1

To get a program on CRAN these are the main issues:

• have a package structure ( not just a bunch of .R files)
• pass R CMD check -- as-cran

Hint: The R package devtools is a really great help in developing packages.

The passing of R CMD check includes quite a lot of things. Like having documentation, ...

Actually the very first check is also not just the R CMD check. Somebody will briefly look over the package.

I remember that I was asked to write "time series" instead of time-series in my description...file.

But in general besides the formal issues, CRAN policy is not too strict.