Closed mdingemanse closed 2 months ago
I am not opposed. It is a very tiny, simple package which does only one thing, namely provide a (so far) text-only version of IFADV as processed by scikit-talk
. Here's what CRAN says:
CRAN hosts packages of publication quality and is not a development platform. A package’s contribution has to be non-trivial.
I guess the non-trivial part is debatable. However, there are precedents for data packages, e.g., igraph
https://cran.rstudio.com/web/packages/igraphdata/ though the level of documentation and utility of that one is not something that ifadv will approach.
I don't love the extra work (and redundant work) it likely requires; was quite happy with the soft release on github, also because as academics we're not necessarily looking forward to maintaining a whole slew of packages.
(Let's not forget that we're only running into this because of superficial clash between the original IFADV license and the licensing requirements of escience center. If that were not an issue, we would have redistributed a portion of IFADV data with talkr
for demonstration purposes, etc., which was the original idea.)
Under these circumstances, I suggest refactoring the way we access ifadv
into a simple download. It's uglier, but it will solve the problem in a satisfying way both for CRAN and for us. If you agree, I can do it.
Agreed!
Let's close this option, at least for now.
I'm working on a fix on #117.
TL;DR:
Remotes
are not allowed in CRAN.Consider publishing
ifadv
package on CRAN too. Would that make sense?Originally posted by @PabRod in https://github.com/elpaco-escience/talkr/issues/117#issuecomment-2312288338