Closed IndrajeetPatil closed 8 months ago
It's also here:
It's been years since that happened, but as I recall, this is the history:
I believe it would be straightforward to remove the dependency on {hypergeo} given that the code already exists in {BayesFactor}, but I'm not sure it would be smart given that the same code will be in the CRAN {hypergeo}, and it is always good to keep code in one place; is there a problem with having the dependency?
Thanks, Richard, for providing a detailed account of the evolution of this segment of the code.
is there a problem with having the dependency?
Not at all. But I created the issue only because I wasn't aware that the JASP module was also using {hypergo}
, even if the R package isn't. It was just to avoid an unnecessary dependency.
I guess it would then make sense to urge the {hypergo}
author to update the CRAN version so that the same functionality is not duplicated in two separate places and {BayesFactor}
can remove its internal implementation in favour of {hypergo}
's.
Looking at the package source code, the only LOC where it is used is commented out:
https://github.com/richarddmorey/BayesFactor/blob/4165b16d05050ee24a134cbc8a2f0e95e196e4ca/pkg/BayesFactor/R/correlation-JASP.R#L23
So I am wondering if it needs to a hard dependency of
{BayesFactor}
?