[Bioc-devel] set.seed()

[Pariksheet Nanda]

Hi Luce,

 From the way you've described it, you don't need the latter.  Why is 
setting the seed in the function any different than setting it outside? 
Instead, you can mention in the documentation about setting the seed for 
reproducible output and also include that detail in the function 
example, which should likely have an example because it's an exported 
function.

For the first case of using a specific seed to match a publication, you 
could make a note of it in your package submission for the reviewer at 
the site of use and maybe also in a GitHub comment.  Some of the 
BiocCheck output is not strictly applied by reviewers; commonly, code 
formatting-related stuff is often left to personal preference.

Pariksheet


On 9/25/26 7:48 PM, Luigi Cerulo wrote:
> [You don't often get email from lcerulo using unisannio.it. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> We are preparing a new package submission. BiocCheck reports set.seed() at two sites. One is in a verbatim port of the frozen manuscript inference implementation, retained and regression-tested for exact method reproducibility. The other is in an exported simulation function with an explicit user-provided seed= argument; the caller RNG state is restored, and seed=NULL leaves RNG control to the caller. R CMD check is clean and all other BiocCheck warnings have been resolved. Would a documented justification for these two uses be acceptable for submission, or would you require their removal despite the reproducibility/API implications?
> Luce
> _______________________________________________
> Bioc-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel