Closed Ironholds closed 9 years ago
I'm already doing that: https://github.com/jeroenooms/openssl/blob/master/R/openssl.R
Yay!
The example is really cool too :-) It shows how you can use the openssl rng to replace rnorm
and rbinom
and stuff by using their cumulative distribution function. We could write about that in the paper to make it more statsy :D
Heh; it's the Journal of Statistical Software: mostly my argument is "using this code to generate SHA512 hashes of strings is 5 times faster than using digest to generate MD5s. Use it for cryptography and dataset releases". We could definitely look into dividing the writeup, though.
Okay; what would you think of me:
rand.R
file, since they're no longer the sole focus of the library;(obviously, each bit step by step, not all in one commit ;p)
I doubt that a (slightly faster) reimplementation of hashing functions is a sufficient innovative contribution for a JSS paper. You have a better chance of getting it published if we make the paper more general about functionality offered by libssl and libcrypto, and how it can benefit statistical software.
I renamed the old files: https://github.com/jeroenooms/openssl/commit/e3e1332083b0e0affa95791c8967cbe55b21c3a3. Make sure to pull so that we don't conflict :-)
Fair point! Thanks for the renames :).
initial tweaking commit. I'll close this issue for now and open more specific ones.
We should probably use roxygen2 for docs. I'll make a stab at adding that now.