RNG R/Splus compatibility

[Paul Gilbert]

Here is a correction to one of the functions I posted yesterday, which did not
work correctly when kind was a list. With a null argument this function returns
current RNG and seed information. Otherwise, it sets the values and returns the
new settings. Thus, in a function one might

  old <- set.RNG()
  on.exit(set.RNG(old))
  effective <- set.RNG(new)   #(= new)

Another approach would be that used by par() (i.e. return the old settings) so
one would

  old <- set.RNG(new)
  on.exit(set.RNG(old))
  effective <- set.RNG()  #(= new)

Does anyone have thoughts on this?

Paul Gilbert
_________

    set.RNG <- function(kind=NULL, seed=NULL, rnorm=NULL)
      {# with a null argument this also serves as get.RNG so it is possible to
       # old <- set.RNG()
       # on.exit(set.RNG(old))
       # set.RNG("default")  can affect both kind and seed, but not rnorm
       if (is.list(kind))
         list(kind=RNGkind(kind$kind), seed=set.seed(kind$seed),
              rnorm= set.RNORMtransform(kind$rnorm) )
       else
         list(kind=RNGkind(kind), seed=set.seed(seed),
              rnorm=set.RNORMtransform(rnorm) )
      }


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._