Ross Ihaka ihaka@stat.auckland.ac.nz
Mon, 4 Jan 1999 13:20:25 +1300 (NZDT)

There is a possible solution to the "single" problem which I've been
thinking of introducing on the "next pass" through the memory
management code.  Since it is largely independent of the memory
management system I guess I should get to it rather sooner.

The solution is to introduce a new data type "single" (internally
SNGLSXP) which is identical to "double" (internally REALSXP) except
when it is passed through a .C or .Fortran call.  There the values
are copied to a genuine single precision array before being passed
to the foreign code.

As far as I know, the "single" type in S only exists for compatibility
with Old-S functions.  That's why there is no corresponding single
precision complex type -- Old-S did not have a complex type.
Would it be worth introducing a "singlecomplex" type to parallel the 
"single" real type?

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