Re: Single/Double precision NA values

[Søren Hauberg]

tir, 29 07 2008 kl. 11:53 -0400, skrev John W. Eaton:
> Given the limited usage and the surprising, system dependent, and, to
> the end user, apparently random problems due to library functions,
> should we even bother to keep the current definition of NA?  Or should
> we just remove it and overload NaN to handle missing values?

Personally I use NA's quite a bit in my code, as I think "Not Available"
is quite different from "Not a Number". I don't really care about
compatibility with R (though it doesn't hurt). As to the "NA turning
into NaN in library functions", I don't think there much to do other
then document the issue.

Just consider this a users plea to keep NA,
Søren