From: Adam B. Thompson <tngamer_at_gmail.com>

Date: Mon, 14 Jul 2008 10:39:57 -0400

Date: Mon, 14 Jul 2008 10:39:57 -0400

Hello,

Looking over the swap_tail macro in pnorm.c, the comment inside the code indicates it is to swap the values of '*cum' and '*ccum'. According to the code below that is taken from the source file, the swap functionality is dependent on what values are provided for 'x' and 'lower'.

#define swap_tail \ if (x > 0.) {/* swap ccum <--> cum */ \ temp = *cum; if(lower) *cum = *ccum; *ccum = temp; \}

It only swaps the corresponding values if and only if 'x' is greater than zero and 'lower' is "true" (non-zero). The following examples will show the actual behavior:

Input 1 Input 2

-------- --------

x: 1 x: 1

lower: 0 lower: 2

*cum: 1 *cum: 1

*ccum: 2 *ccum: 2

Result 1 Result 2

-------- --------

temp: 1 temp: 1

*cum: 1 *cum: 2

*ccum: 1 *ccum: 1

According to the code, Input 1 causes '*cum' to retain its value while '*ccum' loses its own and takes on that of '*cum', performing a half-swap. Input 2 performs the full swap, resulting in '*cum' taking on the original value of '*ccum' while '*ccum' takes on the original value of '*cum'.

Since the logic following the 'lower' conditional statement is not surrounded in braces, this causes the second statement (*ccum = temp;) to be executed even if 'lower' evaluates to "false" (zero). Is this an appropriate result for the macro since this does not actually result in a swapped pair of values, as indicated in the above example (Input 1)? Or should this only result in swapped values if and only if 'x' and 'lower' are greater than zero and non-zero, respectively?

Thank you for your assistance.

-- Adam B. Thompson http://www.cs.utk.edu/~athompso [[alternative HTML version deleted]] ______________________________________________ R-help_at_r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.Received on Mon 14 Jul 2008 - 14:54:35 GMT

Archive maintained by Robert King, hosted by
the discipline of
statistics at the
University of Newcastle,
Australia.

Archive generated by hypermail 2.2.0, at Mon 14 Jul 2008 - 15:01:30 GMT.

*
Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help.
Please read the posting
guide before posting to the list.
*