On Oct 18, 2004, at 7:19 AM, Robert Klemme wrote:
> Do you think the 10000 makes it more convincing? :-)

Obviously! :)

No, I wasn't fully paying attention to this thread, but did see the 
talk about alternating rounding. I didn't know if it applied to the 
sprintf case, so, just in case it was a pseudo-random result I was 
checking if perhaps it varied between 0/1.

> Now, is this due to the Ruby version or rather the libC version (I 
> tested
> with cygwin)?

Dunno; I simply offer my data point for observation, and will let 
others figger out the discrepancy.