[ntp:questions] Re: Tinker and tos configuration commands....

David L. Mills mills at udel.edu
Tue Feb 8 13:12:41 UTC 2005


No, signs of zero disappeared with the Univac 1108 and ones complement 
arithmetic and notwithstanding the Internet checksum. Zero is a 
legitimate state variable. Twos complement arithmetic is the sole 
survivor. Fudge values have nothing to do with it.


brad at shub-internet.org wrote:
> "David L. Mills" <mills at udel.edu> wrote:
>>One thing I try never to do is make a floating point variable into a
>>state variable, with the exception of zero. In other words, 500 - 1 LSB
>>is the same as 500 is the same sas 500 + 1 LSB. If you don't take that
>>care, rounding can cause real trouble.
> Even zero can be problematical in floating point, because there might be
> both positive and negative representations of zero.  When doing floating
> point comparisons, I learned long ago (in my Numerical Methods programming
> class while I was still in college, almost twenty years back), you always
> have to include a fudge factor to take care of the potential rounding
> error.
> In this case, the fudge factor is not being used at all within this part of
> the code, so rounding errors would dominate over either the exclusive or
> inclusive comparisons.
> You are right that this issue should be taken up within the NTP working
> group.

More information about the questions mailing list