| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
On 18 Mar 2012, at 21:53, Kreinovich, Vladik wrote: > OK, I seem to remember the argument: to split so that the number of floating point numbers between a and b is divided equally Yes, but the "length scale" is essential to cut out the bunching of very small FP numbers round zero. Study my output. Even better, run the Matlab. John P