Changes between Initial Version and Version 1 of Ticket #23061, comment 26


Ignore:
Timestamp:
Aug 24, 2017, 10:13:25 PM (2 years ago)
Author:
catalyst
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #23061, comment 26

    initial v1  
    1 The potential problem is less about omitting rare small values and more about missing lots of representable intermediate values.  I think with this kind of approach we're losing half of the representable values between 0.25 and 0.5-2^-52^; 3/4 of the possible values between 0.125 and 0.25-2^-51^, etc., which was kind of the motivation for the Downey paper.  We might decide we don't care, but I think that's highly dependent on what the caller ends up doing with the results.  (We should also document this choice; the comment currently focuses on small rare values instead of "holes" in the range.)
     1The potential problem is less about omitting rare small values and more about missing lots of representable intermediate values.  I think with this kind of approach we're losing half of the representable values between 0.25 and 0.5-2^-54^; 3/4 of the possible values between 0.125 and 0.25-2^-55^, etc., which was kind of the motivation for the Downey paper.  We might decide we don't care, but I think that's highly dependent on what the caller ends up doing with the results.  (We should also document this choice; the comment currently focuses on small rare values instead of "holes" in the range.)
     2
     3edit: fix some exponents I miscomputed due to a sign error