Re: realloc() - frequency, conditions, or experiences about relocation?

Malcolm McLean writes that, given the log-normal distribution
of file sizes with known parameters,

we can work out, given that a file is at least N
characters, what is the prbablity that an allocation of
any size will contain the whole file, and how many bytes,
on average will be wasted.

This is why I thought statisticians might help him: Malcolm
wants to find the aposteriori distribution of the size of a
file, after it has been found to exceed N bytes.  Am I right
that if we take the remaining (N>20) part of the density
function and re-normalise it, we shall obtain the desired
distribution?

My proposition was as follows:

  1.  Find quantile q0 corresponding to the buffer size
      currently requested.

  2.  Calculate new quantile q1 = 1-(1-q0)/k, where k>1 is
      an adjustable parameter, and use its corresponding
      value as the new allocation size.

For example, assuming for simplicity a uniform [0,20]
distribution of file sizez and k=2, a sequence of allocation
may look like this:

                requested allocated
                 2        20-(20- 2)/2 = 11
                12        20-(20-12)/2 = 16
                18        20-(20-18)/2 = 19
--
()  ascii ribbon campaign -- against html e-mail
/\  www.asciiribbon.org   -- against proprietary attachments