version 0.002; 2006-08-05

  * bugfix: rand_flt() with subnormal limits was horribly inefficient

  * in rand(), don't attempt to generate a 48-bit fraction on systems
    where it can't be represented, fall back to the longest possible
    fraction

  * in rand_fix(), use low-level floating point code from Data::Float
    instead of doing it here

  * expand documentation of rand()

version 0.001; 2006-08-03

  * new function rand_flt() to generate a random floating point value
    in a floating point way

  * bugfix: in BigRat handling, use ->as_number instead of ->as_int
    for BigRat->BigInt conversion, because ->as_int is faulty before
    Math::BigRat version 0.14 but ->as_number is available in much earlier
    versions (the version specified in Makefile.PL was one where ->as_int
    was faulty)

  * use standard interval notation in the Data::Entropy::Algorithms
    documentation

  * use "=> 0" instead of "=> undef" in unversioned dependencies in
    Makefile.PL

  * don't include underscore in string version of module version numbers
    in Makefile.PL

  * remove a redundant BigRat->BigInt conversion

version 0.000; 2006-07-19

  * initial released version