masayoshi.okutsu@Eng 1996-12-20
Float.toString(1.0e7f) produces "1.0E+7" instead of "1.0E7", which is
inconsistent with 'The Java Language Specification.'

Float.toString and Double.toString has inconsistence with
'The Java Language Specification' (Gosling, Joy, Steele).

  in 20.9.16 of the 'bible', they wrote:
 
 (cut) .. The magnutide is then represented as the integer
 part of a, as a single decimal digit, followed by '.'
 (\u002e), followed by decimal digits representing the
 fractional part of a, followed by the letter 'E' (\u0045),
 followed by a representation of n as a decimal integer,
 as produced by the method Integer.toString of one argument (20.7.12).

where n is an integer 10^n <= argument < 10^(n+1).

  in 20.7.12, Integer.toString(int) is defined as toString(int, 10).

  in 20.7.13,

  (cut) .. If the first argument is negative, the first character of
  the result will be the character '-' (\u002d). If the first argument
  is not negative, no sign character appears in the result.

So, Float.toString(1.0e+7f) must produce "1.0E7", but current JDK1.1
Float.toString() produces "1.0E+7" (watch out on '+'.)