> 空白先生いる？(;´Д`)
> 浮動小数点計算における正規化の意味って、必ず小数第一位に0以外の値を入れるってことでいいんだよね？

When a floating-point number is normalized, its leftmost digit is 
to be nonzero, a value 1 &#8804; s < b. Thus zero cannot be represented 
in a normalized floating-point notation because no digit string 
for it has any non-zero digit. 

http://en.wikipedia.org/wiki/Floating_point

参考：2006/09/17(日)15時04分15秒