problem with 'double'

[Roedy Green]

Changing the exponent is moving the decimal point!

That is true for BigDecimal, the issue under discussion, but for
double, changing the exponent in the IEEE representation, is moving
the "binary point".

 

[Dale King]

That is true for BigDecimal, the issue under discussion, but for
double, changing the exponent in the IEEE representation, is moving
the "binary point".

Which I already addressed earlier when I asked whether you call it a binary
point for binary numbers. Doing some googling it appears that binary point
is indeed the correct terminology. I specifically used a decimal example to
avoid the confusion.

The idea is the same with binary floating point. The exponent controls
where the division between the integral portion of the mantissa and the
fractional part goes.