Digital's PDP-11 format has a gap around zero.

DEC-PDP11 c.s.:

  expon = ..000   expon = ..001       expon = ..010                   expon = ..011                       expon = ..100
|<------------->|<------------->|<--------------------->|<------------------------------------->|<----------------------
 H---------------H-+-+-+-+-+-+-+-H--+--+--+--+--+--+--+--H----+----+----+----+----+----+----+----H--------+--------+----
 0               A
  |_____________|
        gap

IEEE-754:

  expon = ..000   expon = ..001       expon = ..010                   expon = ..011                       expon = ..100
|<------------->|<------------->|<--------------------->|<------------------------------------->|<----------------------
 H-+-+-+-+-+-+-+-H-+-+-+-+-+-+-+-H--+--+--+--+--+--+--+--H----+----+----+----+----+----+----+----H--------+--------+----
 0 ^             A
|_________________________________|
   integer arithmetics possible
       (except in E-format)

LEGENDA:

^   = minimum un-normalized non-zero value
A   = minimum normalized value
O   = zero value
H   = visible part of mantissa is 000...
H, +   = numbers representable by the computer
--   = all numbers not representable by the computer
..   = series of zero-bits in exponent

Return to the description about the hidden-bit floats

... OR ...

return to that about the proposal for a double-precision emulator