Reverse Polish Notation Calculator
------- ------ -------- ----------

This calculator operates on floating point numbers and
provides the operations +, -, *, and /.  It uses reverse
polish notation; the values are entered first, followed
by the operation to perform.  For instance,

3 2 -

should calculate 3-2.  The result is left on the stack.

The stack is five elements deep.  All elements start out
initialized to zero.  On each new number entry, the
stack is `pushed' (entries moved down, with the last
entry falling off the end).  On each operation, the two
operands are popped off the stack, the operation is
done, and the result is pushed onto the stack.  On each
stack pop, the last element on the stack is duplicated.

The `-' operator subtracts the first element popped from
the stack from the second element popped.  The `/'
operator divides the second element popped by the first
element popped.

Input is multiple lines, each one containing a white-
space-separated sequence of zero or more commands.  Each
command is either a single-character operation or else a
number.  Input numbers can be integers, floating point,
or in scientific representation; they will all be within
the range of a `double precision' type.  After each
input line is processed, the first value on the stack is
printed on a line by itself.  (First is the most recent-
ly pushed number.)  Note that the stack is not cleared
between lines.  Repeat:  do not clear the stack between
lines!

The output should be accurate to at least 0.0001% (6
decimal places) as long as similarly valued numbers are
not subtracted and the number of operations needed to
compute an output number is less than 100.  Again this
is well within the limits of double precision floating
point arithmetic as long as at least 6 digits of
precision are output.

Sample input:

3
2 +
1 2 /
-3.2e10 5 /

Sample output:

3
5
0.5
-6.4e+09