User`s manual

4-20
Effects of Type Conversions on Accuracy
When a number is converted to integer type, it is "rounded down"; i.e., the
largest integer, which is not greater than the number is used. (This is the
same thing that happens when the INT function is applied to the number.)
When a number is converted from double to single precision, it is "4/5
rounded" (the least significant digit is rounded up if the fractional part > =5.
Otherwise, it is left unchanged).
In the following examples, keep in mind that single precision variables are
stored with 7 digits of precision, but printed out with 6 digits (to allow for
proper rounding). Similarly, double precision values are stored with 17 digits
but printed out with only 16
.
Example Programs:
10 A#=1.6666666666666667
20 B!=A#
30 C%=A#
40 PRINT B!,C%
RUN
1.66667 1
READY
>_
When a single precision number is converted to double precision, only the
seven most significant digits will be accurate. And if the single precision
number didn't contain seven significant digits, watch out!
Examples:
10 A!=1.3
20 A#=A!
30 PRINT A#
RUN
1.299999952316284
READY
>_
----------------------------------
10 A#=2/3
20 PRINT A#
RUN