The flavor float describes a random floating-point number in a particular range.

To describe a flavor of a random floating-point number, use either float or float(opts) (where opts is described following) as the argument to RandomTools[Generate] or as part of a structured flavor.

By default, the flavor float describes a random floating-point number logarithmically distributed in the range epsilon..1.0 - epsilon, inclusive, where epsilon = 10e-Digits.

You can modify the properties of the random floating-point number by using the float(opts) form of this flavor. The opts argument can contain one or more of the following equations.

range = a..b

This option specifies the range from which the random float is chosen. The range endpoints a and b are numeric and either a >= 0.0 or b <= 0.0.  All numerics are evaluated by using the setting of the digits option.

If , then  is set to the smallest value of the form  such that . If , then  is set to the smallest value of the form  such that .

If , an exception is raised.

digits = posint

This option specifies a positive integer to use as the Digits setting. The default setting is the current setting of the Digits environment variable.

method = uniform or logarithmic

This option specifies whether the floating-point number should be chosen logarithmically or uniformly from the interval.

The logarithmic method is identical to listing all of the unique floating-point numbers that are found between the endpoints, and then choosing one of these randomly.

The uniform method is similar to sampling from a uniform distribution that is bounded by the endpoints, and then converting this result into a floating-point number.

The default value for this option is logarithmic.