[add information about the way tcl handles floating point numeric values as C type doubles, the range of values supported, etc.]
double is an expr function (or a proc in the mathfunc namespace) [L1 ] for converting tcl numeric strings to floating-point and returning the converted value.
See also Inf.