My reading of the SRFI says that representing a timespec object as a
"seconds.fractional-seconds" float value is a valid implementation
strategy. Can you confirm, and perhaps add an explicit note to the SRFI
that this has been considered?

Gambit's native "time" type is a tagged float, as discussed on the SRFI
170 list previously. It'd be nice if that type can double as a timespec
type with the addition of the procedures in SRFI 174.