Re:Typed nils in Go 2
by kazinator on Friday August 11, 2017 @05:00PM
(#14994225)
(And that empty type is the one named by the symbol nil.)
The nil type is useful at run-time because it constitutes the bottom of the type spindle: just like in set theory the empty set is a subset of every set, including itself, the type nil is a subtype of every type, including itself.
This can be used at run-time; e.g. (subtypep nil 'integer) -> t.
We can't just exclude this value from the type domain on the grounds that it's static only. "Sorry, you don't get a bottom plug on your type spindle at run time ...". :)
The nil type is useful at run-time because it constitutes the bottom of the type spindle: just like in set theory the empty set is a subset of every set, including itself, the type nil is a subtype of every type, including itself.
This can be used at run-time; e.g. (subtypep nil 'integer) -> t.
We can't just exclude this value from the type domain on the grounds that it's static only. "Sorry, you don't get a bottom plug on your type spindle at run time ...". :)