in is a hangover from F#’s OCaml roots and it specifies bound variables, which are subtly different to variable scopes.
Think of variable binding as follows; You have an expression:
first + " " + second
As it stands first and second are unbound – they don’t have any fixed values – so that expression has no concrete value at present. By using
let (...) in
syntax you are specifying how those variables are bound in that expression, so your example will use variable substitution to reduce that function down to
let doStuff x =
x + " " + x
In this example both forms are identical, but imagine the following:
let (x = 2 and y = x + 2) in
y + x
This will not work the same as
let (x = 2 and y = x + 2)
y + x
Because in the former case x is only bound after the in keyword.
In the later case normal variable scoping rules take effect, so variables are bound as soon as they are declared.
Hope that clears things up. In general you should always use the version without in and specify #light at the start of your F# source files