Re: Underscores in numbers for legibility Lassi Kortela 13 Apr 2019 16:10 UTC

> "Any token that is a potential number but does not fit the number syntax
> is a reserved token and has an implementation-dependent interpretation."

You are right. But in practice it seems like the
implementation-dependent interpretation of choice is to read it as a symbol:

     ;;;; clisp
     [1]> (type-of '1000000)
     (INTEGER 0 281474976710655)
     [2]> (type-of '1_000_000)
     SYMBOL

     ;;;; sbcl
     * (type-of '1000000)
     (INTEGER 0 4611686018427387903)
     * (type-of '1_000_000)
     SYMBOL

     ;;;; ccl64
     ? (type-of '1000000)
     (INTEGER 0 1152921504606846975)
     ? (type-of '1_000_000)
     SYMBOL