On Wed, Apr 11, 2012 at 1:24 PM, leppie <xacc.ide@gmail.com> wrote:
>> R5RS doesn't have such permission - it has permission
>> to return #f when the string has _any_ radix prefix, and
>> in a number of other cases related to incomplete numeric
>> towers.  What we explicitly voted on is that these must
>> be unified with `read` behavior.  I don't think this is an
>> unintended consequence.
>
> What will (string->number "#x10" 8) return?
>
> I think the issue is with ambiguity.

That would be an argument for returning #false
on mis-matched radixes as John suggested, not
on just returning #false for any radix prefix as
R5RS specifies.

Personally I don't think the ambiguity is a big
issue - the radix prefix in the string is later and
I don't think people would be surprised that it
acts as an override for the radix argument.  I.e.
your example should naturally return 16.

--
Alex