>@a->[2] is >accepted, without complaint, to apparently mean $a[2]. And I cannot >see why 'scalar(@x)->[2]' produces the 3rd element of array @x. Those are old bugs, which, per my recollection, were judged "very difficult" to address. --tom