On 2 Oct 2006, Marcin 'Qrczak' Kowalczyk wrote:
> If any sorting is adopted, I call for making Schwartzian transform
> available. People must not be advised to pass
> (lambda (x y) (< (car x) (car y)))
> as the comparison. They should pass
> car <
> as two separate arguments (or just car with < being the default).
I find this feature *almost* indispensible, but I would happily
dispense with it (for a few years at least) rather than be left with
the status quo.
Eli Barzilay wrote:
> I think that the `sort' issue is hopeless, since there are many
> features you need to make everyone happy.
This committee knows it isn't going to "make everyone happy", and it
has made bold changes nonetheless. I remain resolutely optimistic.
> Some Lisp implementations will do something similar when
> receiving a `:key' argument.
This feature is actually in Common Lisp. I agree it's useful.
http://www.lisp.org/HyperSpec/Body/fun_sortcm_stable-sort.html
-j
Received on Tue Oct 03 2006 - 10:40:48 UTC