unCLog

I hear from various sources that CLISP and LispWorks both behave as Allegro does (here and here) with respect to delete and express adjustability. I confess to not having thought about it (*) but I can't quite see where the efficiency comes from in having vectors potentially lose their (express) adjustability during a delete. I would have thought that the (naive?) method of swapping the items to be deleted with the items at the end and then shrinking the size would be the fastest general method. And this method, it would seem, wouldn't alter any other vector properties. I guess this is just another question to add to my list.

(*) Saying "I haven't thought about it" is an intellectual's cover when he (or she) is worried that they are about to step in it!

Home | About | Quotes | Recent | Archives