In general, you never "need" to use @ because you can write your own work-alike implementation:
ATop=:2 :0 ([: u v)"v ) That said, there are a couple issues to consider. First, is the use of rank. When rank is significant, and the rank of v is the desired rank and the rank of u is not the desired rank, @ is simpler than the alternatives. Second is the use of special code: http://www.jsoftware.com/help/dictionary/special.htm -- so there will be cases where the use of @ (often with the option of @:) is practical where the alternatives take too many resources. -- Raul On Sun, Feb 5, 2012 at 12:30 PM, Linda Alvord <lindaalv...@verizon.net> wrote: > I thought I had found an example where it couldn't be eliminated but > R.E.Boss proved me wrong. > > In recent posts I have withdrawn my objections so I was trying to find a > situation where you have to use @ instead. > > This wasn't it. > > Linda > > -----Original Message----- > From: programming-boun...@jsoftware.com > [mailto:programming-boun...@jsoftware.com] On Behalf Of Boyko Bantchev > Sent: Sunday, February 05, 2012 9:17 AM > To: Programming forum > Subject: Re: [Jprogramming] Another early morning exercise > > On 5 February 2012 14:11, Linda Alvord <lindaalv...@verizon.net> wrote: >> My goal has been to translate from expressions with @ to ones without it. > > You also mentioned eliminating @ in another thread. > Why do you consider it important? > @ is the composition of functions – and is composition not the most > natural operation on functions that one could think of? > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
