It is explained in the essay http://www.jsoftware.com/jwiki/Essays/Rank
The recursively-defined verb s boxr y produces the list of such cells. The purpose of boxr is to produce the list of such cells. What are "such cells"? The immediately preceding phrase says: r cells y computes the array of cells with shape r fr y , each cell individually boxed and shaped s=: r cs y (r cells y ↔ <"r y) . And in the model: cells =: fr $ cs boxr ,@] So r cells y makes an array of the boxed cells relative to rank r and boxr makes the list of "such cells" (r (cs boxr ,@]) y ↔ ,<"r y). boxr just happens to be recursive. One of its requirements is to avoid overt use of rank. boxr =: ]`(<@$ , [ $: */@[}.])@.([EMAIL PROTECTED]@]) ----- Original Message ----- From: Raul Miller <[EMAIL PROTECTED]> Date: Saturday, June 23, 2007 12:08 pm Subject: Re: [Jgeneral] Let's get explicit about dyadic verb processing > On 6/23/07, Roger Hui <[EMAIL PROTECTED]> wrote: > > Your worksheet uses English prose which is difficult > > to get precise and correct. What is needed is an > > expression of the ideas and rules in an executable > > notation. I have just the thing: > > http://www.jsoftware.com/jwiki/Essays/Rank > > What is the purpose of boxr? More specifically, why is > its definition recursive? ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
