This observation, whether it is true or not, seems weak to me as a reason for omitting ;!.f .
; does the same things (rank extension, scalar replication) that dyad , does. In fact, , can be modeled as ;@:,&< . If the argument for ,!.f was sound, the argument for ;!.f is equally sound, _mutatis mutandis_. My interest in the matter is that I continue to try to document how J works, and the absence of ;!.f is an anomaly that takes more explaining than describing ;!.f would. Henry Rich P. S. I bet most J programmers would not correctly predict the result of the following sentences: ; (i. 2 4) ; (i. 3 2) ; 6 ; (i. 3 0 4) ; (i. 0 2) ; 6 ; (i. 3 0 4) ; (i. 0 2) ; ,6 'assembles along a leading axis the opened elements of the ravel of y' is an incomplete description of what ; does. Henry Rich > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of R&S HUI > Sent: Thursday, June 22, 2006 10:20 PM > To: Programming forum > Subject: Re: [Jprogramming] monad ;!.f > > Expressions requiring a fill for ; are indicative of > a poorly design data structure -- the cells for which a > fill is needed are anomalous compared to the other cells. > > > ---------------------------------------------------------------------- > For information about J forums see > http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
