This is a deliberate choice. http://www.jsoftware.com/help/dictionary/d320.htm
And, APL had different behaviors in different contexts, so to some degree the J choices are based on introducing consistency where APL was inconsistent and/or based on simplifying the syntax of the language. Other aspects are based directly on APL. You might be interested in reading the section on Dialects at http://www.jsoftware.com/papers/APLDictionary.htm -- Raul On Fri, Jan 11, 2013 at 12:42 PM, Steven Taylor <tayl...@gmail.com> wrote: > Maybe this is just a way to deal with fill behaviour? In which case, 2 > cases to remember is easier than three. Sounds like this is a deliberate > choice. Does this come across from APL? > > (0 $ 0),!._. (i.2 2) > _. _. > 0 1 > 2 3 > > (1 $ 0),!._. (i.2 2) > 0 _. > 0 1 > 2 3 > > > > > > On 11 January 2013 17:14, Roger Hui <rogerhui.can...@gmail.com> wrote: > >> Try this: if you join a length-m vector to a length-n vector, the >> resultant vector has length m+n. Make sense? The statement remains true >> when m or n or both are 0. >> >> >> On Fri, Jan 11, 2013 at 9:11 AM, Steven Taylor <tayl...@gmail.com> wrote: >> >> > thanks for the reply. I can see that this is consistent for rank 2 and >> > above. The rank 1 case is the bit that isn't making sense to me right >> now. >> > >> > (n$0) , i. 2 [ n=: 0 >> > 0 1 >> > (n$0) , i. 2 [ n=: 1 >> > 0 0 1 >> > ---------------------------------------------------------------------- >> > 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 ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
