The PArsing and Execution chapters of _J for C Programmers_ discuss this in some detail.
Henry Rich Viktor Cerovski wrote: > > Dan Bron wrote: >> Thanks Raul & Victor. I agree with the assesment that names don't get >> resolved until they're executed (that is, given the correct >> number of arguments). >> > Thank you Dan for posing this interesting problem. > > To me the real question that follows from such an explanation is: when does > the name get executed? Since I lately increasingly tend to think of/in J in > terms of functional programming, where there is no execution, it seems that > the execution is a bit more involved in defining J than I see it necessary > to achieve what J gets done. > > >> As you [Raul] pointed out, that explains the behavior I questioned. At >> first blush, the solution niggled me because pronouns don't have >> to wait for execution to produce their value. But, of course, [pro]nouns >> take 0 arguments, so mentioning them also fully >> parameterizes them, resulting in execution and stacking by value. >> > Ditto. > > >> Now, the harder question: Does the Dictionary predict this behavior (not >> resolving names until they're executed), and if so, >> where? If not, can we defend the behavior using the text of the DoJ? Put >> another way, does the DoJ prescribe this >> implementation, or at at least not proscribe it? >> > It all depends on how precisely you want to define the language. To me it > seems at the present that one rule of J, at the very best, is still implicit ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
