And note also that this does not exhaust the possibilities for this case. Here are a few others:
If the ranks are appropriate, [: f0 f1 f2 f3 f4 f5 might be written as f1 f0@f2 f3 f4 f5 Alternatively if the fork is associated with a concept that's worth considering in isolation, we might use something like f6=: f1 f2 f3 f4 f5 along with f0@f6 Or, maybe we can rewrite f0 as a dyadic verb which needs a constant left argument. Or, maybe we can eliminate f0 entirely if we move the whole process into another domain. I think one of the issues here is that we are struggling to express meaning, and style is only one of our tools for doing so - and it's not always our best tool for that purpose. -- Raul On Thu, Feb 7, 2013 at 3:28 PM, Roger Hui <[email protected]> wrote: >> I feel that it's very hard to win (hard to be clearer) if you replace a > primitive conjunction by something else. > > I can use this to illustrate why it's difficult to have an absolute rule of > style. If I am composing functions f0, f1, f2, f3, f4, etc., and assuming > rank is not a problem, I'd write f0@f1@f2@f3@f4. But suppose I have a > bunch of forks and then the last (leftmost) thing is a f0@ ? In that case > I may write [: f0 f1 f2 f3 f4 f5 instead of f0@(f1 f2 f3 f4 f5). Or, even > more confounding, if the last two things are atops? Then do I write f0@f1@(f2 > f3 f4 f5 f6) or [: f0 [: f1 f2 f3 f4 f5 f6 ? In such cases it is arguable > that replacing a primitive conjunction made the expression clearer. > > > > On Thu, Feb 7, 2013 at 11:31 AM, Roger Hui <[email protected]>wrote: > >> I would write it: >> >> ic1=: , @ (j./&i:/) @ +. >> >> I feel that it's very hard to win (hard to be clearer) if you replace a >> primitive conjunction by something else. In this case & . >> > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
