Roger wrote: > I suspect v"0 would serve. If it > does not, I'd like to see what such a v is.
Viktor replied: > a =: 1 : '$ $ u@,' > v =: 5&, > > v"0 i. 2 2 > [vs] > v a i. 2 2 Roger responded: > The intent of my question [was for] > some practical use that breaks > the equivalence. I cited some practical uses in [1] (and the last citation will tell you how I found them, and how you might find others). One use is in JSoftware's standard library, where a matrix needs to be passed to a library written in a less array friendly language, so the input is raveled to vector, and the output reshaped into a matrix. On a historical note, both Kirk Iverson and I had something analogous to UnderRavel in our personal utilities scripts. So we must have found some repeatable use for the pattern (though I admit I cannot remember an example right now, and also my utilities script has collected some less-than-useful specimens...) -Dan [1] Earlier post in this thread: http://www.jsoftware.com/pipermail/programming/2010-October/020648.html ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
