[Originally sent last night but breached the 100KB limit by top posting on too many old messages. Trimming down and resending so it appears in the archives.]
The most direct translation, including the fixing of A, is: X=: 50 10 ?@$ 10 A=:i. 10 10 ]`(A"_)}"_1 X 0 71 82 93 24 25 56 57 58 39 80 1 2 53 94 65 46 7 68 49 90 81 52 73 44 5 16 27 78 99 30 21 12 23 94 55 36 97 68 99 ... Extending this to allow A to vary would be somewhat convoluted, which is one of the inconveniences of gerund} I referred to earlier, arising from a desire for consistency*. -Dan *In this case, the requirement for gerund} to be applied monadically in order to get the monad behavior of verb}, even though grammar doesn't require that. Note that v0`v1`v2)} y is mapped to (v1`v2)} y , ie v0 is discarded, even though we could have it mean something useful, just because verb}y can't have an x argument.... > On Mar 30, 2014, at 5:27 PM, EelVex <[email protected]> wrote: > > For the record, I wanted to use 'item amend' for a fixed right-hand noun on > a series of precomputed left-hand nouns. > i.e. I was hoping this: > > A&}~ each X > > would mean: > > A&(4 :'y } x') each X > > $A > 100 100 > $X > 5000 100 > > (Taking 5000 "diagonals" from a 100x100 matrix) > > Any ideas? ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
