Thank you for explaining that. But shouldn't something be said in help for d. that it will take non-integer values for n in the expression ^&n ? As it stands now it says it's the same as D. except for rank which then restricts n to be integers.
On Fri, Jan 8, 2010 at 11:44 AM, Oleg Kobchenko <[email protected]> wrote: > > According to the dictionary: "u d. n is like u D. n except that u is > treated > > as a rank-0 function." So why the nonce error? > > If you read further, it talks about partial derivatives. > So you don't want to use D. deliberately wanting them. > > See below. > > ^&3 d. 1 > 3&*@(^&2) > ^&3 D. 1 > (* =/~@(i...@$))@:(3&*@(^&2)) > > 1&o. d. 1 > 2&o. > 1&o. D. 1 > (* =/~@(i...@$))@:(2&o.) > > 1&o. d. _1 > -@(2&o.) > > (1&o. D. 1) 1 2 3 NB. linear operator > 0.540302 0 0 > 0 _0.416147 0 > 0 0 _0.989992 > > (1&o. d. 1) 1 2 3 NB. list of scalars > 0.540302 _0.416147 _0.989992 > > 1&o. D. _1 > |nonce error > | 1&o.D._1 > > So it states here that calculation of partial antiderivative > is not implemented (yet). > > > > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
