Michael's trick plot|:^@-@-:@*:"0 D.(i.3)10%~50-~i.101 NB. replace ^&2 by *: (!)
works nicely! I don't know why. My version is: Engine j807/j32/windows. Thank you very much! Bo. ----- Videresendt meddelelse ----- Fra: 'Michael Day' via Programming <programm...@jsoftware.com>Til: "programm...@jsoftware.com" <programm...@jsoftware.com>Sendt: torsdag den 24. november 2022 kl. 20.05.54 CETEmne: Re: [Jprogramming] Differentiation problem But... as I've just discovered, the equivalent function expressed as ^@-@-:@*: does yield a "smooth" curve for the 2nd differential, so presumably there's a different need to do or level of approximation with ^&2 compared to *: There is a remark Generating higher-order partial derivatives with D. 2 fails. Use repeated application of D. 1 instead in the documentation at https://code.jsoftware.com/wiki/Vocabulary/dcapdot Mike On 24/11/2022 18:36, Raul Miller wrote: > Presumably you're working with an older version of J, like J807. (J9 > tossed the D. conjunction and requires we use the math/calculus > library manually.) > > Anyways, if you increase the size of the plot, you'll see that the > display artifact you're looking at is a consequence of discretization > -- the plot is composed of lines and at that particular point in the > graph these lines do not adequately conform to the underlying curve. > > FYI, > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
