Thank you for the derivatives and integrals.

> In particular, if anyone can shed light on the intended use
> of 8&o. and _8&o. I would appreciate it.

See Paul Penfield's APL81 paper, "Principal values and branch 
cuts in complex APL",
http://portal.acm.org/citation.cfm?id=805368
the section entitled "New Pythagorean Functions" on page 254.



----- Original Message -----
From: John Randall <[email protected]>
Date: Sunday, February 20, 2011 8:25
Subject: Re: [Jprogramming] Derivatives of o. family
To: Programming forum <[email protected]>

> Here are my suggestions for the result of f d. _1 where f is in the
> o. family.
> 
> All of these have been tested by differentiation and comparison with
> the original function on suitable arguments.  For example
> 
>    >./|(((* _1&o.) + 0&o.) D. 1 - _1&o.)"0 ]_1+2*50 
> ?@$ 0
> 1.5497e_7
> 
> is pretty good evidence that (* _1&o.) + 0&o. is an 
> antiderivative of
> _1&o.
> 
> "OK" means that J already knows the integral: "no integral" means
> that there isn't one.
> 
>   f    f d. _1
>   0&o. -:@((* 0&o.) + _1&o.)
>   1&o. OK
>   2&o. OK
>   3&o. OK
>   4&o. -:@((* 4&o.)+^.@(+ 4&o.))
>   5&o. OK
>   6&o. OK
>   7&o. OK
>   8&o. j.@-:@((* 4&o.)+^.@(+ 4&o.))
>   9&o. no integral
>  10&o. no integral
>  11&o. no integral
>  12&o. no integral
>  _1&o. (* _1&o.) + 0&o.
>  _2&o. (* _2&o.) - 0&o.
>  _3&o. (* _3&o.) - -:@^.@(1+*:)
>  _4&o. -:@((* _4&o.) - ^.@(+ _4&o.))
>  _5&o. (* _5&o.) - 4&o.
>  _6&o. (* _6&o.) - _4&o.
>  _7&o. (* _7&o.) + ^.@(0&o.)
>  _8&o. -@j.@-:@((* 4&o.)+^.@(+ 4&o.))
>  _9&o. -:@*:
> _10&o. no integral
> _11&o. j.@-:@*:
> _12&o. -@j.@(_12&o.)
> 
> Obviously these can be expressed in several ways.  Any 
> improvements would be
> welcomed.  In particular, if anyone can shed light on the 
> intended use
> of 8&o. and _8&o. I would appreciate it.

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to