On May 5, 2012, at 12:01 PM, Raul Miller <rauldmil...@gmail.com> wrote:
> On Fri, May 4, 2012 at 6:36 PM, Joe Groff <arc...@gmail.com> wrote:
>> Thanks! I was trying to cook something up like that myself. Here's a
>> tweaked version that works with heterogeneous ranks:
>>
>> ---
>> powmod2 =: 1 :(':';'(m|*)/(|.(i.<:#$b)|:b)^~(m|*:)^:(<{:$b=.#:y) x')
>> 5 6 7 (10 powmod2) 3 2 $ 40 0 0 41 42 0
>> 5 1
>> 1 6
>> 9 1
>
> There's a simpler way to accomplish that:
>
> 5 6 7 (10&|@^) 3 2 $ 40 0 0 41 42 0
> 5 1
> 1 6
> 9 1
Sorry, that was a dumb example. As I said in the original post, m&|@:^
doesn't work when m is a list.
> They can be generalized rather easily, in my opinion, but
> before trying to generalize it's usually better to define
> what it is you are trying to accomplish.
I was trying to write an array-based implementation of the
Miller-Rabin primality test that worked for both a single candidate
and a list of candidates. (I know that the algorithm should short
circuit, and an array-based implementation does extra unnecessary
work; this was just a learning J exercise.) The obvious (to me) way to
do that was with an operator that could handle both rank0 | rank2 ^
rank1 and rank1 | rank3 ^ rank2 operations in the natural way while
still using modular exponentiation.
-Joe
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
