I agree that previous dyadic result was non-sensical.  New behaviour is 
reasonable.  Though the name perm(utations) is not perfect.  "permutations of 
the combinations x of y".  Could be seperate dyadic verb "permcomb"


   On Saturday, August 25, 2018, 10:13:42 p.m. EDT, Ric Sherlock 
<[email protected]> wrote:  
 
 Hi all,
I've made some changes , before I release them I wanted to check that
no-one has any objections or suggestions for improvement?

https://github.com/jsoftware/stats_base

Summary:

 * Add "with replacement" versions of `comb` and `perm`.

 * Use more performant (less concise) version of `perm` as discussed in wiki
Essay <https://code.jsoftware.com/wiki/Essays/Permutations>. I think
performance of the repeated indexing method is further enhanced by the
changes in the 807 beta.

 * Dyadic `perm` used to return a result, but as far as I can see it wasn't
useful? Now it returns permutations of size x from i.y
  2 perm 3  NB. old behaviour
0 1 2
0 2 1
1 0 2
  2 perm 3  NB. new behaviour
0 1
1 0
0 2
2 0
1 2
2 1

 * Added monadic versions for `comb`, `combrev` for consistency of interface
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm  
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to