TODO.txt

1. NSGK results
- current tests has significant results and concordance is less than 1/2
- is there a test that reveals scientifically significant results
  as well as statistically significant

2. Performance (Pythonic)
- numba / Cython ? (Stefan)

3. permuter scripts in pkg / building blocks
- Ultimately want scripts in permute to test.  Are vignettes reusable?
  - enumerate / (combinat from CSP.R)
  - itertools.combinations       
         
4. two_sample
- exact two-sample problem
	- (use MacNell as example/test)
	- see issue + gist
- Write a vignette to show how this can be used to do rank tests, e.g. Wilcoxon

5. k-sample tests
- ks-like test statistic: combine ks for each pair of samples by max distance, sum of distances, npc...
- Permutation ANOVA
- other?

6. n-dimensional k-sample
- MST
- LLQ
- MDL (research topic)

7. Association
- Pearson
- Spearman (replace data by their ranks in call to Pearson)
- Regression ( null beta=0)

8. Misc non-parametrics
- Runs tests
- Other kinds of symmetry
- Romano-style projection

9. Contingency tables
- ecological fixed-margin tests 
	- Stefan's data example - see issue
	
10. Nonparametric combination of tests
- generalize the IRR NPC code