Changes between Initial Version and Version 1 of Ticket #15456, comment 20


Ignore:
Timestamp:
03/06/14 18:33:53 (8 years ago)
Author:
tscrim
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #15456, comment 20

    initial v1  
    1 I don't like having the warning saying "don't override this", especially since I don't see  a good reason for doing this (and there's no "final" semantic in python). I am in favor of saying we must implement either `has_left_descent()` or `has_right_descent()`. However I still think a better fix will be for `has_left/right_descent()` to call `has_descent()` by default.
     1I don't like having the warning saying "don't override this", especially since I don't see  a good reason for doing this (and there's no "final" semantic in python). I am in favor of saying we must implement either `has_left_descent()` or `has_right_descent()`. Also possibly implementing has_left/right_descent()` for `weyl_group.py` to call `has_descent()`.
    22
    33Anyways, I've put my working version on trac at `u/tscrim/15456`. I also gave default implementations of `has_left/right_descent()` to `weyl_group.py` for speed. Although now that I've done the implementation, I'm less convinced of my suggestion. Perhaps what would be best is some category magic. We test upon creation of a Coxeter group to see if any of the methods have been implemented. Thus we do the following: