Changes between Initial Version and Version 1 of Ticket #25388, comment 8


Ignore:
Timestamp:
05/18/18 08:57:06 (4 years ago)
Author:
embray
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #25388, comment 8

    initial v1  
    22> I can think of certain morphisms that I could see benefiting from being `UniqueRepresentation`, but I wouldn't do non-equality comparisons on them. The other would be for small sets of key algebraic objects. We do this in `combinat/crystals/letters.pyx`, but we do it by `@cached_method` on `_element_constructor_` to help improve the construction time and to tie it directly with the parent.
    33
    4 On that note, could you have a look at #25389 and tell me if the use of `UniqueRepresentation` ''doesn't'' make sense in any of those cases.  I felt that it did make sense but clearly there's a lot of hard-earned experience I don't have as to when it does or doesn't make sense to use.
     4On that note, could you have a look at #25389 and tell me if the use of `UniqueRepresentation` ''doesn't'' make sense in any of those cases.  I felt that it did make sense but clearly there's a lot of hard-earned experience I don't have as to when it does or doesn't make sense to use. (Note: The branch in that ticket has this one as a dependency, so it's really only interesting to look at the changes in quaternion algebras: https://git.sagemath.org/sage.git/diff/src/sage/algebras/quatalg/quaternion_algebra.py?id=9ea55d9933a3cd2bedd391068276f804ff868e1f&id2=6fc1e20c666283a301b4ff3f855013de8d206b35)