Hi Robert, William,
On Wed, Mar 04, 2009 at 01:43:23AM -0800, Robert Bradshaw wrote:
> On Mar 3, 2009, at 9:10 PM, Nicolas M. Thiery wrote:
> > Dear Sage developers,
> >
> > Let me recycle this thread to report progress on the bookshelf of
> > standard object behavior. I just posted a small patch:
> >
> > http://sagetrac.org/sage_trac/ticket/5120#comment:3
> >
> > which implements sage.structure.UniqueRepresentation class. Derived
> > classes inherit a unique representation behavior for their instances.
> > Originates from sage.categories.category.uniq with:
> >
> > - Long doc (6 pages for 12 lines of code)
> > - Default implementation of__eq__ and __hash__
> > - Handling of pickling in the simple cases
> > - Avoids multiple calls to __init__
> > - Enables rewritings of the constructor argument
> >
> > This is a call for comments! (and reviews).
> >
> > If the principle is validated, it could be worthwhile to use it quite
> > systematically for parents.
>
> I am a huge fan of uniqueness of parents.
>
> Any comments on how this relates to the code at http://
> hg.sagemath.org/sage-main/file/b0aa7ef45b3c/sage/structure/factory.pyx ?
Thanks for the pointer!
There definitely a feature overlap, but not
complete. UniqueRepresentation can be used by inheritance which has
two advantages:
- Setting up a policy for a full hierarchy of classes
(for example, I will be able to enforce the policy that all
CombinatorialFreeModule's have unique representation. Or I did take
the decision that all categories have unique representation; well,
this later decision deserves debate, but that's another story)
- Hiding this as an implementation detail, and presenting a single
gadget to the user, so that he can do:
bla = Bla(x)
isinstance(bla, Bla)
rather than:
bla = Bla(x)
isinstace(bla, Bla_class)
(no class name digging)
On the other hand, there are a couple cool features only in
UniqueFactory, and UniqueFactory has use cases not covered by
UniqueRepresentation. So I think both should live (and be used
systematically!!!), but features should be merged as much as possible.
William: would you have time to work on this with me next week at MSRI?
Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <nthi...@users.sf.net>
http://Nicolas.Thiery.name/
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
