Re: [sage-devel] CategoryObject vs. Parent

2015-05-12 Thread David Roe
On Tue, May 12, 2015 at 2:43 AM, Nicolas M. Thiery wrote: > On Tue, May 05, 2015 at 10:30:18PM -0700, Robert Bradshaw wrote: > > > Eric: > > > In the same vein, shouldn't base_ring be removed from Parent as well ? > > Yes indeed. > > > As Parents have Elements, this is also the root of operations

Re: [sage-devel] CategoryObject vs. Parent

2015-05-11 Thread Nicolas M. Thiery
On Tue, May 05, 2015 at 10:30:18PM -0700, Robert Bradshaw wrote: > > Eric: > > In the same vein, shouldn't base_ring be removed from Parent as well ? Yes indeed. > As Parents have Elements, this is also the root of operations that > deal with coercion (arithmetic between elements of unequal Paren

Re: [sage-devel] CategoryObject vs. Parent

2015-05-06 Thread Nathann Cohen
> > In the same vein, shouldn't base_ring be removed from Parent as well ? > > sage: Set([1,2,3]).base_ring() > > The return value is None, which is arguably correct, for a generic set has > no base ring, but it might be surprising when using tab completion to > discover functionalities to see t

Re: [sage-devel] CategoryObject vs. Parent

2015-05-05 Thread Robert Bradshaw
On Tue, May 5, 2015 at 2:10 PM, Eric Gourgoulhon wrote: > Hi, > > Le mardi 5 mai 2015 12:44:08 UTC+2, vdelecroix a écrit : >> >> >> IMHO, they do not belong either to parents. A set has generators? >> >> sage: Partitions(5).gens_dict() >> Traceback (most recent call last): >> ... >> AttributeError

Re: [sage-devel] CategoryObject vs. Parent

2015-05-05 Thread Eric Gourgoulhon
Hi, Le mardi 5 mai 2015 12:44:08 UTC+2, vdelecroix a écrit : > > > IMHO, they do not belong either to parents. A set has generators? > > sage: Partitions(5).gens_dict() > Traceback (most recent call last): > ... > AttributeError: 'StandardPermutations_n_with_category' object > has no attribut

Re: [sage-devel] CategoryObject vs. Parent

2015-05-05 Thread Nicolas M. Thiery
On Tue, May 05, 2015 at 12:44:00PM +0200, Vincent Delecroix wrote: > > Very much like for Parent, there probably are things that are in > > CategoryObject for very old legacy reasons (typically pre-categories!) > > and really do not belong there; this could typically be the case for > > generators

Re: [sage-devel] CategoryObject vs. Parent

2015-05-05 Thread Vincent Delecroix
On 05/05/15 11:00, Nicolas M. Thiery wrote: > On Tue, May 05, 2015 at 10:41:42AM +0200, Jeroen Demeyer wrote: >> what are the types CategoryObject and Parent supposed to represent? For >> example, I don't understand why generators and names are dealt with on the >> level of CategoryObject but const

Re: [sage-devel] CategoryObject vs. Parent

2015-05-05 Thread Nicolas M. Thiery
On Tue, May 05, 2015 at 10:41:42AM +0200, Jeroen Demeyer wrote: > what are the types CategoryObject and Parent supposed to represent? For > example, I don't understand why generators and names are dealt with on the > level of CategoryObject but constructing elements is dealt with on the level > of

Re: [sage-devel] CategoryObject vs. Parent

2015-05-05 Thread David Roe
Parent should represent an object in a category that has a forgetful functor to sets; CategoryObject an object in an arbitrary category. I have no objection to moving the generator and name code to Parent. David On Tue, May 5, 2015 at 4:41 AM, Jeroen Demeyer wrote: > Hello, > > what are the typ

[sage-devel] CategoryObject vs. Parent

2015-05-05 Thread Jeroen Demeyer
Hello, what are the types CategoryObject and Parent supposed to represent? For example, I don't understand why generators and names are dealt with on the level of CategoryObject but constructing elements is dealt with on the level of Parent. In fact, in all of Sage, there is only a single cl