Hi Alex, On 2012-03-24, AlexGhitza <aghi...@gmail.com> wrote: > ------=_Part_614_22458373.1332555347386 > Content-Type: text/plain; charset=ISO-8859-1 > Schemes and their points are not yet properly embedded in the category > framework.
Yes. In one of my patches on categories, I ran into it and found it impossible to get Schemes to work in that regard. I mean, how to create a morphism, for example. It ought to be possible to create morphisms if you want to put things into a category! > In the latter case (elements of QQ), things are set up so that Sage > knows that Rational is a subtype of QuotientFields.element_class. But > for schemes, Sage does not know that SchemeMorphism_point_affine is a > subtype of Schemes_abstract.HomCategory.element_class. So how do I fix > this? I have tried to locate how this is done for QQ, but I haven't > succeeded. Can anyone point me to this? Have you had a look at my worksheet/tutorial on it? See http://sagenb.org/home/pub/3837/ As Nicolas has pointed out, the most important thing is that the category of the parent is defined during its initialisation. In sage-5.0, most algebraic parent structures *will* use categories (e.g., all rings). In order to make an element of P inherit methods from P.category().element_class, it actually suffices to initialise the category of P. However, in that case, access to these additional methods may be a bit slow, because some custom __getattr__ method is used. However, if you provide P with an attribute "Element", then P will automatically provided with an attribute P.element_class -- and this element class can access the additional methods by class inheritance, which is of course faster than a custom __getattr__ method. Best regards, Simon -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
