Hi Nicolas,

Could you please comment on Simon's point below?
I did not easily find this in the category primer. Where is it discussed there?

> One detail that I find fishy: It is possible that a category class (say,
> PermutationGroups, but it could be a different name, I don't recall)
> has an attribute (say, Finite) which is a lazily imported class (here:
> FinitePermutationGroups), but then the instance
>   P = PermutationGroups()
> does *not* inherit the "Finite" attribute from its class! Instead, by
> some magic, is is replaced by a cached method, which then returns an
> instance of FinitePermutationGroups.
> 
> That's a detail that I find (a) confusing, (b) likely to result in
> difficult-to-debug errors, (c) not needed, since in the end the cached
> method P.Finite() returns type(P).Finite() anyway.
> 
> Nicolas' counter argument was (a) the cached method is documented, so
> that you can easily introspect the specification of the "Finite" axiom,
> (b) tests pass and people use it without problem, so let's not speculate
> on potential difficult-to-debug errors that have not shown up in the
> existing use cases yet, and (c) it is needed, because of the documentation
> (perhaps Nicolas gave further arguments, I am writing this without
> re-reading the comments on trac).

Also, Volker told me that the proof of concept that the syntax he prefers
can be implemented can be found here

http://trac.sagemath.org/15701

I think the consensus is that the feature that is implemented in 10963
is definitely worthwhile for Sage. So something that actually works with
full functionality to me at the moment seems more convincing than a proof
of concept.

Volker, how quickly can you get your way of doing things implemented
and tested in real life situations? Last time you said within a month,
but that was a over a month ago.

Best,

Anne



-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to