#17194: Minimal bindings for optional arb package
-------------------------------------+-------------------------------------
Reporter: cheuberg | Owner:
Type: enhancement | Status: needs_review
Priority: major | Milestone: sage-6.4
Component: numerical | Resolution:
Keywords: arb, digamma | Merged in:
function, real interval field | Reviewers:
Authors: Clemens Heuberger | Work issues:
Report Upstream: N/A | Commit:
Branch: | a45f691daf51970311b43190496104a774d6518b
u/cheuberg/rings/real_arb | Stopgaps:
Dependencies: #16747 |
-------------------------------------+-------------------------------------
Old description:
> In #16747, the optional arb package is proposed for inclusion in sage.
> That ticket did not have any bindings for the Sage library.
>
> In the present ticket, rather minimal bindings are provided.
>
> As a proof of concept, the digamma function of a
> `RealIntervalFieldElement` is implemented by converting it to an
> `RealBallElement`, computing the digamma function via the arb library,
> and converting the result back.
New description:
In #16747, the optional arb package is proposed for inclusion in sage.
That ticket did not have any bindings for the Sage library.
In the present ticket, rather minimal bindings are provided.
As a proof of concept, the digamma function of a
`RealIntervalFieldElement` is implemented by converting it to an
`RealBall`, computing the digamma function via the arb library, and
converting the result back.
--
Comment (by cheuberg):
Replying to [comment:32 mmezzarobba]:
> * Why call the element class `RealBallElement` and not just `RealBall`?
to me the former sounds more like a point ''in'' a real ball, not the ball
itself...
renamed.
> * Does the parent class really need to be a Cython class?
What are the disadvantages if it is a Cython class?
While working on #17222, I found that I spent most of the time in many,
many calls of `ComplexIntervalField.__call__`. Therefore, I wanted
`RealBallField` to be a Cython class in order to save any bit of
performance possible.
> * I'm not sure if setting up coercion maps from real interval fields is
a good idea. It's probably sound from the mathematical point of view, but
I seem to remember that bidirectional coercions are at least discouraged,
and when computing, say, the sum of an interval and a ball, I'd tend to
expect the result to be an interval. Or perhaps this should depend on the
precisions of the arguments...?
I do not understand your last comment on the precisions of the arguments.
Apart from that, do you have an opinion on how the output of the balls
should look like? See comments [comment:27 27] and [comment:28 28].
--
Ticket URL: <http://trac.sagemath.org/ticket/17194#comment:35>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
