On Sunday, July 1, 2012 5:29:43 AM UTC-4, Slumberland wrote:
>
> The grapher has been defined as an image compositor.
>
> Oops.
> :)
>
> Symptoms:
> -- Render an animated .gif to have an animated graph.
> -- show(A + B) ('plus'? That's a BLIT. Also, the syntax is
> inconsistent. It changes from "add the functions A and B" *precisely* where
> the graph object fails to have any meaningful mathematical input, and
> instead requires "the composition of A && B". Complete change of context,
> at the wrong time.)
> -- all the parameters which (should) belong to the grapher are instead
> duplicated in the various plot() methods. They overlap, and there are then
> *additional* routines to sort out the overlap. e.g. "show axes", "aspect
> ratio".
> -- Where there are *not* routines to sort it out, the behavior is equally
> odd. If the "xmin" and "xmax" values are different for two plot() calls,
> only parts of the function show, when they are added to the same image.
> All but *one* definition of *display range* is accidentally converted to
> a *function evaluation* range.
> -- the solve method is not properly separated from the display method.
> implicit_plot(), plot(), parametric_plot(). But the coordinate space is,
> in reality, completely independent of the method used to solve the equation.
>
>
Truthfully, your remarks are so cryptic that I am not sure what you are
referring to in all of them. To some extent we do try to keep some methods
that are about showing the plot in show() (which I will point out causes no
end of confusion to people who just want to plot graphs), but a lot of
whatever inconsistencies you are trying to point out is the natural result
of taking the original basic functionality and adding methods to it. If
you have the time and expertise to refactor the plot code significantly
(keeping backwards compatibility for a season), that would be a valuable
contribution.
> This seems like a situation where Python could maybe benefit from your
> expertise. Is it possible to talk to the Python developers about replacing
> some of the default methods and definitions over time, if they are
> incorrectly matched to the relationships they are designed to describe? I
> can't see how it is possible to realize the desired improvements to
> Interact, with an image Blitter. :-(
>
I don't know what Python has to do with this. Most of our 2D plot code is
custom wraps of matplotlib; there are some exceptions. If you have
suggestions for Python, probably there is a way to do this.
--
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
