Changes between Version 5 and Version 7 of Ticket #10945


Ignore:
Timestamp:
04/25/11 15:46:25 (8 years ago)
Author:
kcrisman
Comment:

#11028 is taking care of the modularity and ColorPlot issues, I think.

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #10945 – Description

    v5 v7  
    1111   This is hard to follow, though I see what it says.
    1212 * There is also a 'reimann' several times, and a 'correspondance'.
    13  * A lot of the code is redundant with `plot/complex_plot.pyx`.   This should be factored out somehow, or imported, or whatever, unless it would lead to gross slowdowns.
    14     * Is there anything different about `complex_to_rgb` except that it's cdef'd (and the changes needed for that, including some efficiencies like the phase command)? 
    15     * Similarly, the `ColorPlot` class is apparently identical to the `ComplexPlot` class, except that it's better documented, uses only the default interpolation, and won't take options. 
    16     * Note that the documentation for `ColorPlot` only uses `complex_plot`!!!   I view this as an error by mvngu and wdj in reviewing #6648, though that was a long review process so it's easy to see how it slipped through.
    1713 * This also seems to have a transposition in how it's put in.  Presumably the redefinition of `I` is supposed to demonstrate it works with lots of different complex types.  I suggest the following.
    1814{{{
     
    2925}}}
    3026 * Add any information at all about theoretical error bounds, if known.  Since not everyone will be able to just look up that paper referenced, it would be helpful to have at least order of magnitude ideas (e.g., if N=2000 on a map from the unit circle to itself, we expect errors no greater than epsilon=blah).
     27 * From #8867: "It now properly avoids failing with lambda functions, although it doesn't work optimally for them. I'll add some notes on that in #10945."