Hi Doug,
I suspect (in this case), I'd like to build on/leverage plot, because I'd
like things to place nice in DrRacket's REPL (for example), and be able to
play with the plot ecosystem. To save you any work, I'll circle around
later if future me thinks current me was wrong.
This is crossing wit
The plot library creates plots by assembling elements (such as axes and
renderers) together. For example, the plot library does no have a scatter
plot type, instead points can be rendered on a plot, but lines could also be
rendered on the same plot. You can also control many other things, such as
Matt,
I have done many extensions to plots for a variety of reasons. Most of my
extensions are a third option: Build on dc%. I typically make a new widget,
say based on a horizontal or vertical panel with one or more embedded
canvases. Then I use plot/dc to render directly on those canvases. I can
Hi all,
If I want to develop a new plot type, would I do best to:
1. Build on plot, or
2. Build on pict?
I suspect #1.
If I want to (say) have a number line, I would like to have an x-axis,
centered in my plot area (vertically), and no y-axis. I've been reading the
plot code for scatter plots,
The sets are about 20 elements in size.
It seems that ‘member’ before ‘cons’ is marginally faster than
‘remove-duplicates’ after ‘cons’, as one would expect.
But it’s the strangest thing: in some cases switching from sets to lists with
unique elements changes the meaning of the program! I spent
I suspect it will be slow because sets are generics, and generics are slow.
For my application, it has worked well to replace set/seteq with
hash/hasheq mapping to #t; this only works when you have total control over
set representation as an implementation detail, of course! But for me it
sped
6 matches
Mail list logo
