Consider several new improvements by Shel, for possible inclusion

[oskay]

Several improvements made by @ShelMi http://forum.evilmadscientist.com/discussion/comment/2206#Comment_2206

[ShelMi]

I've been continuing to have fun tweaking stipplegen to meet my own conception of how I'd like it to work. Everything about the algorithm is totally unchanged - except for the initial choice of stipple points. The gui has been fiddled with a bit, however.

The only demonstrably worthwhile change is in the randomization method of the initial point selection, This version, 2.40.3.9, uses a low-discrepancy sequence rather than a random generator to place the initial stipple points. https://en.wikipedia.org/wiki/Low-discrepancy_sequence

My guiding principle in the gui changes/additions has been to make it possible for the user to visualize and modify the final stipple pattern accurately from within stipplegen. The changes are documented in the Change log at the start of the file.

Here is the pde, along with a sample options .json file (yes, it's true, you can now save the slider settings):
grace.zip

Note that this version has not been completely cleaned up, nor thoroughly debugged, but is perhaps reasonable as a beta preview. Though Murphy often has the last laugh.

Here is a test look at low-discrepancy vs. random seeds, performed on a uniform 50% gray image, in the first generation:

And here good-old "grace.jpg" as a comparison of random vs. low-discrepancy point selection methods. Best viewed at full resolution to appreciate the differences: