An interesting comparison with web development

[Geoff Canyon]

http://chromakode.com/post/notes-on-xkcd-pixels

That's a description of how XKCD's "Pixels" panel
<http://www.xkcd.com/1416/> was developed using HTML 5 in three days. I
wonder how long it would take in LiveCode.

The basic parameters are:

1. Black and white 600x600 image, zoomable.
2. The zoom must be location sensitive to the pointer, zooming in/out
wherever it is.
3. When the image is roughly 10x zoomed in, replace smoothly (fade in)
replacements for each of the roughly 60*60 = 3600 pixels and continue
zooming.
4. Replacements come from a pool of mostly-black and mostly-white 600x600
images.
5. Replacements are further constrained by the image in which they are
appearing as pixels. So image A might be replaced by White: B,C,D and
Black: Z,Y,X; while B might be replaced by White: C,E,F and Black W,V,U;
and C might be replaced by White: A,B,E, and Black Y,X,W, etc.
6. Replacements must not be the same for each instance of an image; if
image B appears multiple times as a replacement for white pixels in image
A, different instances of B will have their pixels replaced by different
arrangements of images C,E,F for white and W,V,U for black.
7. The whole thing must zoom indefinitely; at least thousands of
replacements deep.
8. The replacements must be in some way deterministic, so the whole thing
can be consistent all the way in and all the way back out.

Apart from anything else, (7) and (8) preclude any sort of simple
coordinate system; a basic X,Y with zoom = Z would break within 8
replacements. Plus I wonder how to break down the data to allow LC to spend
most of its time simply resizing an image.

gc