Knotwork Program

Page 2: Days 4 to 6

All images on this page © copyright T. Shermer. All rights reserved.


Up to Knotwork Index

On to Day 7 (page 3)


Day 6: 06 Nov 2001


SubstanceEarly on yet another pattern. Here I've used the same gradient for the outer part of the threads as I've used for the background; this gives the nice haloing effect.


SubstanceEarly has a parameter known as "earliness" (go figure). This parameter indicates how far into the straight cell one starts the turn. The pictures here show the result of varying the earliness parameter. The top picture has an earliness of 0 (essentially giving SubstanceBasic), the second has an earliness of .22, and the third has an earliness of .5 (the maximum--halfway into the straight cell). Of these, I find the negative shapes of the middle one most attractive, and the border of the bottom one most attractive.



SubstanceEarly applied to the same pattern as shown in Day4a1 below. Note how the horizontal negative shapes are much more pronounced here, but the cross shapes are lost. I plan to add a parameter to SubstanceEarly that allows me to "pull back" sLLs or sRRs segments, which are the ones bordering the cross regions (and also the curves that stick out above and below). I quite like the way those curves stick out further on the top and bottom than the "early" curves, but I do want to see the negative cross shapes. I'm scheming up ways to do both at the same time.



Day6a1 shows a new Substance that I wrote today: SubstanceEarly. SubstanceEarly is so called because it will start some turns early. When it has a "sLs" or "sRs" pattern (see Day5, below), rather than making the turn entirely in the cell for the "L" or "R", it starts the turn in the preceding "s" and ends it in the following "s", giving a much gentler curve. As you can see in the picture, I end up with curves of two different radii: some gentle, some sharp.


Day 5: 05 Nov 2001




There wasn't much to show from Day 5 of coding. Here are some debugging images that show what I worked on--the ability to pick out patterns along a thread. In Day5a1, all sequences of "straight, turn, straight" are shown in black. (Actually, I have to specify left or right in a pattern, so what I'm really showing is all "straight, left, straight" (sLs) and "straight, right, straight" (sRs) sequences.) In Day5a2, I've identified the patterns "sLLLs" and "sRRRs".


Day 4: 31 Oct 2001




Day4b2 and Day4b3 aren't terribly interesting as images, but they do illustrate a new capability for finding, labelling, and linking the different threads in an image, and for assigning different substances to each thread. Here I've just used SubstanceBasic2, with random colors and random widths. Having the threads labelled and linked will allow me to do more interesting things with the geometry.



I snuck in a little time with this before heading off to work. I've written a piece of code that deciphers a set of strings (2d array of characters) into a pattern. For this picture I used the same gradient on the middle of the strands as I used on the background; giving the nice "doubled line" effect. I like the way the negative shapes show up more strongly here than in, say, Day3a5. One of my plans is to work on a Substance that will emphasize the negative shapes even more.


Back to Day 3 (page 1)