I’d like to do a song of great social and political import.
The code that created the illustrations in Tao Te Programming is now available as the TaoTeProgramming package on CRAN.
The sea is just flat hills

And they are both just smooth mountains.
The package contains just over 1000 lines of code. More than half of that is the functions that create the specific pictures in the book. All of the real functionality is produced with about 450 lines of R.
You might think this is because I followed the principles in Chapter 21 to get complexity by writing simple code. Actually it comes from laziness (Chapter 26) — if something wasn’t working well with minimal effort, I threw it out. Mostly it’s randomness that’s the powerhouse.
One call to make them all
The next 6 figures are all produced by the exact same code. The only difference is the random seed that is used.





Yet more
This post comes under the domain of Chapter 65: Play.
Epilogue
That’s it. He-he-he …
— Janis Joplin
Appendix R
Here are the functions (in no particular order) that produce the pictures above (WordPress is not especially good at preserving the formatting).
require(TaoTeProgramming)
P.polytilepurple14 <- function (filename = "polytilepurple14.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
polytile(color=grep("purple", colors(), value=TRUE), seed=14)
if(length(filename)) {
dev.off()
}
}
P.quadtileorange4 <- function (filename = "quadtileorange4.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
quadtile(color=grep("orange", colors(), value=TRUE),
dims=c(2,20), seed=4)
if(length(filename)) {
dev.off()
}
}
P.qtbalturquoise37 <- function (filename = "qtbalturquoise37.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
quadtilebalance(color=grep("turquoise", colors(), value=TRUE),
seed=37)
if(length(filename)) {
dev.off()
}
}
P.darkroad1 <- function (filename = "darkroad1.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
roads(color=grep("dark", colors(), value=TRUE), lwd=2:3, seed=1)
if(length(filename)) {
dev.off()
}
}
P.wheatwave9 <- function (filename = "wheatwave9.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
waves(color=grep("wheat", colors(), value=TRUE),
lwd=2, seed=9)
if(length(filename)) {
dev.off()
}
}
P.violetstick54 <- function (filename = "violetstick54.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
sticks(color="violet", lwd=2, seed=54)
if(length(filename)) {
dev.off()
}
}
P.bendsienna0 <- function (filename = "bendsienna0.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE),
seed=0)
if(length(filename)) {
dev.off()
}
}
P.bendsienna1 <- function (filename = "bendsienna1.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE),
seed=1)
if(length(filename)) {
dev.off()
}
}
P.bendsienna2 <- function (filename = "bendsienna2.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE),
seed=2)
if(length(filename)) {
dev.off()
}
}
P.bendsienna3 <- function (filename = "bendsienna3.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE),
seed=3)
if(length(filename)) {
dev.off()
}
}
P.bendsienna4 <- function (filename = "bendsienna4.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE),
seed=4)
if(length(filename)) {
dev.off()
}
}
P.bendsienna9 <- function (filename = "bendsienna9.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE),
seed=9)
if(length(filename)) {
dev.off()
}
}
P.tree66 <- function (filename = "tree66.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
treeColor(seed=66)
if(length(filename)) {
dev.off()
}
}
P.hills5 <- function (filename = "hills5.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
hillscene(color=grep("green|yellow", colors(), value=TRUE),
seed=5)
if(length(filename)) {
dev.off()
}
}
P.sea51 <- function (filename = "sea51.png")
{
if(length(filename)) {
png(file=filename, width=512)
par(mar=c(4,5, 0, 2) + .1)
}
hillscene(tilt=0,
color=grep("aqua", colors(), value=TRUE), seed=51)
if(length(filename)) {
dev.off()
}
}
random pictures
Here is a function that will draw a random picture from the styles of the ones done in this post (plus a couple more). That is, the type of plot is selected randomly, and then a random picture of that style is drawn.
taoteprog <- function()
{
require(TaoTeProgramming)
cmds <- expression(
bendplotmultcol(xdelta=100, ydelta=200, sd=.001,
color=grep("sienna", colors(), value=TRUE)),
roads(color=grep("dark", colors(), value=TRUE), lwd=2:3),
polytile(color=grep("purple", colors(), value=TRUE)),
quadtilebalance(color=grep("turquoise", colors(), value=TRUE)),
quadtile(color=grep("orange", colors(), value=TRUE), dims=c(2,20)),
treeColor(),
sticks(color=sample(colors(), 1), lwd=2),
waves(color=grep("wheat", colors(), value=TRUE), lwd=2),
hillscene(tilt=0, color=grep("aqua", colors(), value=TRUE)),
hillscene(color=grep("green|yellow", colors(), value=TRUE)),
butterflies(color=colors()),
chessboard(sample(colors(), 1))
)
eval(cmds[sample(length(cmds), 1)])
}
jun 25, 26
The function in question is scriptSearch. I’m not much for superlatives — ”most” and ”best” imply one dimension, but we live in a multi-dimensional world. I’m making an exception. The [...]
jun 25, 26
I recently gave a talk at the R in Finance conference in which I introduced the marketAgent package for R. Here is the source for the package if you'd like [...]
jun 25, 26
Executive summary Surprisingly good. And it’s not like my expectations were especially low. Structure There are 20 chapters. I mostly like the chapters and their order. Hadley breaks the 20 chapters [...]








