哈哈 把你召唤来了
要像你学习这包。

你用的递归思路很好，可以不用借助外部函数
我也发现了错误，已经修正好了，现在图形的趋势是向右上方倾斜。
如果想数据的点能取得密，y的取值范围要相应增大。

An alternative is to view the equation as a 2-variable function z=y-tan(x)-sin(y)-x and plot the solution of z=0:
Results:
A big drawback is that a lot more data have to be generated than actually used in the plot.
@jmpamao The recursion method is brilliant.

请教
对contour用法不了解，是怎么判断 z为0

另外运行此程序后， 我的显示中间为 白板。怪~

A contour plot is just a 2D way of presenting a 3D plot by drawing curves on which z=f(x,y) has the same value. Here the option "levels = 0" ensures the points (x, y) on the curve has the z value of 0. Normally more than one such curves will be drawn to show different contour curves corresponding to different z values, or levels. And I added the option "labcex=0" and "labels = """ just to eliminate the labels completely. Otherwise there will be "0" or short breaks on the curve which I hate. I have no clue why this happened. But I did have issues running the "outer" function in the scripts where huge amount of memory was required. I had to change the resolution of both x and y dimensions to produce the plot.

x,y的维度，已经改小了。但仍然不行。看不到中间的图

好比现在这个：
my.fun <- function(x,y){y - sin(y) - x - tan(x)}
x<-seq(-5, 5, by = 0.01)
y<-seq(-50, 50, by = 0.1)
z<-outer(x,y,my.fun)
options(max.contour.segments=50000)
contour(x, y, z,levels=0, labels="", labcex=0, xlab = "x", ylab = "y")

Will this work?

给个截图你看吧

请问如果只是画y=cosx的代码是怎样的？