The examples in this notebook pass y0 and y (or x0 and x); I’m tempted to rename these options to y1 and y2 respectively, and to remove the other y-options (leaving only y1 and y2). Do we need four y-options? Also, I wonder about using the chart dimensions (the last argument to render) instead of computing the extends eX and eY.

I think both ideas make perfect sense. (We don't need the four ys, it's just that I was reading d3 code.)

done!

Removed the clip option, since it's equivalent to switching y1 and y2