I was looking at the behavior of floating-point sin near %pi/2, so I did:
There are several problems with the result.
First of all, the x and y axes are labelled with only 5 digits of precision, so x goes from 1.5708 to 1.5708, and y goes from 1 to 1. Not very useful. Similarly, the coordinates shown in the bottom left have only 6 digits.
Secondly, about 5/6 of the way across the plot, there is a short line *outside* the plotting box. I don't know if this represents a value larger than the other values or some sort of plotting glitch.
Maxima 5.11.0 GCL Windows Gnuplot 4.0