On Fri, 6 Nov 1998, Sean Breheny wrote:

> At 04:47 PM 11/6/98 +0000, you wrote:
> >On Thu, 5 Nov 1998, Sean Breheny wrote:
> >
> >> when editing that code. The circle it produces is a bit rough, with a large
> >> step size between pixels. I guess that could be changed with some editing
> >> of the code. I really need to take a look at it to see how exactly it
> >> works, when I get some time (yeah, right <G>)
> >
> >Since you are so unhappy with the rough circle, why not plot another
> >circle using BASIC's sin/cos right on top of it for comparison. Tell me
> >what you see ? ;)
>
> Hi Peter,
>
> When I do that, I DO get a better looking circle. However, I don't really
> blame this on the algorythm, but on the constants used in this particular
> implementation. If they are changed, I bet that the step size can be changed.

How did you plot with sin/cos ? dots, back-joining lines and closure, what
step did you use ? Using the machine's native circle command is cheating
because it is optimized to look good not be 'right' vs. sin/cos.

Peter