Re: [ft-devel] Fast bisecting BBox_Cubic_Check

[Alexei Podtelezhnikov]

On Sun, Feb 24, 2013 at 1:08 AM, Werner LEMBERG <address@hidden> wrote:

> Attached are more thorough tests. Each line contains bboxes for the
> the current and proposed algorithms as well as their difference.
> The bboxes are for two popular fonts at size 24 with rotation of 30
> degrees as dumped from modified ftbench.  You will see that the
> differences rarely reach 3/64.

Great! ... Have you compared with, say, a
floating point implementation which returns bboxes with really high
precisions?

I anticipated this question. Why would you compare with theoretical results when your rasterizer uses discrete and less accurate bisections? Shouldn't you compare with the rendering results? One way or the other 3/64 of a pixel is really small difference.