Re: [Algorithms] Cutting Ears Algorithm

[Bill B. <wb...@gm...>]

On Wed, Jan 28, 2009 at 10:11 AM, Jon Watte <jw...@gm...> wrote:
> Bill Baxter wrote:
>> In the situation he's describing he's left with just 4 collinear
>> vertices because of his initial choice of ear to snip.
>>
>
> Great! He's done!
>
> Unless he wants to avoid T junctions. In that case, he'll need slivers.
> Whether you want slivers plus a big triangle, or a number of small
> triangles, probably depend on a bunch of situation-specific parameters,
> though. The naive ear clipping can generate these slivers for him, too
> (by relaxing the "no zero area triangles" rule when there is no other
> ear to clip).

Good point.  Some algorithms are perfectly happy to accept degenerate
tris as input.

--bb