Re: [Fontforge-devel] Conversion to quadratic splines

[George W. <gw...@si...>]

On Mon, 2006-02-20 at 11:35, Alexey Kryukov wrote:
> Well, I understand this is not a bug, but rather a limitation of the
> algorithm you use. 
I don't even understand why it's a limitation. Making something
invisible pretty doesn't seem worth any effort.
> However, the problem is serious enough, because
> it causes all "round" splines (like in "O") look so dirty, that manual
> corrections are necessary before any attempt to instruct them. 
I guess I don't understand why this should be so. I'd think that most of
those points would just be positioned with IUP. And if not they'll just
be moved perpendicularly to the vector to the control points. Behavior
that is the same whether the points are symmetric or not. I don't see
how it's any harder to instruct them. Since there are fewer points I
should think it would be easier.

> And of course you are right that your algorithm produces fewer
> points, but I think pretty point arangement is much more important
> (both for instructing and from a purely aesthetical point of view)
> than reducing the file size.
And that was not even a minor desideratum in the list of things I
thought important.
It's not as though anyone is going to be looking at the points when they
view a page of text.

However. No one seems to share my opinion. So the cvs tree now contains
a conversion algorithm which tries bisecting, trisecting, quadrasecting
... a cubic curve until it finds a good quadratic fit.