PixelSnap not working on some simple paths - any way to improve it?

[Julien]

Hi,

First, thanks for building this extension! This is exactly what I've
been looking for.

I tested the extension on a few SVG files from The Noun Project
(http://thenounproject.com). Their SVG documents are typically
100x100px and are pixel-aligned, so I resized the objects (for example
to 60x60) and then fired PixelSnap.

(I tried both the version that's bundled in Inkscape, and the download
from Google Code -- however by looking at the Python scripts, they
seem to have mostly converged by now, right?)

It worked in a few cases, however, most of the time, nothing happened,
even with Objects that have lots of horizontal/vertical segments, such
as this one: http://thenounproject.com/noun/inbox/#icon-No406

Is there something I can do to improve the chance of PixelSnap
working?

In the long run, what I would like to do is try and automate the
creation of "pixel-perfect" icons with custom sizes, from the SVG
documents from The Noun Project. Do you think the code for PixelSnap
could be used to do that?

Thanks for reading,

Julien

[Julien]

Hi,

Just thought I'd reference here a very enlightening related article:
http://www.pushing-pixels.org/2011/11/04/about-those-vector-icons.html
and the corresponding discussion happening right now at HackerNews:
http://news.ycombinator.com/item?id=3720363

In my case (monochromatic icons from the Noun Project) I guess font-
face rendering with good hinting support would actually be the way to
go.

Any thoughts?

Julien