The code is stand-alone (although the aem stuff has been more or less coded with Inkscape in mind) and does not accept file based definitions, the images are hard coded (although relatively easy to alter, I just didn't get around to writing a parser). For integration with Inkscape there are still quite a few things missing, like syntax :) Also, I don't think the method I developed is mature enough to put in Inkscape as is. In my opinion the quickest route to get this in Inkscape would be to figure out a syntax+semantics (how to deal with "exterior" boundaries!) and then write a hybrid solver that uses an analytic solution for dealing with discontinuities and a grid-based solver for the rest.

Another question is whether diffusion curves are really the "best" way to go about this. There are some cases where diffusion curves might not give the most intuitive answer and I am looking at alternatives (and so are some other people). If you have specific ideas about what a system like this should support and/or have interesting cases where you feel diffusion curves do not meet your requirements (or where they perform surprisingly well), then please speak up.

BTW, if you want to play with diffusion curves, use one of the earlier programs to edit/view them: http://artis.imag.fr/Publications/2008/OBWBTS08/

Also, we have some ideas on the wiki, feel free to augment/edit them: http://wiki.inkscape.org/wiki/index.php/Diffusion_Curves

On 2011-03-07 17:47, Josh Andler wrote:

Hey,

I'm a little confused by the code page... is there a "test" version of this integrated with Inkscape or 2geom toy or something? Either way, wowza to the math in your thesis paper. :D

Cheers, Josh

On Mon, Mar 7, 2011 at 1:35 AM, Jasper van de Gronde < th.v.d.gronde@...528...> wrote:

...

Hi,

For those interested in Diffusion Curves, I recently finished my thesis on Diffusion Curves. It can be downloaded here: http://home.hccnet.nl/th.v.d.gronde/ThesisJvdGFinal2.pdf (Initially it was also available from my university's site, but there are some problems with the repository at the moment.)

Associated code can be found on LaunchPad: https://code.launchpad.net/~jaspervdg/

In short: I found that it's not that easy to render diffusion curves with high accuracy and looked at the cause of this, as well as possible solutions. In particular, I implemented a renderer/solver based on a different concept than the previously existing solvers, and this does show some promise. Also, I found a way to make it easier to handle the color discontinuities across diffusion curves (which are essential to the usefulness of diffusion curves but can be tricky to handle correctly in a solver), essentially by first removing them and then adding them back in.

Comments, suggestions, questions, etc. appreciated.

• Jasper

BTW, most graphics were made using Inkscape (often saving to PDF+LaTeX).

---

What You Don't Know About Data Connectivity CAN Hurt You This paper provides an overview of data connectivity, details its effect on application quality, and explores various alternative solutions. http://p.sf.net/sfu/progress-d2d _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel