Gradient meshes in Scribus and Inkscape
In case you don't read all these fancy blogs, Scribus 1.5.0svn now features gradient meshes support. Logically support for them was added to the Adobe Illustrator importer in Scribus as well.
The feature is implemented largely thanks to gradient-mesh branch of Cairo (Scribus's SVN now contains its copy). In order to properly see what you do you will need Pixman v0.15.16 or newer. You don't need exactly that version of Pixman for the output to PDF: it will just work.
Of course, then very next question is whether we really need it in Scribus in comparison to, say, Inkscape or sK1. Last December there was a lengthy discussion about gradient meshes in inkscape-devel@ mailing list which resulted in quasi silent agreement not to do anything about it just now.
Three main reasons for that, as I see them, are:
- There is no support for meshes in SVG specification, and it is not recommended to implement it as extension because of lack of a reasonable fallback.
- There are different approaches to meshes, none of them ideal usability-wise.
- Jasper van de Gronde, who wrote SVG spec compatibility tests for Inkscape, is rather for diffusion curves.
As a matter of fact, in just one week Jasper is talking about diffusion curves at Libre Graphics Meeting 2010 in Brussels. Apparently, he has finally managed to solve basic issues of mating diffusion curves and SVG (of which he's been talking since last year).
Now, one fun thing about diffusion curves is that the original application (written in Qt) that demonstrates the technology actually uses some of Inkscape's icons from v0.37 or so. One of the researchers actually told me two years ago that some people in the lab used Inkscape for various tasks.
The other fun thing is that two years later this commercially tempting technology still does not seem to be used in any existing application, while the research goes on (the team of researchers is different now).
The next release cycle (0.48 is due this summer) for Inkscape is supposed to be rather long, so in case Jasper solves all other SVG related issues, there will be plenty of time for him or somebody else to implement diffusion curves for 0.49.