On Sat, 2005-11-12 at 09:56 -0600, aaron@...749... wrote:
The itch I would most like to scratch is fixing the bugs that cause objects to disappear with boolean ops. I can't believe how often I run into these bugs when I try to do real work in inkscape. (I almost never try to do real work.) I've also noticed that newbies often try boolean ops with simple paths to see the results, in many cases simple paths are the ones that cause boolean ops the most problems.
Tracking boolean ops down in doxygen I get about this far before I realize that I don't understand anything: http://www.inkscape.org/doc/doxygen/html/splivarot_8cpp.php#689a3eb52238706a... http://www.inkscape.org/doc/doxygen/html/classShape.php#280456eade08d59a110d...
I'm not going to be able to track and squash these bugs on my own. I have no idea what process to follow or what to look for. But I do pledge to assist in any way the person or persons who will offer to help me.
Unfortunately, in practical terms this is one of the hardest areas of computational geometry to do robustly, and then the livarot code is pretty much incomprehensible to boot.
I don't expect things to get much better unless we decide to write a new computational geometry library from scratch that is very clear, well-abstracted, well-understood, and uses robust methods.
-mental