[Bf-committers] a question on possible future nurbs-bmesh convergence plans

[Sergey Kurdakov]

Hello all,

while I attempted to start discussion on forum, it brought no response.

Still, starting to look at Blender features, and being impressed by new
opportunities which both bmesh and nurbs branch present I started to look
further.
I might be wrong at places, because I just starting to familarize myself
with Blender and it's codebase, still if I'm completely off, it is easy to
point.

But back to my thoughts, I asked myself is it possible to bring them
together?

Looks like yes, and much details are in this ( 91 Mb  ) course

http://alice.loria.fr/publications/papers/2007/SigCourseGeoProc/modeling-course.pdf

there are even two libraries which might be of  use
cgal ( might be download here https://gforge.inria.fr/frs/?group_id=52 ,
official site www.cgal.org - some code is LGPL but some QPL, but still can
be looked at )
and openmesh http://www.openmesh.org which is completely LGPL

I would like to mention  few sources which nurbs team addressed
http://www.cl.cam.ac.uk/research/rainbow/projects/subdnurbs/nurbswep.html -
source code, paper and demo video
(more docs of the same author are here  http://www.people.usi.ch/cashmant/ )
and which I personally found
http://www.levien.com/phd/thesis.pdf approach to fit curves to points
better
http://hss.ulb.uni-bonn.de:90/2008/1614/1614a.pdf
approach to tessellate trimmed surfaces without visual errors ( code is in
OpenSG scene graph and can be used besides OpenSG whole code)
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.87.4855&rep=rep1&type=pdfconverting
Catmull-Clark Subdivision Surfaces with Bicubic Patches by Loop
C. and Schaefer S.  - can be used to combine both ease of use of subsurfaces
for artists with power of bi-cubic patches.
something here too
http://faculty.cs.tamu.edu/schaefer/research/
( sorry for lot of links it is just I did not seen that there is something
like wiki page or like in this direction, so maybe too much noise, but still
I have just one idea in mind - if the roadmap would be laid out in this
direction ).
 <http://faculty.cs.tamu.edu/schaefer/research/>
looks like a lot of work, still freely blend nurbs and meshes, deform,
simplify these combined models will be quite a great feature for artists.

looks like Blender is almost there, but will there be a roadmap to bring it
all to final desired feature set and what could/should be used , made as
steps to achieve  that ( considering available codebase )?

Regards
Sergey