[Bf-committers] Freestyle branch status report May 2012

Tamito KAJIYAMA rd6t-kjym at asahi-net.or.jp
Tue May 22 00:51:56 CEST 2012


Hi all,

The scope of this message is to report the up-to-date status of the Freestyle branch and
discuss the possibility of the merge into the trunk.

Since the January 2011 status report [1], the Freestyle development has focused on
stability improvements and the completion of selected unfinished components
including an artist-friendly GUI for interactive manipulation of Freestyle parameters.

The stability of the Freestyle renderer is the highest priority of the project.  Thanks to
intensive testing by branch users, many instability issues were sorted out and important
bug fixes were made, such as a fix for occasional straight lines [2], improved handling
of face smoothness in view map construction [3], and a fix for degenerate triangles in
imported mesh data [4].  It should also be highlighted that substantial performance
enhancement of view map construction was contributed by Alexander Beels [5].

The new Freestyle GUI has been improved in many aspects, including the addition of
new stroke geometry modifiers (e.g., [4, 6, 7]), and easy duplication of line sets and line
styles [7].  Moreover, new edge and face attributes (Freestyle edge/face marks) were
introduced for fine-control of feature edge selection [3].

As mentioned in the January 2011 report, the Freestyle integration project listed three
unfinished components that were planned to be done before the merge into the trunk
is asked:
1. The aforementioned artist-friendly GUI for Freestyle parameters (90%)
2. Freestyle Python API Improvements (80%)
3. Feature edge detection at face intersection (0%)
Percent values in the parentheses are subjective achievement rates as of this writing.

In the meantime, I have been receiving an increasing number of positive opinions from
branch users about the merge.  In fact, the time when the merge will happen is the most
FAQ of the branch these days.  Aware of the documented incompleteness of the branch,
however, I personally have a mixed feeling.  Here are some of pros and cons.

Downsides of the merge:

- The feature edge detection at face intersection is a must but not implemented yet.

- Documentation is not ready.

Upsides of the merge:

- More testers of the Freestyle renderer are expected, which will accelerate stabilization
and improvements of the renderer.

- Branch users no longer have to rely on multiple versions of Blender, which makes their
workflow simpler.

- More development force can be addressed to the unfinished and other Freestyle
components.  So far a significant amount of time has been regularly spent for merging
changes of the trunk into the Freestyle branch, which is not so productive.

In conclusion, I tend to think that the merge is a good idea even at this stage of
development.  The Freestyle branch is fairly stable and useful.  It has been kept well
synchronized with the trunk.  For these reasons, the merge can be made at any time.

I am open to discussions about whether or not the merge should take place right now.

Any feedback on this matter would be much appreciated.

With best regards,

-- 
KAJIYAMA, Tamito <rd6t-kjym at asahi-net.or.jp>


[1] http://code.blender.org/index.php/2011/01/freestyle-integration-report-january-2011/
[2] http://freestyleintegration.wordpress.com/2011/09/25/development-updates-on-september-25/
[3] http://freestyleintegration.wordpress.com/2011/10/25/development-updates-on-october-25/
[4] http://freestyleintegration.wordpress.com/2012/04/03/development-updates-on-april-2/
[5] http://freestyleintegration.wordpress.com/2011/03/18/development-updates-on-march-17/
[6] http://freestyleintegration.wordpress.com/2011/08/28/development-updates-on-august-28/
[7] http://freestyleintegration.wordpress.com/2011/12/19/development-updates-on-december-18/


More information about the Bf-committers mailing list