On Jun 10, 2014, at 10:57 AM, Daniel Roßberg wrote:

> Zhao,
> 
> I just tested facetize with a simple region (see the attached file).
> Then I wanted to see how many shells it has produced.  I thought
> looking at the ASCII version of the database would be a good idea
> (created with g2asc).  However, the ASCII NMG has always only one
> shell: rt_nmg_adjust() modifies only the first shell.
> 
> Therefore I'm more and more convinced that we don't need the
> model-nmgregion structures.  Even the work-around with the
> combinations may be unnecessary (but I'm not sure).
> 
> At least the ASCII import and export should be no problem with the
> single-shell nmg.  I'ts ASCII format knows only one shell.

Not a disagreement, but a caution:  most NMG are facetized to BoT for export 
and visualization (using other g-* converters, bot_dump command, shaded_mode, 
facetize, etc).  Someone implementing the ASCII export hook almost certainly 
just put in a quick stub as .asc is not involved in any production workflow 
(not since v5).  I wouldn't use it as a basis for deciding anything.

I would use a poor layman's perspective.  What happens if you chop a sphere 
into quarters by subtracting a "+" from the middle?   Will that facetize (-n)?  
Can it be exported?  With a converter (e.g., g-stl) that runs triangulate? With 
one (e.g., g-x3d) that does not run triangulate?

Zhao, maybe you can find some answers and report back?  Try trunk or a prior 
release and your branch to compare behavior.

Cheers!
Sean

p.s. Note that the g-shell-rect command might be useful for testing, creates a 
single shell manually by ray tracing.


------------------------------------------------------------------------------
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing & Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
_______________________________________________
BRL-CAD Developer mailing list
brlcad-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-devel

Reply via email to