On Apr 3, 2012, at 9:00 PM, Edward wrote:
> I have submit my proposal.
>
> http://www.google-melange.com/gsoc/proposal/review/google/gsoc2012/plussai/1
> would you please give me some advise or comment,thanks.
Thanks for sharing it Laijiren! You know, I laughed a little when I read your
comment that you dream you can one day make a contribution to an open source
project. You can do that ANY day, *any* time, regardless of GSoC or job or
school! Contributing to open source doesn't require permission or even
approval.
You proposal looks to have a good level of detail. Some minor implementation
details specified will probably change (such as creating a "draw mesh" tool in
MGED -- already exists), but on the whole I like how you tie your approach to
the research that followed our original implementation. There's probably some
more research papers I can get you on NURBS tessellation as well.
The one major technical aspect that I didn't see covered is solidity. As a
solid modeling system, guaranteeing the preservation of solidity is usually
implicitly more important than speed. Basically, you don't have any
accommodation in your proposal for verifying and validating the accuracy and
fidelity of the mesh being generated.
Speed is certainly important, but it's more valuable (to our users) to first
have a slow "correct" tessellation that produces fully closed solid geometry
before having a fast tessellation that is only suitable for visualization
purposes. Both are good to have, but I'd propose including some steps and time
for ensuring proper mesh closure. Maybe do a few quick nurbs paper searches to
see if you can find any recent ones that talk about solidity, or just one or
two additional NURBS paper references that speak of quality and edge alignment.
In your timeline, it'd be nice to see more detail in June. That's arguably the
hardest part but there aren't any specific "stepping stone" deliverables
identified to indicate whether you're making adequate progress. Otherwise, the
detail looks good there.
Cheers!
Sean
------------------------------------------------------------------------------
Better than sec? Nothing is better than sec when it comes to
monitoring Big Data applications. Try Boundary one-second
resolution app monitoring today. Free.
http://p.sf.net/sfu/Boundary-dev2dev
_______________________________________________
BRL-CAD Developer mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-devel
