> On Nov 14, 2016, at 1:21 PM, Michal Hanus <mikeha...@seznam.cz> wrote:
> 
> In my opinion, BRL-CAD does not have to replace the slicing software, which
> is quite complicated and specialized. Direct g-gcode export would thus not
> be necessary for 3D printing. An error-free export g-stl would be enough for
> using BRL-CAD with any type of a 3D printer.

Mike, it’s very true that BRL-CAD does not have to replace the slicing 
software.  I don’t see BRL-CAD having good gcode export or direct printer 
integration soon also, as you noted there are other tools that do this already. 
 Consider, though: there is not any CAD system that guarantees error-free STL 
export.

There’s not even one that does an exceptionally good job at it.  Pro/E, NX, 
CATIA, etc, all fail, often spectacularly for seemingly simple cases.  The 
general rule of CAD is to not move between systems and if you must, to avoid 
format conversions too.  Longer-term, CAD software is in a far better position 
to slice as that is where the actual geometry resides.

Everyone reads STL because it’s simple, not because it’s good.  Converting from 
CAD geometry to STL invariably results in errors (e.g., invalid geometry due to 
new overlaps).  A primary reason STL took off for 3D printing (aside from 
makerspace hobby coders being able to implement STL easily) is because printer 
resolution was far more coarse than triangle tessellation errors.  That’s 
quickly changing with the latest printers.

For BRL-CAD, we still need to strive for some of the best polygonal export 
regardless, so getting STL robust is still needed.  The best option for that is 
to finish our NURBS Boolean evaluation project.  There’s a great doc in our 
repo detailing how to sort through the remaining issues, and that should give 
us the best possible export capability.  Separating and cleaning up our NMG 
library and replacing its Boolean evaluation support with a different 
implementation is a close second.  For absolutely robust, a point sampling 
approach based on ray tracing might take the cake near-term.  Alas, they all 
need investment of effort. :)

Cheers!
Sean


------------------------------------------------------------------------------
_______________________________________________
BRL-CAD Developer mailing list
brlcad-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-devel

Reply via email to