Ok, hopefully this helps.
Make a complex shaped solid with a loft. You have created a BREP surface topology. B-Rep solids are Euler topological boundaries). This could also be described as manifold surface in a Reimann space. This is a parameter driven model, hence, parametric.
To better understand a parameter model. Make a circle and extrude it. You have a primitive function driven by two parameters u and v. r(u,v)=(acos(u), a sin(u), v) where v is your extrusion length. Parametric Feature based modeling uses primitive geometry like a spheres or torus, NURBS curves and surfaces, it uses parent/child design trees (much like LOSP), and is full of Geometric relations, equations, and dimensioning. NURBS are controlled by knot vectors. Curves can establish Positional Continuity, Tangential Continuity, and Curvature Continuity; (C0, C1,C2) respectively. G0, G1, and G2 continuity is positional, tangential, and continuity respectfully, can be established in surfaces. Solidworks is built for this stuff.
Solid Modeling is actually just a set of feature based surface creation macros which form closed solids.
Now, what you are actually viewing on your display is polygons. Your graphics card takes this geometry and tessellates it. Your display is showing you rendered triangles of the solid or surface body. Here is a good test for you. Take a complex solid -> go to save as STL -> go to options -> move this box out of the way - > and click preview. Change the tolerances and watch the difference in the tessellation. Save this as an .stl. Your geometry is now a bunch of triangles that are a representation of your geometry within a certain tolerance. If you are low on memory you can crash your app just by setting the tolerances very high and tying to view a preview. Now, save the same complex solid as a parasolid or step. Compare. The stl triangles will take a long time to come in, if at all. They may form a solid, they may not. It doesn't matter, because you can’t do much with these triangles in SW anyway other than use the vertices to try create something. The package was not built for large polygon counts, it treats them as bilinear nurbs patches. I tested the actual limits to be 100K triangles in 2004. They have said 2007 will cater to working with scan data, which becomes polygons, which then can be surfaced with the right tools and algorithms. Geomagic and rapid form were built to do this but they can barely stay afloat because the needs for this are minimal........But.....now import the step or parasolid file that you made of that complex geometry. Solidworks is built for this geometry, not large polygon files, and reads this geometry right in. Although, your graphics card is is turning it right back into polygons so you can view it.
People will often say, well, if I convert the file to this format, and solidworks accepts this format then why won't it work. Curves and surfaces can be represented in IGES and DXF, and Triangle files can be represented in Iges and DXF format, they can be represented in Geomagic has its own triangle format. .x, .3ds....there are hundreds of triangle formats.
So why do people use them. Of course, graphics (open GL, direct x). All display is triangles, then gridded as pixels. Triangles can be decimated to reduce data size without noticeable loss in detail. Triangle counts can be reduced and shaders can display them as smooth on your display (New DirectX 10 for Vista, Holy Cow). You can Add texture to them and you have all the geometry for graphics and shaders are applied in the graphics card pipe. TINs are used in GIS data. FEA uses triangle meshes. rapid prototyping, even CAM and CNC is interpolating across these splines or surfaces to a tolerance to create linear moves. There are hundreds of uses. Tessellation of NURBS geometry can be done in a number of methods. Quad tree methods are commonly used. Point clouds can be tessellated with Delaney triangulation and Voronoi diagram methods.
I have tested all the auto surfacing tools in all the reverse engineering packages. None of them are perfect.
What you more than likely have with your car model, is someone who used Catia, or UG, or SW, or ProE, to create a Car, and hit save as tessellating it to some triangle format for the guys using Maya, or Lightwave, of 3ds to put in their animations, games, renderings, etc. These high end graphics guys can Make nurbs and use NURBS too, but they aren't as catering to us MCAD guys.
So, Do what Scott said. Make the model yourself. You will thank yourself for it. Do it in surfaces or do it in solids. Using the stl graphics body to view as a reference helps guide you splines. Solidworks is powerful enough for some heavy industrial design surfacing and concept modeling. It will only keep getting better too because there are so many of us using it. If they want to take over reverse engineering and inspection with this next enginge system then they could if there was real money in it.
I'm done. Thanks for the star Jabberwocky. Hope I'm not coming off as someone who runs my mouth with this long post. CarbonWerkes, I'm just saying what I leaned when I asked the same question as you did and then researched for years about everything out there. Just trying to save you some time on parsing triangle files with solidworks.
A note:
to surface points brought in with the point import macro -use selection filters, view orientations, and insert curve through reference points. Loft these curves. Make surfaces and then trim the surfaces with planes. You can do a lot to surface points in SW. It just takes a lot of patience and fairly organized data.
RFUS
![[pc2]](/data/assets/smilies/pc2.gif "[pc2] [pc2]")
