Lutz Kretzschmar <lut### [at] stmuccom> wrote in message
news:06d1lsok316cotu3bvljlun8osf8ibqals@4ax.com...
> > better yet, a separate tab - to work with the local coords of an object.
> Ouch, that would create a whole lot of other headaches. And then you'd
> have to look at the top of the tabs to see whether you're in local or
> global mode.

Yeah, you're probably right that another tab might be too much...

> Absolutely, please feel free to post suggestions for features, etc.

There are two features that would help speed up modeling significantly. The
first, as I've mentioned, is running object snap modes. To allow the user to
manipulate objects using existing geometry as reference points is tons
easier than having to always come back to the global coordinate system to
make these manipulations. It helps in two ways: accuracy and speed. I know
accuracy is not so important, but the speed benefits are really tremendous.

For example, say I wanted to scale a sphere to the same diameter of a cube's
width. Rather than having to locate the sphere properly by typing coords,
going back to check the actual size of the cube and then type in the
appropriate diameter, I could merely reference the quadrant of the sphere
and snap to the endpoint of the  edge of the cube - boom, its done. (Yes,
this is a little simplified, but you get the idea.) If you operate on the
idea that geometry's relation to itself is more important than its relation
to a global scale or coordinate system while modeling, its easier. You can
always come back and scale it as desired later. And this size accuracy is
really only important in CAD, isn't it? - not so much in the
visually-oriented field of rendering. Relational size, however, is always
important.

Some snap modes could include: endpoint, midpoint, intersection, center,
quadrant, tangent, perpendicular, and nearest. (These are some of the
mainstays in AutoCAD.)

The other feature is what I call slicing. If you could add the ability to
simply slice any primitive along any 2D plane, breaking it into two separate
entities, this would be quite handy for modeling. I realize that you can
already do this, albeit in a roundabout way, with CSGs. What I'm talking
about here is a shortcut way to turn primitives into more complex entities
outside of the CSG method (tho it doubtlessly would be used in conjunction
with it as the modeler desired/required). If this feature is to be
implemented, tho, it needs to be done with efficiency as a high priority.
Ways that the user could define the slicing plane include entering 3 points
(yet another area where the snap modes above could be very useful),
selecting a face of existing geometry, or simply defining an offset of the
three existing x, y and z planes.

Adding slicing along curved surfaces would facilitate fillets and chamfers.
Adding slicing along any bezier patch, if even possible, could add some
interesting capability.

While I'm throwing in thots, how about expanding the extrude entity by
allowing extruding of the 2D outline along a 3D line/spline path, or the
edge of a primitive or entity?

Just some thots,
-Kelly (kla### [at] usacom)

Post a reply to this message