Subject:
|
Type 5 lines at the edge of primitives
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Sat, 4 May 2002 23:37:18 GMT
|
Viewed:
|
452 times
|
| |
|
|
I'm not sure if this has been discussed before, but I discovered a problem
with the use of type 5 lines at the edge of primitives. The basic problem
is that there is no way for the primitive to know which direction any
attached polygons are going to go, so no way for them to include proper type
5 lines.
A good example of this is the sphere primitive. The included type 5 lines
at the edge are ... well ... unique. Look at
http://home.san.rr.com/tcobbs/LDView/test2.gif and
http://home.san.rr.com/tcobbs/LDView/test3.gif to see what I'm talking
about. In these shots, a line is drawn from the first point of each type 5
line to its two control points. (In addition, all type 5 lines are drawn,
not just the ones that should be.)
The problem is that they won't be drawn correctly if you (say) put 4 of
these together to form a hemishere. (They'll be omitted from some angles
when they should have been there.) On the other hand, if you set it up so
that a hemisphere works, then they'll be drawn when they're not supposed to
be if you put the hemisphere on top of a cylinder.
Unfortunately, the only solution I can think of is to remove the type 5
lines from the edges of all the primitives and then add them back with a
correct 4th control point in all places they get used. Obviously this would
be a HUGE task, and not one we are ever likely to undertake.
In case you're wondering, I came across this little problem after adding
type 5 line generation to my sphere primitive. Don Heyse successfully
transferred the type 5 line drawing code from LDGLite into the LDView
source, and I'm adding code to generate type 5 lines for primitives. The
type 5 lines I generate at the edge of my eighth sphere are consistent with
a continuation of the sphere. This means that a full sphere will always
look right, but a hemisphere on top of a cylinder will have an outline
around the join when looked at from a steep angle. (You can look at the
minifig air tanks for an example once I've released the next LDView version.)
--Travis Cobbs (tcobbs@REMOVE.halibut.com)
|
|
Message has 1 Reply:
 | | Re: Type 5 lines at the edge of primitives
|
| (...) I'm not sure if I understand this properly - if the second "test" point is down the line where the cylinder joint (between two rects) would be if it was attached to a cylinder, won't that work? It'll mean the same line may occasionally be (...) (22 years ago, 5-May-02, to lugnet.cad.dev)
|
12 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
