Programmer 7500/MAX_PROGRAMMERS.iso/INFO/IRIT/IRITS.ZIP/WISHLIST.DOC

Text File | 1990-05-05 | 1.7 KB | 37 lines

This wish list will be updated according to the current releases.

This is release 2.1.

   The main problem in IRIT which used to be the convex.c module, has been
solved (this module was totally rewritten). No error will result in
non convex to convex decomposition (at list I didnt find any).

1. If two objects have no intersection in their boundary, IRIT assumes they
   are disjoint: in union simply combines them, and in other operators return
   NULL object. One should find FAST way (3D Jordan theorem!?) to find the
   relation between the two (A in B, B in A, A disjoint B) and according to
   that decide. You probably wants to look at Bool-Hi.c module.

2. Co-planar polygons that intersects, are NOT supported, and the result is
   undefined. The lower level Boolean operations should be modified (a lot)
   to support such case. This is a hard challenge.

3. Interactive profile editor. One can create objects from profiles (much
   as IRIT was created in solid0.irt) by extruding them or by using them in
   surfaces of revolution. Currently the profile is created manually, and
   having a simple interactive editor to manipulate the profile will be neat.

4. The polygonal representation of the created object has no real vertices
   normals, so that Gaurd/Phong shading can be used. Such improvment will
   increase the memory usage significantly (each vertex will need to carry
   a normal as well). A minimum memory usage method (especially important
   for the IBMPC) should be developed.

This list is open, so if you find you really want IRIT to have something, but
you do not have the time to implements it, send me mail, and I will add it to
the list.

Gershon

[e-mail]