[Gzz-commits] journals jvk

[Janne V. Kujala]

CVSROOT:        /cvsroot/fenfire
Module name:    journals
Branch:         
Changes by:     Janne V. Kujala <address@hidden>        03/07/04 14:38:43

Modified files:
        .              : jvk 

Log message:
        this week

CVSWeb URLs:
http://savannah.gnu.org/cgi-bin/viewcvs/fenfire/journals/jvk.diff?tr1=1.52&tr2=1.53&r1=text&r2=text

Patches:
Index: journals/jvk
diff -u journals/jvk:1.52 journals/jvk:1.53
--- journals/jvk:1.52   Sun Jun 29 06:09:06 2003
+++ journals/jvk        Fri Jul  4 14:38:43 2003
@@ -46,6 +46,40 @@
 Past:
 =============
 
+2003-07-01 - 2003-04:
+    - fillets 3d blending work (change modes with "m"/"M", see also "i")
+       - Implemented an interpolated lookup table for computing the distance
+         from the center of the node to the fillet at a given direction
+       - Trying Dicer:
+               - Starting from a Delaynay triangulation based topology 
+                 (mode 3)
+                       - non-dicable polygon rings at the connection mid 
+                         points
+                        - for each Delaynay triangle (consisting of three
+                         connection directions), the relevant spans of the
+                         fixed rings of each connection are connected
+                         to form a polygon, which is triangulated as a 
+                         triangle-fan (the triangle mid-point is added)
+                       - dummy connections added as necessary
+                         to surround a three-dimensional volume
+                       - Dicer seems to have trouble with the thin triangles
+               - Also tryid Icosahedron based topology (modes 4 & 5)
+               - Dicer seems very slow
+       - Another approach (modes 0 & 1)
+               - Draw each filletoid with the cylinder tesselation 
+                 but blend each point with all overlapping filletoids
+               - preventing too large triangles:
+                 those quads whose corners are "deeper" than than
+                 the surface of some other filletoid are not drawn
+    - spiderman fillets using a texture ("x" key):
+       - the thinnest part of the connection looks bad because 
+         stretching distorts the line-width of the grid anisotropically
+        - computing lower-detail mipmap levels with thicker
+         liners produces the same effect as the line polygon mode:
+
+2003-06-30:
+    - fillets npr edge work and abstracting
+
 2003-06-26 - 2003-06-29:
     - fillets work with tjl
        - 3D fillets without shape blending