Raul,
I have been getting better resuls with the new
definition of tAkool shown next.
tAkool=:3 :0
assert 0 0 0 1-:3{y
eye=. -3 7 11 {,y %. y*-.(i.4 4)e.3 7 11
scale=.+/&.:*: eye
's up nf'=. 3 3 {.y
center=. eye-scale*nf
eye,center,up
)
There seems to be a flaw still because I always get
0 0 0 = center . For example, when I supply the following
inputs to gluLookAt
eye =. 0 0 6
center =. 0 1 0
up =. 1 1 0
the corresponding M produced (I call it Y) is
Y =. |:4 4$0.707107 0.707107 0 0 _0.707107 0.707107 0 0 0 0 1 0 0 0 _6 1
round =: [ * [: <. 0.5 + %~
0.001 round 3 3$tAkool Y
0 0 6
0 0 0
0.707 0.707 0
The *output* results for eye and up are fine, but
not for center. Do you have any ideas for getting center?
Btw, I don't really understand why the more
complicated version of eye is required, but I have a case
for which the complicated version gives a better answer.
(B=)
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
