I want to recognize the "object" im clicking on and forexample
reference textual information belonging to that OpenGL
primitive / point / line / etc.
Is there a way to do that, that i am missing?.

cognacc wrote:
I want to recognize the "object" im clicking on and forexample
reference textual information belonging to that OpenGL
primitive / point / line / etc.
Is there a way to do that, that i am missing?.

Although OpenGL provides a so called "selection mode", I strongly discourage
its use. Instead perform ray-intersection tests yourself:

   Inversely project your mouse position (gluUnProject) at Z_near and Z_far,
this gives you the vector of the ray, your mouse pointer is casting into the
scene.

   Test this ray against (all) the objects in the scene. To speed things up,
manage your objects in a space subdivision structure (Octree, Kd tree, BSP).
These spatial subdivision structures provide very fast tests through which
object's bounding areas the ray passes.

For simple object selection (like
files in a file manager), this simple test normally is sufficient.
   For more fine grained intersection tests, work down from easy (=fast) to
hard (=slow). In this order: Ray<->Bounding Sphere, Ray<->Bounding Box (Axis
Aligned), Ray<->Convex Hull, Ray<->Triangle Soup. The last two tests can of
course be again accelerated by sorting the triangles into their own spatial
subdivision structure.

OPCODE homepage: <http://www.codercorner.com/Opcode.htm

I think i'm to newbie in OpenGL to go there just yet. :)

Well, that's not OpenGL at all. It's only math and programming and nothing
OpenGL or other renderers. Though you will find such math/code oftenly in
close proximity to OpenGL or other renderer's code.

I was good at vector calculation/operations, but that was some years
ago (2-3)

I think i'm to newbie in OpenGL to go there just yet. :)

Well, that's not OpenGL at all. It's only math and programming and nothing
OpenGL or other renderers. Though you will find such math/code oftenly in
close proximity to OpenGL or other renderer's code.

Well i (kinda) know that, but some times just adding another complex
subject at
the same same adds to the confusion, so i start out small.
And then build / add on that later on.
I was good at vector calculation/operations, but that was some years
ago (2-3)

Regards michael

Excellent answer. Always use the simplest thing that works for you. Octrees
or whatever for a barchart is way too much code, wayy too much chance to
make a mistake. If you only want to know which bar they are pointing at,
that's even simpler to do -- just keep track of the min/max for each bar &
compare to the gluUnproject() of the mouse position.

Excellent answer. Always use the simplest thing that works for you.
Octrees
or whatever for a barchart is way too much code, wayy too much chance to
make a mistake. If you only want to know which bar they are pointing at,
that's even simpler to do -- just keep track of the min/max for each bar &
compare to the gluUnproject() of the mouse position.

jbw

You only need a stack depth of 2. You don't keep pushing the name stack.

That total number of entities can probably be picked with GL_SELECTION in a
trivial amount of code & time.

You only need a stack depth of 2. You don't keep pushing the name stack.

That total number of entities can probably be picked with GL_SELECTION in
a
trivial amount of code & time.