[Glmom-devel] Re: HI.. Liu..

[Yaxun L. <yx...@ho...>]

Hi, iceberg,
  Yes, it should be wxWindows. I made a typo.
  Actually I did not program with wxWindows and gtk+
before. Since they lack lots of things MFC has, there
may be some difficulties in the porting. It will be
better if we can find an open-source MFC clone :-)
  It may be still early to consider the porting, but
it may be useful to keep it in mind when introducing
new things in the document class. I think you are
right that we should avoid using doc/view specific
mechanisms.
  I did not use python before. Since you mention it
so many times I feel curious about it :-) I think I'll
learn it and see what's its advantage to C++.
  Currently there are some significant bugs in GLMoM
which need be removed. The first bug is in the camera
and navagation part. If you load bottle.raw you will
know it at once. The object is not located in the
center of the scene, and when you use arrow key to
rotate, the rotation center seems somewhere below the
bottle and it is easily rotated out of scene.
  Previously I use a simple camera/navigation system.
The camera is moved on the surface of a sphere shell,
therefore its position is determined by the 6 parameters,
3 of them is the cartisen coordinates of the center of the
sphere (xc,yc,zc), the other 3 is the sperical cooridnates
of the camera on the sphere shell (r, theta, phi). We can
use array key to change r, theta, phi, use shift+arrow to
change xc,yc,zc. The camera is always look at the center
of the sphere, and the view angle is automatically adjusted
when a .raw file is loaded. This view angle is calculated
by using the size of the bounding box of the objects and
its distance to the camera so that it just fills the whole
view port. The modelview matrix is implemented simply through
glTranslate and gluLookAt.
  Later I need to add a dragging-a-box-then-zooming feature.
To implement this I need to isolate the view matrix and
the model matrix. I made some modifications, which may mess
up the camera/navigation part.
  Another problem is with the background grid. Its size seems
not correct in some cases.
  Now I am considering re-writing the camera/navigation part
by emulating the camera/navigation system of rhinoceros 3D.
I'm also considering using a top/front/side/3D four-view
user interface. It seems some more versatile modelview
functions such as glFrustum is needed, also I need to do
some geometrical calculations. If you have any suggestions
please tell me.
Yaxun

----Original Message Follows----
From: iceberg <ic...@mw...>
To: Yaxun Liu <yx...@ho...>
Subject: Re: HI.. Liu..
Date: Wed, 28 Jun 2000 02:42:05 +0900
MIME-Version: 1.0
Received: from [165.132.116.86] by hotmail.com (3.2) with ESMTP id 
MHotMailBB2383C000A3D820F3A2A5847456050A0; Wed Jun 28 10:41:23 2000
Received: from mwant.yonsei.ac.kr ([165.132.116.90])by mwant.yonsei.ac.kr 
(8.9.3/8.9.3) with ESMTP id CAA04548for <yx...@ho...>; Thu, 29 Jun 
2000 02:49:11 +0900 (kst)