programming (fast) 3D in Squeak

Andreas Raab andreas.raab at gmx.de
Wed Dec 4 20:08:10 UTC 2002


> Thank you to make the picture much more clear :-))  But why 
> not put such essential information in Swiki ;-)

Feel free to put it up there.

> But this drives me back to another old question: does it make 
> sense to work with balloon deeply,  if  croquet is in the door
> in near future (which has as far as I know other or new concepts
> for let the programmer make 3D ?

No and yes and perhaps. The question doesn't make sense as stated (hey,
I've written this before ;-) 

When you talk about Croquet you talk about a system. Balloon3D (or
rather its equivalent) is but a tiny part of Croquet. Croquet *uses*
certain aspects of Balloon3D and (at least right now) there is no
problem intermixing the low-level aspects of B3D within Croquet. In
particular the concepts used at this level are almost identical since
we're really only talking about a very thin layer on top of a platform
3D API.

Once you start looking into a retained mode framework (aka, "3d objects"
and not just vertices, polygons, material and light parameters) you will
find significant differences. For one thing, there is no "builtin"
retained mode framework for Balloon3D - I always wanted to have
scriptable framework so once Wonderland was there I figured there's
simply no point in writing another one for Balloon3D. Croquet, on the
other hand, will have its own retained mode framework (starting at
TFrame and friends) which will be significantly different from
Wonderlands (in particular more efficient and *much* more powerful). And
while there are certain concepts that are similar (such as having local
coordinate frames ;-) there are others which will be vastly different -
in particular since this is where TeaTime comes into play and some other
deep aspects of collaboration.

Cheers,
  - Andreas




More information about the Squeak-dev mailing list