[Ray-devel] RAY development roadmap (* please read *)

[Wolfgang W. <ww...@gm...>]

Hello everybody!

As we all know, RAY aims to be a full-featured extensible raytracer 
portable among POSIX and Windows with native support for multi-threading 
and network distribution. 

Up to now, quite a lot has been done (mainly by implementing libraries which 
are largely core independent) but we need to go for the main parts now, and 
hence, it is probably a good idea to decide roghly on how to go on. 

This is my suggestion concerning the next steps in the development roadmap. 
I would like to encourage everybody to comment on that. 

* RT core design. Before we implement it, we should get a good idea of 
  the flow of information, flow of execution and required classes and 
  interfaces. 
  
  Andrew offered to do some more overall RT core design / planning some time 
  back and I'd like to kindly invite him especially because he has much more 
  experience about raytracers than me. 
  
  Apart from the "RT core event model", I did not do much more planning, 
  however, I am very confident to nicely implement a slightly modified 
  event model fixing the portability problems. 
  More on that in a separate email. 

* General discussion about rendering on lots of computers: How things should 
  operate, how to distribute information (scene graph,...). 

* VM and Compiler implementation. 
  This is what I'll do in my free time now. 
  After more reasoning, I decided to implement the "real pointer VM" (I 
  probably said that already) which requires a change in the assembly format 
  in order to be portable accross 32 and 64 bit platforms. 
  Hence, I'll re-implement the linker as well (thereby tidying up the 
  code, too). 

I'm still deeply convinced that the project will eventually be a great 
raytracer and that more supporters can be found once it is operating. 
However, as usual, getting it started is hard and that is what we're 
currently going through. 

Yet another important point is the amount of time everyone is able 
or willing to invest. 
I'll have two important exams after this semester and therefore will be 
much more limited in the amout of free time compared to the months before. 
However, as always, I'll do the best I can. 

Regards,
Wolfgang