It started well as the guy I was ment to be meeting was around 20mins late, I got the impression he'd kinda forgotten about it, heh, always a good start [wink]
Anyways, one of the first things to come up was why do I think I've got the ability to do this, which was a fair question given the guy doesn't know me (he would have taught me in the first year but as I skipped that year he'd never met me before), however this was quickly taken care of when I explained I'd been doing 3D stuff for a few years and I threw in the bit about writing the GLSL chapter for More OpenGL Game Programming. It was intresting to watch his face go from a kinda 'who is this guy' to a 'ah, he knows his' stuff then' look. So, I think any doubts as to if I can do this has been put to one side (although I did hi-light that I'd be doing it with D3D instead of OGL).
After that it was pretty much just making sure we've got the same 'vision' as to where things were going, which we did it seems.
The idea is to compare my work to the work of the guy who did it last year, which means I'm going to need to rip his code out and place it in my own framework and get details out of it.
So, the basic plan of action is;
- build framework
- convert Kristian's code to C++ and the framework (maybe do some SSE work to parrallise things)
- make my own version with TLM on CPU and mesh adjustment performed on GPU
- make my own version with everything on a D3D9 GPU
- make a final version with everything on a D3D10 GPU (software emu to start with).
The final report will be a lovely compare and contrast of the methods used.
Time and hardware availibity might well be the most pressing matters for this project, in reality things shouldn't take that long and I don't have to have it handed in until mid-may so I've got 8 months or so, plenty of time right? *chuckles*
Ofcourse I need to place a loud of research into all this as well, so TLM is a main part of that and probably GPU arch (well, pipeline D3D9 and D3D10, for which there are plenty resources out there).