Fluid Control's Time log

All times are BST (GMT-1) :

9.00 Just got the theme 'Preperation'. Hmm, it's one of the ones I had a good idea for yesterday but I'm a bit upset that it wasn't evolution. I managed to come up with a great, but difficult, idea and was looking forward to trying to do it in 48 hours. Oh well, perhaps I'll have a go at doing it some other time.

My basic idea for this compatition is to have a 3D surface with some bumps and dips in it and one or more taps and buckets. The objective is to add in more bumps and dips so that the water that flows out the taps arrives in the buckets. Possibly more than one set of water at once and possibly with other obsticles in the way.

10.40 Okay I now have a program that draws a white rectangle on the screen. As major steps go this isnt very much, but it is being done in allegro and OpenGL, neither of which I've used before so it's looking good.

12.40 I've cheated on the OpenGL transforms a bit but I've got an object to move around used user control. I'll try and get the transforms done properly after lunch.

14.10 Okay. Eventually got the transformation done properly and added some frame rate testing information as I am noticing some small jumps in the animation. Currently I am getting about 900 frames per second. Neat!

15.40 Finished makeing the basic rectangle move around now. Small problem with nasty things happening when the user click the Window Manager close button. I've tries everything that I can think of for now, perhaps I'll get back to it later.

I now need to concentrate on the grid simulation, i.e. drawing the grid properly (not just 1 polygon) and making it go up and down when poles / zeros are placed.

17.30 Okay, got a better grid running. Havn't got round to making it go up or down yet but it is now composed of lots of triangles (1089) and it is still running at 500fps.

18.45 I've put some temporary bumps on the surface and they look good, but I need to sort out back face culling. I could just set up a z buffer but I'm reasonably certain that I can do much better by exploiting the grid nature of the system to do it faster.

19.50 Finished the hidden surface removal stuff. Looks execentent. There may be some poping if there are very steep ridges as the division of each grid square into triangles changes depending on the view angle but I don't think that this will be a problem.

23.00 I now have the system so I can scroll about on a virtual grid larger than the visible one. I've also added a perminate slope to the grid to make the water flow.

23.45 I've got a cursor, but I'm not entirly happy with it.

Okay, tommorrows work : Put the real bump function in and allow the user to alter it. Stop recomputing the function as we scroll and just store it all. Put in bits for the tap and bucket. Add in the water level system. Do the water simulation Actually make it a game!

Sleep now!

12.20 I've fixed the visual grid to move across a main grid and not to recompute all the data every move now. I've also make the cursor scroll the grid about. Next I've got to allow the user to alter the grid height.

13.40 Got user control of the grid. Things can now go up and down...

16.15 Damn. This is where I wish I'd spent Friday afternoon looking up FEM of fluid dynamics rather than genetic algorithms and Alife. However, if I had done that I know that evolution would have been picked as a theme.

19.30 Good God! I have eventually got a simulation that is almost exceptable. Except that it's not water but little blue things that really don't behave quite like a fluid...

22.30 Excelent. Got a decent fluid model once I figured out that the system had numerical precision problems. I was loosing more than have the water to rounding errors.

1.30 Okay, got it working, a couple of gauges and a few levels. It also cross compiles. Make it ready to go.