What a wild ride this project has been. Creating the characters was one thing - getting them to interact was something completely different. Considerations needed to be made for the complexity of the characters - each one a jenga stack of modifiers just waiting for an excuse to say "I don't want to play today!" Once everything was set up progression was reasonably steady. It all seems worth it in the end when we look back on what we've achieved though.

This flash uses an experimental "loop pool" system whereby there are a bunch of loops in each stage, and one is picked at random - against a few percentage values for the sake of a little order. So it uses fuzzy logic to go from loop to loop - each one being a variation on the current stage. This breaks up the monotony of the same loop playing back to back forever.

The events are run-once loops you activate manually. They give a sense of involvement with the character. When you activate one, it plays once before going back to the current loop pool.

The quick menu allows you to jump to a stage immediately. This is in contrast with the keyboard shortcuts, which play to the end of the active loop before progressing via transition to the next stage. So if you want to jump around quickly, use the menu. If you want a more cinematic, fluent experience, try out the keyboard.

Sounds were scavenged from the 'net, as usual. We still don't have someone to do lines for Bridgett.

We tried simulating physics for the hair objects, but the simulations didn't work and time was too pressing (the deadline was S.G.'s Birthday) so they were omitted. We look forward to implementing physics in the future.

Photorealistic lighting was not available due to incompatible technologies. To rectify this, we're dropping one solution in favour of another - thus we should be able to create even more realistic-looking animations. Global illumination techniques such as ambient occlusion will be employed to provide a lighting model which closer resembles real life.

Shadows weren't included due to time constraints - it would have taken too long to render. As it was, the animation took a couple of weeks to render, and this was 24 hours a day, 7 days a week. The workstation used to render it only has a single processor at 3.4Ghz and 1Gb of RAM. The largest amount of work involved is in rendering the hair and fur - it bumps up the rendering time by about five times, and that's without shadows. Shadows bump it up about 4.5 times again, as each individual strand needs to cast a shadow, which can hit many more strands before they hit the base model.

We plan on doing more stuff in the near future, though you shouldn't get your hopes up. While it's not so difficult to create the animations, it's the hardware we have that's restricting us. So what does everyone think of the new animation? We'd love to hear from you as we thrive on feedback.