ws Lv 1
- A lot of people seem to run into the 16-bit color selection bug. Perhaps that should be documented prominently somewhere?
- The chat feature is barely usable on slow computers. It appears to only process one key press per screen redraw, so at around 3 frames per second it takes way too long to input a sentence.
- Consider giving the amino acids appearing in the protein numbers according to the linear sequence and displaying these numbers to help people discuss folds.
- Bending and twisting a protein presumably costs points (energy), but it's hard to tell how much because the energy change from the between-chain interactions dominates the change due to self-interactions when making local changes. It's also hard to tell what the "rest state" of the protein's backbone is. This makes it hard to guess what high-level structure is appropriate. Here are two ideas to help solve this problem:
- Display the bending and torsional stress of each part of the backbone somehow.
- Allow users to temporarily turn off most of the forces, leaving only the interaction of backbone amino acid i with i+1 (for all i); i.e., the forces related to the primary structure. One could use this mode to explore what the self-interaction energy function looks like. </ol>
- The "wiggle" feature seems to do hill-climbing with a little inertia. Adjustable noise of some sort (simulated annealing) might help it get out of local minima. From another perspective, it would be nice to have a gradual tradeoff between human and computer effort. If I step away from the computer for an hour, the computer should keep itself busy folding. </ol>
