Why I'm not learning much

Cubey Terra has a piano that you can play freeform - though only one note can be played per av per click. You can have two or more avs playing in sync if you're good enough.

Also, Richard Stern created drumsets that can be played freeform - though limited to the too few key inputs LSL supports... and he's giving them away in Maxim.

I have an, well, I think it is currently in the form of a sphere, but it will allow you to play music by utilizing your computer keyboard and saying keys (ala old school tracking programs). It also will allow a (limited) dump from the Modplug tracker and play it. Currently only has 3 instrument sets though, piano, guitar, and pipe organ.

Yes yes, but honestly there are still better ways to go about a SL musical instrument

For example, an enterprising person should utilize the raw button presses for notes, and utilize the mouse to define stresses and flats. Or how about embedding time-tracked recording and playback into the script itself? Or displaying the notes as sheet music after as well, so people could literally compose their own musical tunes in Second Life?

Come on folks, I'm getting disappointed here.

-find an effective way to make 31+ prim vehicles work gracefully (multimove is the ugliest thing in SL, moreso than crazy sex sims)

/15/da/37403/1.html

You should be able to get 31 prims per joint, so in theory one could use a static animation as a base for that.

Ooh, that sounds interesting. Reckon you'll be selling it any time soon?

Oneironaut Escher has a 3x3x3 Rubik's cube that works. Very neat It works by unlinking and relinking.

Wednesday Grim has a 2x2x2 Rubik's cube that works on child rotations. Also very cool.

Both implementations have a "solver". I believe it works by reversing your moves.

As for the sim-to-sim pathfinder, the most effective solutions I've seen depend on an off-world server. My own solution is a breadth-first search. I considered A* and Dykstra, but they were more "heavy-weight" solutions, and the sim topology is so simple.

I went with a breadth-first search, because of the simplicity of the algorithm, which makes it easier to debug and optimize.

In the end, I don't think it matters what you choose. I think the runtime of all 3 algorithms is O(n) on the number of nodes, and the map of SL is simple enough that it doesn't matter