foldit is way coolEdit
modify the design of the thumb
modify the structure of ther anvil
there is alot of room to modify and customize, to create custom variation.
These may not be to the point.
Maybe there are better examples of emerging structure.
But we think it starts to emerge here.
Just complex enough to be predictable. We never really thought about this before. Quantum numbers went 1, 2, 3, 4, ... classical limit. But here we begin to see that as quantum complexity increases by orders of magnitude stability and predictability increase too. They become emergent properties of complexity.
When the goo begins to flow into legos. Or Lincoln Logs or Erector sets. The Renton Catalog below tries to follow this line of thinking.
dane2010 posts - fess up
interesting community - the peopleEdit
- from kids to grandmothers
- smart, dedicated, hooked
- good place to be
exploration and discoveryEdit
This is my first video. It is more a statement of intent than anything else. To look at the protein in terms of what it does. Many proteins will likely not be important as static objects but rather as components of a system of automata that do things. Those automata have characteristic ways of doing things. They are dynamic (function of time) building blocks. Legos.
Another video experiment this time mixing fraps video with a sound track. The music is a great tune but its relevance is simply that I had it and it was long enough. Silent videos seemed exceedingly booring.
These segments are experimental to see how it looks and how to piece things together in 30 sec shots.
Hope to try voice over next.
Renton suggest the idea of a catalog to enumerate the ways structures can combine and by ommission the ways they can't. The first level of structure from the goo is the protein chain. The second level is the division of the chain into sheets, helices and loops.
The third level is to see what ways sheets and helices combine with loops as the default or slack state.
As the first entry into third stage, the well known fact that sheets can combine with each other side by side with the same polarity. Sheets have a poolarity - a sided ness. There are typically more hydrophobic chains on one side of the structure than on the other. Call the polarity in and out, with in meaning more hydro-phobic and out meaning less.
This a an edge on combination of sheets with their polarities aligned. We are naming them Renton's 1, 2 etc cause he is the one who suggested a catalog.
Not sure how far we'll get before we run out of steam. In part it depends on how helpful it proves to be.
We love foldit. So. How to make it better? Some thoughts.
When you build a proof, you build from both directions, from premise to conclusion and also from conclusion to premise. Design tools should be available to the solvers. Solvers should be able to build the goal state amd "unfoldit" towards what you can reach from the stretched or tangled state. this is a t-symmetry argument. Changes should be reversible. Forward and backward in time are both physical.
The current level of abstraction is too low. It is like coding with a soldering iron. We suspect that higher level representations will be the best way to thump combinatorial explosion. We suspect also that folding boundaries will weave in and out of the the story. That they co-evolve would seem a certainty.
We think the hydrophobic/hydrophillic distinction is something that should be a build in, like a special field that draws hydrophobic points to places without water automatically and always on in wiggle.
We think certain other operations should be built-in. We spend a hellah amount of time making helices and sheets Making those should be constructions not random searches in rebuild.
Also, we are wary of a seductive path. Trying to make the best by nibbling at the good. When you find a good path, the best choice might be to run far away. Then your may avoid the same well as your neighbors.
More flexibility in the UI. And less real estate. We want to see the molecules and caress them. and smack em if needed.