Extracting Quantitative Information with Lua
|Opened on:||Friday, June 8, 2012 - 17:39|
|Last modified:||Thursday, March 14, 2013 - 20:18|
I have written scripts to force sidechain angles in disulfide bridges, measure dihedral angles, twist sheets, make perfect (pi!) helices. I've done this because the tweak/rebuild tool is too violent, or I can't intuit WHY a segment is scoring poorly. A lot of other discussion has been going on among the scriptors about doing similar things themselves. Clever scripting and a lot of trig can get you pretty far.
Quantitative information has been intentionally hidden from the players. Being able to make measurements allows the creation of scripts that converge to a particular conformation, be it a superior evolved foldit solution or a pdb extract.
What information is the foldit team willing to provide? What limits should we impose upon ourselves in using that information? There are lots of stakeholders in this:
The science team wants novel solutions, not copies of other servers.
The players want a fair competition with integrity of the solo / evolved ranking system
The games team wants to compare humans to computers and study crowdsourcing and gamification.
The dev team has to implement it.
I think some guidelines need to be laid out so that everybody gets what they want out of Foldit. For example, from a scripting folder's point of view, http://fold.it/portal/node/991316 really should be implemented, we're doing crazy work-arounds to emulate it, yet it is ultimately a very powerful function in regards to the information it provides.