Question?/Suggestion: Precision control issue

Case number:671071-752776
Topic:Game: Other
Opened by:jas0501
Status:Closed
Type:Suggestion
Opened on:Monday, December 22, 2008 - 03:29
Last modified:Wednesday, June 1, 2011 - 20:29

Issue: Often after setting bands and starting a wiggle I can't stop it quick enough.

Possible solution: It would be nice to be able to step through the wiggle result and then be able to undo the steps. Maybe a "stepped wiggle" with + and - advancing and treating the wiggle.

(Mon, 12/22/2008 - 03:29  |  5 comments)


axcho's picture
User offline. Last seen 8 years 9 weeks ago. Offline
Joined: 12/18/2007
Groups: None

Thanks for the suggestion. How often do you find yourself needing to step back through a wiggle, and how significantly does it affect your score to not be able to do so?

Joined: 12/08/2008
Groups: None

axcho asks:
How often do you find yourself needing to step back through a wiggle, and how significantly does it affect your score to not be able to do so?

----------
Good questions, and I'm not sure.

Consider that I am a newbie and what I may be doing may be impractical or inefficient first off.

My point of asking is based on the following hypothisis:
The position may be on a local peak. In order to get to the next summit's foothills, I need to go down to the valley and start up the next slope. More ticks ob the wiggle may take me past multiple valleys.

The wiggle with bands gets me going however the first couple ticks may not take me far enough and once I deactivate the bands, and wiggle, I climb right back to the original position. However, if I wiggle long enough with the bands, but not too long, once I deactivate the bands, I will wiggle to a new and different peak, and sometime significantly higher peak.

In order to answer your question some experimenting with the added precision would help determine the answer.

--------------------

The above approach is similar to simmulated annealing, but done manually I guess. The band charge, distance, and number of wiggle ticks would be the "temperature" control.

Joined: 09/14/2008
Groups: Void Crushers

you know that, if the wiggle got too far, a simple 'ctrl-b' will set the puzzle to the best result you have?
even if the wiggle get yiou a lower score, ctrl b will 'reset' the puzzle to its best result.

Joined: 10/19/2008
Groups: Oma Gawd

FJ, Jas may be leading towards the same suggestion made in the past, ie, the "New Best" feature that would allow a user to reset a secondary "best score" buffer" to the current user score, and then track the best score following that reset point in exactly the same manner as the global best score ie, using peak detection, and allow the user to reset to that point via a secondary "restore best" command key combo.

He may also be describing another feature previously discussed and submitted (and resubmitted clearly in a post I just made) suggestion for user-specified undo buffer lengths.

but I'm not exactly sure. However, if either are accurate assumptions, then I agree, imo, these are tools that all users could benefit from having.

In My Opinion>>>As for actually stepping thru the individual iterations of the wiggle function itself, I would personally be concerned that such a change may bog down other more important development and/or lead to problems without a very thorough thinking and testing period, requiring redevelopment of the buffering from file-based to memory-based, need for concomitant GUI changes, etc., and so may be better back-shelved until such a time as it is more practical and the outstanding list of current bugs and user suggestions are handled. Further, in my understanding of the design paradigm for Fold.it, such minutia should not be necessary for a human-powered "quantum-thinking" solution, although there is no question that it would be of clear benefit were the application redefined for those proponents that seek to change Fold.it into a variant on Rosetta or other hands-off BOINC applications.

Joined: 06/17/2010
Status: Open » Closed

Band strength can be adjusted. Done.

Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons