I wonder if a similar game for labeling video would work well. As the video plays, players would type words corresponding to whatever they see at that moment. Word matches between players would only occur within a small time interval (e.g., within 5 seconds say). Players can type the same word multiple times and matches with the same word may occur multiple times. Of course, as with the ESP Game, you would need various mechanisms to detect cheating.
Del.icio.us users bookmarking helpful/timely URLs (as evidenced by others subsequently bookmarking those URLs) have greater influence on the search rankings.
The Speculative Search Game is a bit like BlogShares.
However, the Speculative Search Game allows you to make predictions about any web page -- not just blogs. Moreover, The Speculative Search Game has a much simpler model than BlogShares and most other artificial game markets -- and perhaps this would encourage more people to play.
The Speculative Search Engine is not available yet, but will be if/when the game attracts many players and generates some interesting data.
I believe that alpha/beta applications can be "stabilized" by users. See the Stabilizer project for buggy GUI applications: http://stabilizer.sf.net. Perhaps something similar can be done for buggy web sites.
You may wish to look into programming by
example/demonstration, which does as you
suggest and more: it generalizes user
examples into complete programs/macros:
However, that's not the purpose of DRT. DRT
is a tool to help you understand other people's
code by considering concrete scenarios/actions.
As for automatic instrumentation without code
changes, that would be difficult right now
even if we focus just on qt/gtk.
DRT currently patches applications to ensure
they do not perform too much polling (which
confuses DRT action detection) and also that they
have a non-blinking cursor of a unique
color (to allow cursor tracking).
Amir
Slashdot Mirror
I wonder if a similar game for labeling video would work well. As the video plays, players would type words corresponding to whatever they see at that moment. Word matches between players would only occur within a small time interval (e.g., within 5 seconds say). Players can type the same word multiple times and matches with the same word may occur multiple times. Of course, as with the ESP Game, you would need various mechanisms to detect cheating.
BTW, you may be interested in this search engine for del.icio.us:
/
http://collabrank.web.cse.unsw.edu.au/del.icio.us
Del.icio.us users bookmarking helpful/timely URLs (as evidenced by others subsequently bookmarking those URLs) have greater influence on the search rankings.
This may be of interest: http://collabrank.org
However, the Speculative Search Game allows you to make predictions about any web page -- not just blogs. Moreover, The Speculative Search Game has a much simpler model than BlogShares and most other artificial game markets -- and perhaps this would encourage more people to play.
The Speculative Search Engine is not available yet, but will be if/when the game attracts many players and generates some interesting data.
I believe that alpha/beta applications can be "stabilized" by users. See the Stabilizer project for buggy GUI applications: http://stabilizer.sf.net. Perhaps something similar can be done for buggy web sites.
You may wish to look into programming by example/demonstration, which does as you suggest and more: it generalizes user examples into complete programs/macros:
http://lieber.www.media.mit.edu/people/lieber/PBE/
However, that's not the purpose of DRT. DRT is a tool to help you understand other people's code by considering concrete scenarios/actions.
As for automatic instrumentation without code changes, that would be difficult right now even if we focus just on qt/gtk.
DRT currently patches applications to ensure they do not perform too much polling (which confuses DRT action detection) and also that they have a non-blinking cursor of a unique color (to allow cursor tracking). Amir
Hi,
The tool is designed to help you understand
other people's code. How would replaying
actions help in that regard?
Could you be more specific in your second question?
The DRT manual already describes how to instrument
most X applications to work under DRT.
Amir
Please see the following paper for motivation and examples:
http://citeseer.nj.nec.com/536115.html
(See the links in the top right to download/view.)
Amir