next up previous
Next: 4.4 An Example of Up: 4 Experimental Setting Previous: 4.2 The Tasks

4.3 Collected Data

In the collaborative sessions, we collect logs of everything typed to the MOO, as well as quicktime movies of the Whiteboard window. Excerpts of the MOO-logs for two of these sessions are shown in Tables 2 and 4. The basic information recorded includes the time a command was typed (shown in [hours:]minutes:seconds from start of the experiment), which subject performed the command (S for Sherlock, H for Hercule), which MOO-room the subject's character was in at the time, and what the subject typed. Further derived columns were also included for typing that was understood as valid commands by the MOO. This includes a column for the action name, one for the other arguments of a command (e.g., what is being looked at, where a player is moving from and too, who is being spoken to, etc.), and, in the case of direct communication, what has been said. This kind of logging allows us to reconstruct aspects of the state of both individuals throughout the collaboration. Table 4 includes some examples of typing that did not result in successful commands. Most of these, such as at 1:12:20, are due to simple typos, but some are due to misconceptions about the domain (at 1:13:55, S tries to read about the gun, but the detective's notebook only presents information gathered according to the witnesses who provided it).

In addition, from the whiteboard movies, we derive logs of all whiteboard creations and manipulations, time synchronized with the MOO logs. This allows us to easily view the relationship between grounding-related actions in the MOO and whiteboard, showing that the pair forms one collaborative system, rather than acting as individual media. This point is elaborated upon in [Dillenbourg et al. 1996].

 
time place who action args said text typed command
10:23corrH *say* --I see you are fastsay I see you are fast
11:23blueS *who* who
11:56blueS *page* Hwhat are you doingpage hercule what are you doing
12:18corrH *say* --how do you know about this f.. chair?say how do you know about this f.. chair?
12:47corrH *moveto* corr,bluewalk to blue room
12:58blueS *who* who
13:42blueS *say* Hdo you agree that red and blue room do not communicate?say do you agree that red and blue room do not communicate?
14:05blueS *who* who
14:37blueS *say* Hwhy dont you never answer my messages?say why dont you never answer my messages?
15:13blueS *look* bluelook
15:24greenS *moveto* blue,greens
15:35blueH *say* --because I am trying to draw the obvious exits of blue room, Isay because I am trying to draw the obvious exits of blue room, I
17:27blueH *page* Sbecause I am trying to draw the obvious exits of blue room, Ipage sherlock because I am trying to draw the obvious exits of blue room, I
Table 2: Excerpt of MOO Log 5T1

 


next up previous
Next: 4.4 An Example of Up: 4 Experimental Setting Previous: 4.2 The Tasks

David Traum
Thu May 23 21:12:30 MET DST 1996