next up previous
Next: Parsing Up: User Testing Previous: Working Styles

Time Spent Entering and Correcting Formulae


  
Figure 6.1: Times for people entering Formula 1.
\includegraphics[width=0.95\linewidth]{figures/F1.eps}


  
Figure 6.2: Times for people entering Formula 2.
\includegraphics[width=0.95\linewidth]{figures/F2.eps}


  
Figure 6.3: Times for people entering Formula 3.
\includegraphics[width=0.95\linewidth]{figures/F3.eps}


  
Figure 6.4: Times for people entering Formula 4.
\includegraphics[width=0.95\linewidth]{figures/F4.eps}


  
Figure 6.5: Times for people entering Formula 5.
\includegraphics[width=0.95\linewidth]{figures/F5.eps}


  
Figure 6.6: Times for people entering all the formulae.
\includegraphics[width=0.95\linewidth]{figures/Ftotals.eps}


  
Figure 6.7: Time spent by users entering and correcting formulae.
\includegraphics[width=0.95\linewidth]{figures/pie_graph.eps}

Figures 6.1, 6.2, 6.3, 6.4, 6.5 show the drawing and correction times for each of the five test formulae. Figure 6.6 shows the total times taken by each participant across all the formulae. There is no total time for participant seven, as they did not enter formula two or formula three. Figure 6.7 is a pie graph showing the proportion of time spent by the average user in various stages entering and correcting formulae.

Time spent by users fixing their own mistakes, for example writing the wrong symbol and then having to delete it and rewrite it correctly, is included in this drawing time. The drawing time also includes waiting for the interface to time out and recognise the outstanding strokes that the user has drawn.

Over the period of entry for the five formulae, the total entry times for the participants in the user testing ranged from 117 to 236 seconds, with an average of 194 seconds. An expert user who is familiar with the system is able to enter the same formulae in around 68 seconds.

The difference between the ideal and observed time is due to the fact that most participants did not write quickly with the pen and tablet, due to unfamiliarity and tentativeness. The users also occasionally chose to delete and rewrite characters instead of using the modify stroke groups or modify character options for correction when errors occurred.

Due to the fact that the character recogniser had not been trained for each participant and the participants' discomfort with the pen and tablet, the number of misgroupings that occurred was quite high. As a result, a sizable proportion of the user's time was spent making grouping corrections.

Of the total time the users spent entering and correcting each formula, 13% of their time was making grouping corrections. This also includes time that the users spent checking over the formula and looking for grouping errors. Actual grouping corrections took about a second each.

As with the automatic stroke grouping, the low performance of the character recogniser increased the number of symbol misrecognitions. Of the total time spent entering and correcting formulae, 22% of the users' time was spent looking for and correcting character recognition errors. If the correct alternative was on the pop up menu, the correction took less than a second. If the user had to type the correct character in from the keyboard, it took five to ten seconds.


next up previous
Next: Parsing Up: User Testing Previous: Working Styles
Steve Smithies
1999-11-13