Information Technology Reference
In-Depth Information
Table 11.6
Example application of the KLM to a simple editing task
Method
1. Delete 3rd clause. H[mouse] K[mouse down] P K[mouse up] M K[D]
2. Insert it in front of 1st clause. P M K[l] K[ESC]
3. Replace '': o'' with ''O''. P M 2 K[SHIFT R] H[keyboard] 2 K[O ESC]
4. Replace ''T'' by '': t''. H[mouse] K[mouse down] M K[R] H[keyboard] 4 K[: SPACE t ESC]
5. Delete 3rd clause. H[mouse] P K[mouse] M K[D]
6. Insert it in front of 2nd clause. K M K[l] K[ESC]
7. Find next task. M K[F]
Time Predictions
Texecute = [23 t K +4t P +5t H ]+7t M
= 22 (0.15) + 4 (1.1) + 5 (0.4)] + 7(1.35)
= 19.15 s
Note that in this example the same average time is used for each keystroke
interfaces should choose keys that are fast (and presumably easy to type). This
suggestion includes choosing words that use both hands for touch typists and
avoiding punctuation characters (which are slower to press).
(2) The gross predictions of mouse moves and drags can be made more accurate
by using Fitts' law to make predictions based on the target size and distance.
When the initial location of the mouse is not known, it can be approximated.
(3) The system response time can be computed using tools to analyze key press
times. The response times are still important in devices such as ATMs, and in
systems which may be required to present the same data simultaneously on
different continents, within a multinational company, for example.
11.6.2 Example Application of the KLM
Table 11.6 shows an example KLM analysis worked out for a simple text editing
task of cutting a phrase (Step 1), inserting it (Step 2), changing the case of two
letters (Steps 3 and 4), and taking up the next task (Step 5). The analysis suggests
that the time to perform this task is not so long (24 s). Thus, if this edit takes
longer, it is not the interface that is restraining the user's progress, it is the
cognition to come up with the letters or to think about other changes. It is the
writing, essentially, to come up with the changes that takes so long!
11.6.3 Summary
Although closely related to GOMS, note how keystroke-level analysis is closer to
time-and-motion (chronological) analysis than goal/subgoal analysis. It assumes a
simple world: concentration on one task at a time, a fully specified interface, no
Search WWH ::




Custom Search