Document

GOMS
Timing for WIMP interfaces
When (fine-grained) speed matters
•
•
•
•
•
•
K - keypress
P - point with mouse
C - click with mouse
H - home hands on new device
M - mentally prepare
R(t) - system response time
2
K - keypress
P - point with mouse
C - click with mouse
H - home hands on new device
M - mentally prepare
R(t) - system response time
3
•
•
•
•
•
•
K – keypress
.08 - 1.20
P - point with mouse
.8 - 1.5 (Fitt's Law)
C - click with mouse
.2
H - home hands on new device
.4
M - mentally prepare
1.35
R(t) - system response time
?
4
NOTES:
• M before K/C or P except
• MPMC ... becomes just …
• MPC because C “anticipated”
– e.g. move mouse to target and click
• MKMKMK ... MKKK for cognitive unit
– e.g. type “cat”
5
Method 1 – highlighting “the cat”
• Assumptions:
•
•
•
•
•
•
•
•
•
hands were on keyboard
•
R=0
•
Experienced typist
H - 0.40 - Reach for mouse
M – 1.35 – mentally prepare
P - 1.10 - Point to the left of "the"
C - 0.20 – Click mouse down and hold
M – 1.35 – mentally prepare
P - 1.10 - Point to right of "cat"
C - 0.20 - Release mouse
Total = 5.7
6
About K, P, R
 You need to consider each of these




What assumptions can you make?
Record them
Then select the value(s) that are reasonable
You may need to creates branches which assess
different values
Method 1 cont – bolden keyboard shortcut
•
•
•
•
M – 1.35 – mentally prepare
K - 0.60 - Press and hold "Control"
K - 0.60 - Press "B"
K - 0.60 - Release "Control"
• Total = 3.15
8
Method 2 - use menu
Assumptions: as before
 M – 1.35 – mentally prepare
• P - 1.10 - Point to "Format" menu
• C – 0.20 - Click and hold
• M – 1.35 – mentally prepare
• P - 1.20 - Point to "Bold" menu item
• C – 0.20 - Release mouse
• Total = 5.40
9
Conclusion for this case
• Assumptions: Hand position, R, K, P
• Common part is 5.7 (sweeping out “the cat”)
• Rest:
– Keyboard shortcut takes 3.15 seconds
– Mouse menu method takes 5.4 seconds
10
Other cases to consider?
• Consider each assumption:
• Hand position, R, K, P
• Other representative tasks
11
Summary of approach
•
•
•
•
Focus on speed
Known sequence of operations
Can predict performance for experienced users
Walkthrough steps, calculate time for each step,
sum
• Can sometimes predict choices of method
12
Summary of uses
• Relatively inexpensive
• Can be used to compare “methods”
• Challenging to apply for conventional
interfaces .... pervasive?
• Expert users only
• Would you expect software that assist in
this?
13
Postconditions (exam questions)
• Critique the relevance of GOMS in a particular
context: user expertise, timing in project, usability
targets, available expertise of evaluators
• Compare and contrast with other UE methods
• Application of keystroke analysis for a given task
• Application timing analysis, with particular care
about assumptions
• Critique a keystroke or timing analysis: strengths
and limitations
14