Erro r C.W. Johnson, Univ ersit y of Glasgo w, Glasgo w, - - PDF document

▶
erro r c w johnson univ ersit y of glasgo w glasgo w g12
SMART_READER_LITE
LIVE PREVIEW

Erro r C.W. Johnson, Univ ersit y of Glasgo w, Glasgo w, - - PDF document

Jun 22, 2023 202 likes •380 views

Erro r C.W. Johnson, Univ ersit y of Glasgo w, Glasgo w, G12 8QQ. Scotland. johnson@dcs.gla.ac.uk, h ttp://www.dcs.gla .a c.uk/ johnso n Octob er 2001 c CS-1Q: HCI (Lecture 9) C.W. Johnson, 2001 1 Erro rs

slide-1
SLIDE 1 Erro r C.W. Johnson, Univ ersit y
  • f
Glasgo w, Glasgo w, G12 8QQ. Scotland. johnson@dcs.gla.ac.uk, h ttp://www.dcs.gla .a c.uk/ johnso n Octob er 2001 CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001 1
slide-2
SLIDE 2 Erro rs
  • Slips,
lapses, mistak es and violations.
  • Erro
r tolerant design
  • Erro
r detection and recovery .
  • Organisational
facto rs. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-3
SLIDE 3 Slips, Lapses and Mistak es
  • Junio
r trader sells $16m
  • f
German b
  • nd
futures. Ac kno wledgemen t: BBC (No v. 1998)
  • T
rader thought it w as a training screen.
  • Whose
fault is this? T rader
  • r
Company? CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-4
SLIDE 4 Slips, Lapses and Mistak es
  • Erro
r:
  • unwitting
deviation
  • f
actions from intentions.
  • Violation:
  • delib
erate deviation
  • f
actions from regulations.
  • Slip:
  • visible
failure in the execution
  • f
a plan;
  • a
slip
  • f
the tongue is
  • bservable.
  • Lapse:
  • invisible
failure in the execution
  • f
a plan;
  • fo
rgetting someone's name.
  • Mistak
e:
  • a
failure
  • f
intention;
  • trying
to use W
  • rd
to maintain complex accounts. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-5
SLIDE 5 Why do Erro rs Occur?
  • F
atigue and circadian rhythms:
  • mistak
es a re very lik ely last thing
  • n
a F rida y!
  • Stress
(light, heat, noise, domestics):
  • environmental
facto rs and distractions induce mistak es. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-6
SLIDE 6 Why do Erro rs Occur?
  • Alcohol
and drugs:
  • long
  • ce
lunches don't help interaction
  • W
  • rkload
(physical, mental etc):
  • time
p ressures can impair p erfo rmance.
  • Individual
dierences:
  • some
p eople actually do mak e mo re mistak es than
  • thers;
  • sp
ell-checking indicato r
  • co
rrect no w
  • r
at end? CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-7
SLIDE 7 P erception, Cognition, Physiology CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-8
SLIDE 8 Situation Aw a reness
  • Level
1: p erception
  • f
elements in environment.
  • Level
2: comp rehension
  • f
current situation.
  • Level
3: p rojection
  • f
future states. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-9
SLIDE 9 Erro r T
  • lerant
Design
  • Greying
  • ut
menu items:
  • users
can't select inapp rop riate item.
  • T
raining wheels:
  • p
revent users from making an erro r. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-10
SLIDE 10 Erro r T
  • lerant
Design
  • What
happ ens if y
  • u
undo and undo?
  • Conrmation
  • f
irreversible actions:
  • do
y
  • u
really w ant to refo rmat this disk?
  • do
y
  • u
really w ant to end this task? CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-11
SLIDE 11 Erro r T
  • lerant
Design
  • V
alidation
  • f
input p rio r to entry .
  • As
so
  • n
as p
  • ssible
to reduce frustration. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-12
SLIDE 12 Erro r Detection
  • Erro
r messages:
  • recap
  • n
Winograd and Flo res' b reakdo wn.
  • Must
consider b
  • th
frequency and consequences:
  • a
ra re erro r message ma y need mo re explanation;
  • users
must receive indication
  • f
seriousness. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-13
SLIDE 13 Erro r Detection
  • Interface
design:
  • some
users will fail to
  • bserve
erro r messages;
  • some
cannot interp ret the meaning
  • f
the message.
  • F
  • cus
fo r user testing: `no w get
  • ut
  • f
that'. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-14
SLIDE 14 Observing Erro rs
  • Will
user testing rep ro duce erro rs?
  • p
eople strive to please investigato r;
  • p
eople kno w they a re b eing w atched (Ha wtho rne eect)
  • Logging
and tracking
  • f
user's b ehaviour.
  • Problems:
  • will
logs distinguish b et w een slip and mistak e?
  • will
logs help to detect lapses at all?
  • ethical
issues and legal issues... CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-15
SLIDE 15 Organisational F acto rs
  • James
Reason:
  • Human
Erro r (1990);
  • Managing
the Risks
  • f
Organizational Accidents (1997).
  • Who
causes the erro r:
  • the
user, the designer, the manager? CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-16
SLIDE 16 Summa ry
  • Slips,
lapses, mistak es and violations.
  • Erro
r tolerant design
  • Erro
r detection and recovery
  • Organisational
facto rs. CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001
slide-17
SLIDE 17 F urther reading
  • Again
Shneiderman skims this issue.
  • Shneiderman
  • n:
  • erro
r messages
  • 373-379.
  • T
ry to read Reason's Human Erro r?
  • Y
es, seriously ... CS-1Q: HCI (Lecture 9) c
  • C.W.
Johnson, 2001