[PDF] - T ask Analysis Ov erview What is task analysis? T ask PDF Document

SLIDE 1 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (1) T ask Analysis Ov erview What is task analysis? T ask Analysis Metho ds

task

decomp

sition
kno

wledge based analysis

en

tit y-relati

nshi

p tec hniques Sources

f

Information Uses

f

T ask Analysis

SLIDE 2 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (2) What is T ask Analysis? Metho ds

f

analysing p eople's jobs:

what

p eople do

what

things they w

rk

with

what

they m ust kno w Example: in

rder

to clean the house

get

the v acuum cleaner

ut
x

the appropriate attac hmen t

clean

the ro

ms
when

the dust bag gets full, empt y it

put

the v acuum cleaner and to

ls

a w a y Must kno w ab

ut:

v accum cleaners,their attac hmen ts, dust bags, cupb

ards,

ro

ms

etc.

SLIDE 3 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (3) Approac hes to task analysis

T

ask decomp

sition

splitting task in to (ordered) subtasks

Kno

wledge based tec hniques what the user kno ws ab

ut

the task and ho w it is

rganised
En

tit y{relat io n based analysis relationships b et w een

b

jects and actions and the p eople who p erform them General metho d:

bserve

unstructured lists

f

w

rds

and actions

r

ganize using notation

r

diagrams

SLIDE 4 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (4) Dierences from

ther

m

uc h

v

erlap in general

dierences

ha v e exceptions.

SLIDE 5 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (5) T ask Decomp

sition

Aims:

describ

e the actions p eople do

structure

them within task subtask hierarc h y

describ

e

rder
f

subtasks F

cus
n

Hier ar chic al T ask A nalysis (HT A) It uses:

text

and diagrams to sho w hierarc h y

plans

to describ e

rder

SLIDE 6 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (6) T extual HT A description Hierarc h y description ... 0. in

rder

to clean the house 1. get the v acuum cleaner

ut

2. x the appropriate attac hmen t 3. clean the ro

ms

3.1. clean the hall 3.2. clean the living ro

ms

3.3. clean the b edro

ms

4. empt y the dust bag 5. put v acuum cleaner and attac hmen ts a w a y ... and plans Plan 0: do 1 { 2 { 3 { 5 in that

rder.

when the dust bag gets full do 4 Plan 3: do an y

f

3.1, 3.2

r

3.3 in an y

rder

dep ending

n

whic h ro

ms

need cleaning N.B.

nly

the plans denote

rder

SLIDE 7 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (7) Generating the hierarc h y

get

at list

f

tasks

group

tasks in to higher lev el tasks

decomp
se

lo w est lev el tasks further Stopping rules Ho w do w e kno w when to stop? Is \empt y the dust bag" simple enough? Purp

se:

expand

nly

relev an t tasks Error cost: stop when P

C

is small Motor actions: lo w est sensible lev el

SLIDE 8 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (8) Diagrammatic HT A

make a cup of tea boil water empty pot put tea leaves in pot pour in boiling water wait 4 or 5 minutes pour tea fill kettle put kettle

n stove

wait for kettle to boil turn off gas do 1 at the same time, if the pot is full 2 then 3 - 4 after four or five minutes do 5 1.1 - 1.2 - 1.3 when kettle boils 1.4 0. plan 0. 1. 2. 3. 4. 5. 6. plan 1. 1.1. 1.2. 1.3. 1.4.

Line

under b

x

means no further expansion.

Plans

sho wn

n

diagram

r

written elsewhere.

Same

information as: 0. mak e a cup

f

tea 1. b

il

w ater : : :

SLIDE 9 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (9) Rening the description Giv en initi al HT A (textual

r

diagram) Ho w to c hec k/impro v e it? Some heuristics: paired actions e.g., where is `turn

n

gas' restructure e.g., generate task `mak e p

t'

balance e.g., is `p

ur

tea' simpler than making p

t?

generalise e.g., mak e

ne

cup

r

t w

:

: :

r

more

SLIDE 10 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (10) Rened HT A for making tea

make cups

f tea

boil water empty pot make pot wait 4 or 5 minutes pour tea fill kettle put kettle

n stove

turn on and light gas wait for kettle to boil turn off gas warm pot put tea leaves in pot pour in boiling water put milk in cup fill cup with tea do sugar ask guest about sugar add sugar to taste do 1 at the same time, if the pot is full 2 then 3 – 4 after 4/5 minutes do 5 1.1 – 1.2 – 1.3 – 1.4 when kettle boils 1.5 3.1 – 3.2 – 3.3 5.1 5.2

empty cups ? for each guest 5.3

NO YES

5.3.1 — if wanted 5.3.2 0. plan 0. 1. 2. 3. 4. 5. plan 1. plan 3. plan 5. 1.1. 1.2. 1.3. 1.4. 1.5. 3.1. 3.2. 3.3. 5.1. 5.2. 5.3. plan 5.3. 5.3.1. 5.3.2.

SLIDE 11 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (11) T yp es

f

plan xed sequence e.g., 1.1{1.2{1.3

ptional

tasks e.g., if the p

t

is full 2 w aiting for ev en ts e.g., when k ettle b

ils

1.4 cycles

5.1 5.2 empty cups ? for each guest 5.3 NO YES Plan 5.

time-sharing e.g., do 1; at the same time : : : discretionary e.g., do an y

f

3.1, 3.2

r

3.3 in any

r

der mixtures most plans in v

lv

e sev eral

f

the ab

v

e

SLIDE 12 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (12) Kno wledge Based Analyses F

cus
n:

Ob jects | used in task Actions | p erformed T axonomies represen t lev els

f

abstraction Example: motor con trols steering ste ering whe el, indic ators engine/sp eed direct ignition, ac c eler ator, fo

t

br ake gearing clutch, ge ar stick ligh ts external he ad lights, hazar d lights in ternal c

urtesy

light w ash/wip e wip ers fr

nt

wip ers, r e ar wip ers w ashers fr

nt

washers, r e ar washers heating temp er atur e c

ntr
l,

air dir e ction, fan, r e ar scr e en he ater parking hand br ake, do

r

lo ck radio n umerous!

SLIDE 13 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (13) TDH notation TDH { T ask Description Hierarc h y Three t yp es

f

branc h p

in

t in taxonom y: X OR | normal taxonom y

b

ject in

ne

and

nly
ne

branc h AND |

b

ject m ust b e in b

th

represen ts m ultiple classications OR | w eak est case can b e in

ne,

man y

r

none Example: w ash/wip e AND function X OR wip e fr

nt

wip ers, r e ar wip ers w ash fr

nt

washers, r e ar washers p

sition

X OR fron t fr

nt

wip ers, fr

nt

washers rear r e ar wip ers, r e ar washers

SLIDE 14 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (14) Larger TDH example kitc hen item AND / shap e X OR / j dished / j mixing b

w

l, c asser

le,

sauc ep an, / j soup b

w

l, glass / j at / plate, chopping b

ar

d, frying p an / function OR f preparation f mixing b

w

l, plate, chopping b

ar

d f co

king

f frying p an, c asser

le,

sauc ep an f dining X OR j for fo

d

j plate, soup b

w

l, c asser

le

j for drink glass N.B. `/|{' used for branc h t yp es.

SLIDE 15 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (15) More

n

TDH Uniqueness rule: can the diagram distinguish all

b

jects? e.g., plate is: kitc hen item/shap e(at)/functionfpreparation,dining(for fo

d)g/

nothing else ts this description Actions ha v e taxonom y to

:

kitc hen job OR j preparation j b e ating, mixing j co

king

j frying, b

iling,

b aking j dining p

uring,

e ating, drinking

SLIDE 16 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (16) Abstraction and cuts After pro ducing detailed taxonom y `cut' it to yield abstract view. That is, ignore lo w er lev el no des. e.g., cutting ab

v

e shap e and b elo w dining, plate b ecomes: kitc hen item/functionfpreparation,diningg/ This is a term in Kno wledge Represen tation Grammar (KR G) These can b e more complex: `b eating in a mixing b

wl'

b ecomes kitc hen job(preparation) using a kitc hen item/functionfpreparationg/

SLIDE 17 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (17) En tit y{Relati

nshi

p Based T ec hniques Emphasis

n
b

jects, actions and their relationships Similar to

b

ject-orien ted analysis, but : : :

includes

non-computer en tities

emphasises

domain understanding not implemen tation Running example: `V era's V eggies' { a mark et gardening rm Owner/manager: V era Bradsha w Emplo y ees: Sam Gummage and T

n

y P eagreen v arious to

ls

including a tractor `F ergie' t w

elds

and a glasshouse new computer con trolled irrigation system

SLIDE 18 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (18) Ob jects Start with list

f
b

jects and classify them: Concrete

b

jects: simple things: spade, plough, glasshouse Actors: human actors: V era, Sam, T

n

y , the customers what ab

ut

the irrigation con troller? Comp

site
b

jects: sets: the team = f V era, Sam, T

n

y g tuples: tractor ma y b e < F ergie, plough > T

the
b

jects add attributes: Ob ject Pump3 simple | irrigation pump A ttributes: status:

n/o/fault

y capacit y: 100 litres/min ute N.B. need not b e computationally complete

SLIDE 19 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (19) Actions List actions and asso ciate with eac h: agen t | who p erforms the actions patien t | whic h is c hanged b y the action instrumen t | used to p erform action Examples: Sam (agent) plan ted (action) the leeks (p atient) T

n

y dug the eld with the spade (instrument) Note: implicit agen ts | r e ad b ehind the wor ds `the eld w as ploughed' | b y whom? indirect agency | the r e al agent? `V era programmed the con troller to irrigate the eld' messages | a sp ecial sort

f

action `V era told Sam to : : : ' r^

les

| an agen t acts in sev eral r^

les

V era as worker

r

as manager

SLIDE 20 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (20) E/R Example I {

b

jects and actions Ob ject Sam h uman actor Actions: S1: driv e tractor S2: dig the carrots Ob ject V era h uman actor | the proprietor Actions: as w

rk

er V1: plan t marro w seed V2: program irrigation con troller Actions: as manager V3: tell Sam to dig the carrots Ob ject the men comp

site

Comprises: fSam, T

n

yg Ob ject glasshouse simple A ttribute: h umidit y: 0{100% Ob ject Irrigation Con troller non-h uman actor Actions: IC1: turn

n

Pump1 IC2: turn

n

Pump2 IC3: turn

n

Pump3 Ob ject Marro w simple Actions: M1: germinate M2: gro w

SLIDE 21 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (21) Ev en ts Ev en ts are when something happ ens

p

erformance

f

action `Sam dug the carrots'

sp
n

taneous ev en ts `the marro w seed germinated' `the h umidit y drops b elo w 25%'

timed

ev en ts `at midnigh t the con troller : : : '

SLIDE 22 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (22) Relationships

b

ject{ob ject so cial | Sam is sub

rdinate

to V era sp atial | pump 3 is in the glasshouse action{ob ject agent | (listed with

b

ject) p atient and instrument actions and ev en ts temp

r

al and c ausal `Sam digs the carrots b e c ause V era told him' T emp

ral

relations

also

use HT A

r

dialogue notations.

sho

w task sequence (normal HT A)

sho

w

b

ject lifecycl e (see page 241)

SLIDE 23 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (23) E/R example I I { ev en ts and relations Ev en ts Ev1: h umidit y drops b elo w 25% Ev2: midnigh t Relations:

b

ject{ob ject lo cation ( Pump3, glasshouse ) lo cation ( Pump1, P ark er's P atc h ) Relations: action{ob ject patien t ( V3, Sam ) { V era tells Sam to dig patien t ( S2, the carrots ) { Sam digs the c arr

ts

: : : instrumen t ( S2, spade ) { : : : with the spade Relations: action{ev en t b efore ( V1, M1 ) { the marro w m ust b e so wn b efore it can germinate triggers ( Ev1, IC3 ) { when h umidit y drops b elo w 25%, the con troller turns

n

pump 3 causes ( V2, IC1 ) { the con troller turns

n

the pump b e c ause V era programmed it

SLIDE 24 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (24) Sources

f

Information

Do

cumen tation N.B. man uals sa y what is supp

se

d to happ en but, go

d

for k ey w

rds

and prompting in terviews

Observ

ation formal/informal, lab

ratory/eld

(see Chapter 11)

In

terviews the exp ert: manager

r

w

rk

er? (ask b

th!)

SLIDE 25 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (25) Early analysis

Extraction

from transcripts list nouns (obje cts) and v erbs (actions) b ew are tec hnical language and con text `the rain p

ur

e d' `I p

ur

e d the tea'

Sorting

and classifying grouping

r

arranging w

rds
n

cards ranking

b

jects/actions for task relev ance (see Ch. 11) use commercial

utliner

Iterativ e pro cess: data sources ! analysis But costly , so use c heap sources where a v ailable

SLIDE 26 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (26) Uses

f

T ask Analysis I Man uals and Do cumen tation Pro cedural `ho w to do it' man ual

from

HT A description

useful

for extreme no vices

r

when domain to

dicult
assumes

all tasks kno wn Conceptual man ual

from

kno wledge

r

en tit y/relation based analyses

go
d

for

p

en ended tasks Example: tea making man ual from HT A T

mak

e cups

f

tea b

il

w ater | se e p age 2 empt y p

t

mak e p

t

| se e p age 3 w ait 4

r

5 min utes p

ur

tea | se e p age 4 | page 1 | Mak e p

t
f

tea

nc

e water has b

ile

d w arm p

t

put tea lea v es in p

t

p

ur

in b

iling

w ater | page 3 |

SLIDE 27 Human{Com puter In teraction, Pren tice Hall A. Dix, J. Finla y , G. Ab

wd

and R. Beale c

1993

T ask Analysis Chapter 7 (27) Uses

f

T ask Analysis I I Requiremen ts capture and systems design

lifts

fo cus from system to use

suggests

candidates for automation

unco

v ers user's conceptual mo del Detailed in terface design

taxonomies

suggest men u la y

ut
b

ject/action lists suggest in terface

b

jects

task

frequency guides default c hoices

existing

task sequences guide dialogue design NOTE. task analysis is never complete rigid task based design = ) inexible system