Vous êtes sur la page 1sur 12

AUTOMATION TESTING CONCEPTS

1 Introduction.......................................................................................................................3
2 Why Test Automtion.......................................................................................................3
3 Automtion !i"e Cyc#e.....................................................................................................3
3.1 Automtion $esi%i#ity An#ysis................................................................................3
3.2 P#nnin& nd 'esi&n..................................................................................................3
3.3 Test En(ironment Setu).............................................................................................3
3.* Automtion Scri)t Genertion...................................................................................3
3.+ Test E,ecution............................................................................................................3
3.- 'e"ect An#ysis . $i,in&..........................................................................................3
* Cost In(o#(ed in utomtion.............................................................................................3
+ Myths . /e#ities.............................................................................................................3
- $esi%#e Strte&ies........................................................................................................3
-.1 $unction# 'ecom)osition Method...........................................................................3
-.2 Test P#n 'ri(en Method...........................................................................................3
0 /esource Concerns...........................................................................................................3
1 Thin&s to remem%er..........................................................................................................3
2 /e"erences.........................................................................................................................3

1
1 Introduction
A"ter certin )oint o" time3 the re)etiti(e mnu# testin& o" test scri)ts %ecomes (ery
monotonous nd mundne. A#so3 "or certin ty)es o" testin& #i4e re&ression testin&3 it
#5ys he#)s i" some%ody or somethin& cou#d &o on testin& the sme thin& &in nd
&in. With the current trend o" so"t5re re#ese cyc#es %ecomin& shorter nd the #ist o"
tests in the re&ression test suite %ecomin& #r&er3 it %ecomes im)ro%%#e to do com)#ete
mnu# re&ression testin& 5ith e(ery re#ese. Incom)#ete nd ine""icient testin& my
cuse ne5 %u&s 6de"ects introduced 5hi#e "i,in& %u&7 &oin& undetected %e"ore the
re#ese. 8y utomtin& the )rocess 6re&ression testin& in this cse73 the time needed to
test run the "u## suite reduces3 there%y )ro(idin& the tester more time to "ocus on mnu#
testin& o" ne5 code nd on res o" more ris4.
The conce)t o" utomtion "inds its use in ## such cses. In short3 it9s min#y %out
Usin& so"t5re too#s to e,ecute tests 5ithout mnu# inter(ention
Use o" other so"t5re too#s to contro# the e,ecution o" tests3 the com)rison o"
ctu# 5ith the e,)ected outcome nd other test contro# . test re)ortin& "unctions.
Automtion Testin& s such re"ers to utomtin& the mnu# testin& )rocess. Automtion
cn %e ))#ied to
: We% ))#ictions
: C#ient :ser(er en(ironments.
The )rocess %sic##y consists o" test cses tht contin e,)ected resu#ts %sed on
%usiness re;uirements nd #so n e,c#usi(e test en(ironment h(in& test dt%se in
5hich the test cses cn %e re)eted 5hene(er the ))#iction under&oes chn&e.
2 Why Test Automation
The "o##o5in& )oints e#ucidte the resons %ehind the need "or utomtion 6se#ecti(e nd
not e,husti(e7<
/unnin& the e,istin& tests on ne5 (ersion6s7 o" )ro&rm=))#iction
More num%er o" tests cn %e run in #esser )eriod o" time
8u&s tht cn %e unco(ered on#y 5ith #on& runs
8etter use o" resources
Testers cn do %etter >o%s3 mchines cn %e run ## the time

2
Consistency3 /e#i%i#ity nd /e)et%i#ity o" tests
The sme tests cn %e re)eted e,ct#y in the sme mnner3 e(ery time 5hen the
utomted scri)ts re e,ecuted thus e#imintin& mnu# testin& errors
3 Automation Life Cycle
St&es in(o#(ed in utomtion )rocess re 6in order7
1. Automtion $esi%i#ity An#ysis
2. P#nnin& nd 'esi&n
3. Test En(ironment Setu)
*. Automtion Scri)t Genertion
+. Test E,ecution
-. 'e"ect An#ysis . $i,in&

Ech st&e hs s)eci"ic Entry Criteri3 Acti(ities3 Mesurements nd E,it Criteri de"ined
"or them. The deti#s re s "o##o5s<
3.1 Automation Feasibility Analysis
O(er(ie5
/e;uirements . e,)ecttions h(e to %e de"ined c#er#y . e,)#icit#y %e"ore strtin& o""
ny )ro>ect nd so is the cse in(o#(in& utomtin& )ro>ect. Thin&s #i4e
?o5 nd 5ht is to %e utomted
Which ## modu#es cn %e utomted
shou#d %e decided in the %e&innin& itse#" to )re(ent ny con"usion "ter the utomtion
)rocess hs %een "#&&ed o"". This st&e hence re;uires &ret de# o" ttention.

Entry:criteri
C#er#y de"ined /e;uirements
Sound understndin& o" c#ient e,)ecttions
Acti(ities
/e;uirement n#ysis "or thorou&h understndin&
Identi"y s)eci"ic re;uirement tests tht cn %e utomted
Identi"y test cses tht cnnot %e utomted due to too# #imittions nd = or
com)#e,ity o" the "unction#ity
Understndin& the System nd the en(ironment in 5hich the ))#iction hs
to %e inst##ed
An#yse ))#ic%i#ity o" suit%#e too# to crry out the utomtion
/ou&h estimte o" the si@e3 e""ort nd cost in(o#(ed in this utomtion
An#y@e i" utomtion is ctu##y )ossi%#e nd %ene"it in the #on& run
Pre)re "esi%i#ity n#ysis re)ort

3
A#idtion
/e(ie5 o" $esi%i#ity An#ysis /e)ort

E,it:criteri
A))ro(ed $esi%i#ity An#ysis re)ort
3.2 Planning and esign
O(er(ie5
Once the re;uirements . e,)ecttions h(e %een hi&h#i&hted c#er#y nd the "esi%i#ity
studies yie#d )ositi(e resu#t3 methodic# )#nnin& hs to %e done. This 5ou#d inc#ude
de"inin& the ))roch3 decide on 5hich "rme5or4 is to %e used nd su%se;uent#y3
estimtin& the time nd "innce needed to com)#ete the sme.

Entry:criteri
Automtion $esi%i#ity n#ysis re)ort
Acti(ities
Identi"y ))#ic%#e ))roches
An#yse the suit%i#ity o" ech ))roch nd se#ect the %est mon& them
Identi"y the "rme5or4s tht cn %e used
An#y@e e(ery "rme5or4 "or their )ros nd cons
$in#i@in& the "rme5or4 to %e used
$orm test )#n %sed on the %o(e.
An#yse mnu# test scri)ts "rom n utomtion stnd)oint.
'e"ine utomtion scri)tin& stndrds 6Common #n&u&e "unction#ity7
Identi"y common "unction#ity cross test cses nd desi&n com)i#ed modu#es
Cte&ori@e test scri)ts %sed on "unction#ity
Identi"y "unction#ity 5hich need to %e )rmeterised
Pre)re the deti#ed cost3 si@e nd e""ort estimte
A#idtion
/e(ie5 o" Test P#n %y the c#ient

E,it:criteri
A))ro(ed utomtion $rme5or4
A))ro(ed utomtion test )#n
3.3 Test !n"ironment #etu$
O(er(ie5
En(ironment set:u) is one o" the critic# s)ects o" testin& )rocess. Test tem my not %e
in(o#(ed in this cti(ity i" the customer=de(e#o)ment tem )ro(ides the test en(ironment
in 5hich cse the test tem is re;uired to do rediness chec4 o" the &i(en en(ironment.

*
Test en(ironment decides the so"t5re nd hrd5re conditions under 5hich 5or4
)roduct is tested.
Entry:criteri
System 'esi&n nd rchitecture documents re (i#%#e
En(ironment set:u) )#n is (i#%#e
Acti(ities
Understnd the re;uired rchitecture3 en(ironment set:u)
Pre)re en(ironment setu) chec4#ist
Pre)re hrd5re nd so"t5re re;uirements
$in#i@e connecti(ity re;uirements
Setu) Test En(ironment
A#idtion
E""ort "or setu)
'e"ects in setu)

E,it:criteri
En(ironment setu) is 5or4in& s )er the )#n nd chec4#ist
3.% Automation #cri$t &eneration
O(er(ie5
The )ro>ect tem 5i## &o throu&h the test )#n . o%tins c#ri"ictions3 i" needed. It 5i##
then decide u)on the ty)e o" utomtion to %e used3 it cou#d %e c)ture=)#y%c43
scri)tin& etc. The )ro>ect tem 5i## shre the 4no5#ed&e nd e,)erience throu&h
)resenttions nd=or discussions t ))ro)rite inter(#s durin& the )ro&ress o" the
)ro>ect.
Entry:criteri
Test )#n is (i#%#e
Test too# (i#%#e
Mnu# test cse i" ny

Acti(ities
Crete test scenrios
Identi"y reus%#e test scenrios
Grou) the common test cses to&ether
$in#i@e Test Scri)ts
A#idtion
/e(ie5 o" scri)ts

E,it:criteri
Scri)ts re tested nd re(ie5ed
'e"ects re "i,ed

+
3.' Test !(ecution
O(er(ie5
The )ro>ect tem mem%er6s7 5i## crry out testin& %sed on the test re;uirements.
Entry:criteri
8se#ined Test )#n is (i#%#e
Test en(ironment is redy
Test dt set u) is done
Test Scri)ts re (i#%#e

Acti(ities
Per"orm the testin& s )er the test )#n
Per"orm the %tch test 5here(er necessry
U)dte the sttus in the sttus mtri,
!o& the de"ects

A#idtion
Testin&

E,it:criteri
A## test cses re e,ecuted.
'e"ects #o&&ed in the de"ect trc4in& too#
3.) efect Analysis * Fi(ing
O(er(ie5
The )ro>ect tem mem%ers 5i## n#y@e the de"ects #o&&ed nd 5i## "i, the de"ects. The
)ro>ect tem mem%ers 5i## #so )re)re document on the de"ects "ound3 their )riority
nd #so the so#ution "or the sme.
Entry:criteri
Testin& hs %een done
Test records re (i#%#e
'e"ect #o& is (i#%#e

Acti(ities
An#y@e the de"ects "or se(erity nd )riority
$i, the de"ects
'ocument the de"ect #on& 5ith the so#ution

A#idtion
Testin&

E,it:criteri
'e"ects re n#ysed nd "i,ed
A))#iction is redy to %e retested "or %u& "i,es

-
% Cost In"ol"ed in automation
$i,ed Cost
Automtion "esi%i#ity n#ysis cost
Too# se#ection nd Ac;uisition cost
?irin& s4i##ed resources O/ trinin& e,istin& tem mem%ers
Time in #ernin& the ))#iction=%usiness )rocesses
Pi#ot )ro>ect identi"iction e""ort nd Proo" o" Conce)t
$irst time utomtion o" the identi"ied )rts o" ))#iction=)roduct
Test suite 'ocumenttion e""ort
Ari%#e cost
Test scri)t nd documenttion mintennce cost
Automted Test in"rstructure mintennce cost
E,ecution cost
' +yths * ,ealities
Automted testin& is >ust n dd:on to the testin& )rocess to m4e the testers9 #i"e esier.
Un#i4e )re(#ent myths3 it doesn9t necessitte rm):do5n o" resources in the testin&
de)rtment. A#so3 utomtion is not s ine,)ensi(e s 5e thin4. As mtter o" "ct3 it
my t4e much #on&er to de(e#o)3 (eri"y3 nd document n utomted test cse thn to
crete nd e,ecute mnu# test cseB This is e(en truer i" the record=)#y%c4 "eture is
se#ected s the )rimry testin& techni;ue. In "ct3 record=)#y%c4 is the #est cost:
e""ecti(e method o" utomtin& test cses.
?(in& sid tht3 it doesn9t men tht utomtion testin& cn9t %e mde cost e""ecti(e t
##. 'e)endin& on the ty)e o" testin& re;uirements o" the com)ny3 the too# cou#d %e
decided u)on. I" there %e cse 5herein the sme "unction#ity is chie(ed %y t5o
di""erent too#s 5herein one is si&ni"icnt#y cost#ier thn the other3 the #ess e,)ensi(e one
cou#d %e esi#y )re"erred o(er the other. A&in3 the too# chosen shou#d %e ser(in& the
interests o" the com)ny 5e##. Other5ise3 e(en sm## mount o" money s)ent on C#ess
e,)ensi(e too#D 5on9t do ny &ood. Sim)#y )ut3 it9s the re;uirements tht come "irst
5hi#e choosin& too# nd then ## other monetry nd other considertions.
A)rt "rom the monetry s)ect3 there re other issues re#tin& to time mn&ement s
5e##. Tryin& to utomte (ery com)#e, test cses 5ou#dn9t ser(e much )ur)ose s they
5ou#d consume #ot o" time nd t the end o" the dy3 no one cn %e sure i" the
utomtion 5ou#d ctu##y %e success"u# time nd &in. So rther thn %re4in& heds
o(er some (ery com)#e, test cses3 the tester 5ou#d do 5e## to utomte si&ni"icnt
num%er o" re#ti(e#y esier test cses in the men time. This method is the most cost#y 6in
terms o" time consum)tion7 mon& ## methods. The record=)#y%c4 "eture o" test too#
is use"u# "or determinin& ho5 the too# intercts 5ith the ))#iction under test3 nd cn
&i(e ides re&rdin& de(e#o)in& test scri)ts3 %ut its use"u#ness ends there. It cn9t hnd#e
ny une,)ected )o):u) tht mi&ht sudden#y ))er or ny such une,)ected %eh(ior "or
tht mtter tht 5s not )rt o" the recordin&.

0
) Feasible #trategies
Other thn the record=)#y%c4 strte&y 65hich hs more disd(nt&es thn d(nt&es73
there re "e5 more methodo#o&ies 5hich re (ery e""ecti(e "or utomtin& "unction# or
system testin&. A "e5 o" such strte&ies re<
).1 Functional ecom$osition +ethod
Essenti##y3 it in(o#(es reducin& ## test cses to their most "undment# ts4s3 nd 5rite
user:de"ined "unctions3 %usiness "unction scri)ts3 nd su%:routine or uti#ity scri)ts 5hich
)er"orm the ts4s inde)endent#y o" one nother. In &ener#3 these "undment# res
inc#ude n(i&tion3 s)eci"ic %usiness "unction3 nd dt (eri"iction . return n(i&tion.
$or chie(in& this3 it is mndtory to se)rte dt . "unction. This ##o5s n utomted
test scri)t to %e 5ritten "or %usiness "unction3 usin& dt:"i#es to )ro(ide %oth the in)ut
nd the e,)ected:resu#ts (eri"iction. The hierrchic# rchitecture em)#oyed3 hs
structured or modu#r desi&n. The en&ine o" the test is the dri(er scri)t 5hich is t the
hi&hest #e(e#. This dri(er scri)t hs num%er o" c##s to one or more test scri)ts. These
test scri)ts contin the ctu# #o&ic3 c##in& the %usiness "unction scri)ts necessry to
)er"orm the ))#iction testin&. $unctions . uti#ity scri)ts re c##ed s needed %y
'ri(ers3 Min3 nd 8usiness $unction scri)ts.
Whi#e the dri(er scri)ts do the initi#i@tion nd c## the scri)ts in the correct order3 the
test cse scri)ts )er"orm the ))#iction test cse #o&ic usin& %usiness "unction scri)ts
5hich in turn )er"orm s)eci"ic %usiness "unctions 5ithin the ))#iction. The su%routine
scri)ts )er"orm s)eci"ic ts4s re;uired %y t5o or more %usiness scri)ts 5hi#e user:de"ined
"unctions re &ener#3 ))#iction:s)eci"ic nd screen:ccess "unctions. The %usiness
"unction nd su%routine scri)ts in(o4e user de"ined "unctions to )er"orm n(i&tion. The
Test Cse scri)t 5ou#d c## these t5o scri)ts3 nd the 'ri(er Scri)t 5ou#d c## this Test
Cse scri)t some num%er o" times re;uired to )er"orm ## the re;uired Test Cses o" this
4ind. In ech cse3 the on#y thin& tht chn&es re the dt contined in the "i#es tht re
red nd )rocessed %y the %usiness "unction . su%routine scri)ts.
This method3 ho5e(er3 re;uires on#y tht 5e dd the dt:"i#es re;uired "or ech test3 nd
these cn esi#y %e u)dted=mintined usin& Note)d or some such te,t:editor. U)dtin&
these "i#es does not re;uire ny 4no5#ed&e o" the utomted too#3 scri)tin&3
)ro&rmmin&3 etc. menin& tht the non:technic# testers cn )er"orm this "unction3 5hi#e
one technic# tester cn crete nd mintin the utomted scri)ts.
Ad(nt&es<
The %est )rt is tht scri)ts cn %e creted 5hi#e ))#iction de(e#o)ment is
sti## in )ro&ress. I" the "unction#ity 5ere to chn&e3 on#y the s)eci"ic %usiness
"unction scri)t 5ou#d h(e to %e modi"ied..
8ein& te,t records3 the dt in(o#(ed 6%oth in)ut nd out)ut7 . the e,)ected
resu#ts re esi#y mintin%#e.
'ecreses du)#icity in cretin& utomted test scri)ts s it uses modu#r
desi&n . uses "i#es to in)ut nd (eri"y dt.
8etter error hnd#in& s the "unctions return (#ues to the c##in& scri)t3 rther
thn %ortin&3 there%y incresin& the ro%ustness o" the test scri)ts.

1
'isd(nt&es<
Prior 4no5#ed&e needed in the scri)tin& #n&u&e used %y the too#
Mu#ti)#e dt:"i#es re needed "or e(ery test cse.
Mintennce o" the 'eti# Test P#n 5ith s)eci"ic dt %ecomes incresin&#y
im)ortnt.
).2 Test Plan ri"en +ethod
It is #so c##ed the 4ey:5ord dri(en ))roch. This method uses the ctu# test cse
document de(e#o)ed %y the tester usin& s)redsheet continin& s)eci# EFey:WordsE.
This method )reser(es most o" the d(nt&es o" the E$unction# 'ecom)ositionE
method3 5hi#e e#imintin& most o" the disd(nt&es. In this method3 the entire )rocess
is dt:dri(en3 inc#udin& "unction#ity. The Fey Words contro# the )rocessin& here.
Architecture
The rchitecture o" the Test P#n 'ri(en method ))ers simi#r to tht o" the
$unction# 'ecom)osition method3 %ut in "ct3 they re su%stnti##y di""erent.
The dri(er scri)t )er"orms the initi#i@tion i" needed 6>ust s in "unction#
decom)osition method7 nd c##s the ))#iction s)eci"ic contro##er scri)t3 5hich
in turn )rocesses the "i#e nme recei(ed "rom dri(er nd mtches on 4ey 5ords
contined in the in)ut "i#e3 %ui#ds )rmeter:#ist "rom the records . c##s uti#ity
scri)ts ssocited 5ith the 4ey 5ords. The uti#ity scri)ts in turn )rocesses in)ut
)rmeter:#ist recei(ed "rom the contro##er scri)t3 )er"orm s)eci"ic ts4s3 rise
errors tht mi&ht h(e occurred nd then return to the contro##er scri)t. User
de"ined "unctions3 %oth &ener# nd ))#iction:s)eci"ic ones cn %e c##ed %y ny
o" the scri)ts "rom %o(e.
Ad(nt&es
A)rt "rom the d(nt&es tht the "unction# decom)osition method hs3 it hs "e5
more )#us )oints too nd they re<
Ese o" 5or4in& s the deti# test )#n need %e 5ritten on#y once in the
s)redsheet "ormt continin& ## in)ut nd (eri"iction dt.
The %est )rt is tht e(en i" the tester doesn9t h(e )rior 4no5#ed&e on the
too#3 the 5or4 cn %e strtedB I.e. %e"ore the )#n is 5ritten3 i" the uti#ity
scri)ts re creted %y someone )ro"icient in the too#9s scri)tin& #n&u&e3 then
the tester cn use the test too# immedite#y (i the s)redsheet:in)ut method3
5ithout h(in& to #ern the scri)tin& #n&u&e. A## he needs to 4no5 re the
s)eci"ic 4ey:5ords nd the "ormt to use 5ithin the test )#n. This s(es time
s there is no dire need o" s)endin& time on trinin& the tester %out the too#
"irst nd then strtin& o"" 5ith codin&.
/e:us%i#ity is nother desir%#e thin& in this ))roch. A"ter num%er o"
&eneric Uti#ity scri)ts h(e #redy %een creted "or testin& n ))#iction3
most o" these cn %e re:used to test nother ))#iction.
'isd(nt&es

2
Prior 4no5#ed&e o" the too# s)eci"ic scri)tin& #n&u&e is needed "or cretin&
customi@ed "unctions nd uti#ities. So i" someone technic##y sound is )resent
to do it3 then "ine other5ise tht 5ou#d %ecome %ott#enec4 s in the
d(nt&e o" the tester not h(in& to 4no5 nythin& %out the too# 5ou#d no
more ho#d &ood.
In the e(ent o" ))#iction re;uirin& more thn "e5 customi@ed uti#ities3 the
tester is re;uired to #ern num%er o" 4ey 5ords nd s)eci# "ormts 5hich
cn %e )retty time:consumin&. Once the testers &et used to this3 ho5e(er3 the
time re;uired to )roduce test cse is &ret#y im)ro(ed.
Automtion ))#ied to di""erent ty)es o" testin&
Test #tage Ty$e of testing Amenity to
Automation
Factors that influence
/e&ression
Testin&
8#c4 8o, ?i&h $ocus is to test the im)ct on the st%i#ity
o" the ori&in# ))#iction
Tests re re)etiti(e 6Sme "unction#ity
need to %e tested "or e(ery re#ese7
Mode# O""ice
Testin&
8#c4 8o, ?i&h A))#iction 5i## %e hi&h#y st%#e
$in# #e(e# o" cce)tnce tests nd
there"ore #ess num%er o" outstndin&
de"ects
Per"ormnce
Testin&
8#c4 8o, ?i&h A))#iction 5i## %e "ir#y st%#e
Tests re re)etiti(e in nture 6Sme tests
need to re)eted "or (rious )rmeters7
User Acce)tnce
Testin&
8#c4 8o, Medium to ?i&h A))#iction 5i## %e "ir#y st%#e
'e)ends on the G o" outstndin& de"ects
"rom System Testin&
Num%er o" test cyc#es )#nned 6At #est 2
cyc#es shou#d %e )#nned7
System Testin& 8#c4 8o, Medium No re)et%i#ity = No st%i#ity in terms o"
"unction#ity
?i&h cost "or utomtion
Inte&rtion
Testin&
Grey 8o, !o5 System #e(e# inte&rtion testin&
White %o, nd not "unction# testin&
No re)et%i#ity nd no st%i#ity in terms
o" %oth technic# nd "unction# s)ects
Unit Testin& White 8o, !o5 Com)onent #e(e# 5hite %o, testin&
No re)et%i#ity nd No St%i#ity s
system is sti## in construction )hse
- ,esource Concerns
As discussed er#ier3 the myth tht utomtin& test )rocess 5ou#d m4e com)ny to
cut do5n on its st"" si@e is not e(en distnt#y true. A## tht these too#s do is cce#erte the
testin& )rocess. These test too#s use scri)ts 5hich utomtic##y e,ecute test cses. These
test scri)ts re nothin& %ut )ro&rms. They re 5ritten in the too# s)eci"ic scri)tin&
#n&u&e. Since these re )ro&rms3 their mn&ement hs to %e done in s much the
sme 5y s ny other code. To m4e this )ossi%#e3 n e,)ert in tht )rticu#r too# must

1H
%e hired or some5ht simi#r ro#e must %e creted nd rmed 5ith t #est one senior:
#e(e# )ro&rmmer. This e,)ert or 5hoe(er tht )erson is must %e %#e to desi&n3 de(e#o)3
test nd de%u& the code )rt "rom documentin& the sme. A#so3 )rt "rom ## this 5or4
re#ted to cretin& scri)ts nd "unctions3 this )erson must %e res)onsi%#e "or de(e#o)in&
stndrds nd )rocedures "or scri)t de(e#o)ment3 cretin& chn&e:mn&ement
)rocedures "or scri)t="unction im)#ementtion s #so testin&3 im)#ementin&3 nd
mn&in& the test scri)ts s 5e##.
. Things to remember
Su""icient )re)rtions must %e mde %e"ore strtin& the testin& utomtion )rocess.
More o"ten thn not3 this is not the cseB $o##o5in& re the thin&s tht must %e done to
)re)re "or the e""ortI.
An e,ct re)#ic o" the ctu# )roduction en(ironment must e,ist in 5hich the
sme ty)e o" hrd5re nd )ro&rms re inst##ed.
These utomtion scri)ts must h(e ## e,c#usi(e com)uters 6dedicted ones7
to %e run on. This is %ecuse s such they 5ou#d %e consumin& #r&e mount
o" memory nd 5ou#d thus %e s#o5. So i" ny e,tr )rocess 5ere to run on the
sme com)uter tht 5ou#d s#o5 do5n the rte o" scri)t e,ecution e(en more
resu#tin& in more time de#y.
It must %e )ossi%#e to restore the test en(ironmentJs dt%se to 4no5n
%se#ineK other5ise tests )er"ormed &inst the sme cn9t %e re)eted3 s the
dt 5ou#d h(e chn&ed
Initi##y the com)ny 5ou#d need some )ro"ession# triners to trin the
testers on the trinin& too# to %e em)#oyed3 %ut &oin& "or5rd3 it is u) to the
testers themse#(es s to ho5 much (#ue they cou#d dd to the 5or4 %y honin&
their s4i##s in the shortest )ossi%#e time. As they sy3 the sooner the %etterB
Tht 5s 5ith re&rds to the )re)rtion re;uired %e"ore ctu##y strtin& o"". 8ut there
re other thin&s s 5e## tht mtter. 'ecisions #i4e 5hich too# to use nd ho5 to
im)#ement it re mde %y the mn&ement3 o"ten 5ithout consu#tin& the )eo)#e 5ho re
ctu##y &oin& to test. This sometimes meets 5ith certin mount o" resistnce "rom the
testers. The thou&ht tht the too# cn re)#ce the testers is de(oid o" ny %it o" truth. A## it
does is he#)s the tester in doin& the >o% %etter nd "ster. Initi##y the testers mi&ht "ce
"e5 di""icu#ties in #ernin& ne5 thin& "rom the scrtch #to&ether 6in this cse3 ho5 the
too# 5or4s etc.73 %ut then )rctice m4es mn )er"ectB To strt o""3 )ro"ession# triners
cn %e hired to o(ercome the initi# hiccu)s %ut 5ith time3 the )eo)#e 5ho re ctu##y
&oin& to 5or4 on the too# must )ic4 u) the "iner deti#s "rom the triners nd mo(e on. I"
the test )#n dri(en method is used3 the testers 5i## not h(e to #ern ho5 to use the too# t
## i" not needed. A## tht they h(e to #ern is di""erent method o" documentin& the
deti#ed test cses3 usin& the 4ey:5ord=s)redsheet "ormt. And "rn4#y3 techies tht they
re3 it 5on9t t4e them #on& to #ern such thin&s.
A## sid nd done3 there sti## re mny thin&s tht need to %e t4en cre o" )rt "rom the
ones discussed %o(e. $or e.&. thin&s #i4e h(in& de"ined nd reson%#e &o#s s to 5ht
cn nd 5ht cnnot %e chie(ed And o" course3 %e"ore strtin& o""3 eductin& onese#"
%out the ctu# utomtion )rocess 5on9t %e %d ide t ##B Gettin& c#er
understndin& o" the re;uirements 5ou#d #5ys he#). It #5ys he#)s to identi"y the tests

11
tht re &ood cndidtes "or utomtion. Too com)#e, tests re %etter #e"t "or mnu#
testin&. Cost e""ecti(eness is #so o" s much concern s the "unction# e""ecti(eness o"
the too#. It is thus im)erti(e to se#ect too# tht im)#ements the )rocess in 5y tht
ensures smooth #on&:term testin&. Fno5in& ## these thin&s no53 utomtion shou#d no
#on&er remin n unherd o" conce)t ny more.
/ ,eferences
htt)<==555.s;:test.com=5L))er1.htm#
htt)<==555.testin&.com=5ritin&s=utomte.)d"
Automation Testing

12

Vous aimerez peut-être aussi