Você está na página 1de 10

International Journal of Advance Foundation and Research in Computer (IJAFRC)

Volume 1, Issue 7, July !1"# I$$% &"' ( "')&



1* + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
A Ris0 1ased $tory 2rioriti3ation 4echni5ue In An A-ile
6nvironment
Anita/Research Scholar*, Dr. Naresh Chauhan/ Chairman and Head of Department
YMCA Uniersit! of Science " #echnolo$!
anitaarora%&'(rediffmail.com*, nareshchauhan)*($mail.com

A 1 $ 4 R A C 4
A-ile soft.are development is an emer-in- area in this era as customer is the centre of attraction
.hile doin- development or any other customer centric tas0# 7ne of the representatives of the
customer is al.ays present on the development site so as to -ive the instant feed8ac0 and for any
future improvement in the sprint# At the same time, customer can furnish ne. re5uirement that
may distur8 the ori-inal functionin- of the e9istin- system# Althou-h, unit testin- and acceptance
testin- is on-oin- activity durin- the sprint as a part of re-ression testin-, still some 8u-s may -o
unnoticed due to lac0 of ris0 measure of any ne. re5uirement disclosed 8y the customer# In this
paper, an approach is discussed .hich is 8ased on complete user story matri9 that helps in
evaluatin- the overall desi-n measure of the user story# Further to add on, this measure predicts
the ris0 factor of any user story and may have su8stantial effect on the e9istin- stories# 4his
prioriti3ation on the 8asis of ris0 factors ensures the optimum utili3ation of resources such as
cost, time and effort#
Inde9 4erms: A-ile soft.are development, Complete user story matri9, ;esi-n measure, Ris0
measure, Re-ression testin-, $tory prioriti3ation

I# I%4R7;<C4I7%
#he A$ile pro+ect mana$ement is the approach used in the soft,are or manufacturin$ industr! so as to
attract more customers. -n this approach, processes and principles are d!namic in nature. Here, team
mem.ers and customer /&0 1one of the sta2eholder in an a$ile culture3 are the ma+or parties inoled
durin$ the .usiness deal. Dependin$ on their need, ,or2in$ rules, st!le and processes ta2e different
route so as to ma2e proper utili4ation of resources. Resources mainl! refer to people, time, mone! and
infrastructure. 5or2 to .e deliered to customer is fre6uent and response is also fre6uent from the
customer side. #his response ma! comprise of improement in the e7istin$ s!stem, ne, re6uirement,
ne, ,or2 st!le, scala.ilit! of the e7istin$ s!stem etc. 8ut of all these, ne, re6uirement is considered
throu$hout this paper so as to hi$hli$ht the importance of ne, chan$e framed .! customer durin$ or
after the sprint. #hese later introduced chan$es ma! hae seeral unnoticed effects in the ,or2in$
s!stem. #hese effects must .e controlled in a planned manner .! the team mem.ers so as to delier the
6ualit! deliera.le to the customer on time. Controllin$ of the e7istin$ s!stem is the first priorit! as per
the definition of the re$ression testin$ ,hich sa!s that ori$inal modules should not re$ress .!
introduction of ne, functionalit!/modules/user stories. Re$ression testin$ is performed .! retest all,
re$ression test selection /)0, test suite reduction or test case prioriti4ation /90 /:0. #est case prioriti4ation
/;0 is most popular one and ma! .e done .! usin$ arious factors. -n this paper, an approach is proposed
for findin$ the ris2 factor of the e7istin$ stories and ne, stor! introduced .! the customer. #his
measurement is useful for its outcome to the customer. #his measurement is done .! considerin$ the
stor! point of the user stor! and complete user stor! matri7. #hese results are used to proide ans,er to
6uestion li2e if one user stor! is chan$in$ ,hat other user stories of the s!stem should .e e7amined. -n
other ,ords, ,hat other modules hae potential to chan$e ,hen an! ne, module is introduced .!
International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

17 + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
customer after the sprint or closure of the pro+ect. Also, this approach can .e used for doin$ testin$ /<0 of
the most suita.le module/user stor! of the s!stem.
#his paper discloses prioriti4ation of modules on the .asis of ris2 /=0 /*0 /)'0 factor of the different user
stories of the s!stem. #he su.se6uent sections of this paper are prior art in this a$ile area, proposed
desi$n model, methodolo$!, results and discussion.

II# 2RI7R AR4
-n a$ile soft,are deelopment /))0, incremental and iteratie st!le of ,or2in$ is adopted .! most
professionals. 8ne of the ,ron$ misconceptions is that there is no need for ris2 mana$ement ,ithin the
a$ile st!le of ,or2in$. #his has led man! to .eliee that ris2 mana$ement />0 is irreleant in an iteratie
model. Some follo, the approach of i$norin$ ris2s until the! manifest into issues? the! then mana$e them
throu$h the natural sprint pro$ression.
@undamentall! spea2in$, a risk is somethin$ that ma! occur and cause une7pected or un,elcomed
outcomes. #he outcome ma! hae a positie or a ne$atie effect on the oerall ,or2in$. A positie effect
is an opportunit!, ,hile a ne$atie effect ma! lead to do,nfall in .usiness/dissatisfaction to a customer.
As there is a pro.a.ilit! aspect attached to ris2, its e7act occurrence is un2no,n .ut ris2 aries from
pro+ect to pro+ect. #hatAs ,h!, mana$in$ ris2 is trou.lesome for ne,.ieAs in an a$ile ,orld.
-n traditional st!le of deelopment, ris2 mana$ement ma! .e done .! maintainin$ a ris2 re$ister for
controllin$ different t!pes of ris2s. A simple ris2 re$ister ma! consist of the follo,in$ attri.utesB
Description of ris2B A oneC or t,oCline oerie, of the ris2. -t should .e precise one.
-dentification dateB Date ,hen the ris2 ,as identified.
Di2elihoodB Estimated pro.a.ilit! of occurrence of the ris2.
Seerit!B #he seerit! of the ris2 is assessed .ased on impact of the undesired outcome.
Friorit! 1optional3B #his could .e either $ien an independent alue or set as a product of
li2elihood and seerit!. A hi$hCseerit! ris2 ,ith a hi$h li2elihood should receie more
importance than a hi$hCseerit! ris2 ,ith a lo, li2elihood.
ActorB #he person ,ho mana$es, controls, and ta2es action in response to the ris2.
ActionB #he response defined to mana$e/control the ris2.
StatusB -ndicates ,hether the ris2 is re$istered or open or closed or .ein$ monitored.
-n an a$ile enironment, a colla.oratie approach for ris2 mana$ement is $enerall! used. #his approach
is open to all team mem.ers. -t is imperatie that the ris2 re$ister .e made aaila.le for the team so that
it can .e mana$ed, controlled and monitored colla.oratiel!. At eer! sprint meetin$, the ris2 re$ister is
reie,ed and updated ,ith an! ne, information o.tained oer the sprint. #his ,a! ris2 mana$ement
.ecomes an inte$ral part of A$ile.
Another interestin$ techni6ue, introduced .! Gohn Hrothers 1Agile Times, &''<3, relies on usin$ a Ris2
HurnCdo,n Chart. As ela.orated .! Mi2e Cohn, the ris2s are collated into a ta.le similar to a ris2 re$ister.
-t consists of the follo,in$ elementsB
Ris2B Description of the ris2 in a fe, lines.
Fro.a.ilit!B Di2elihood of the ris2.
Si4e of lossB Amount of time lost ,hen the ris2 occur. #his could .e represented in da!s or stor!
points.
E7posureB #his is computed as a product of the pro.a.ilit! and si4e of loss 1a.oe3.
International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

1' + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
#his ris2 re$ister is reCealuated at eer! sprint meetin$? its alues are ad+usted .ased on the current
assessment of the e7istin$ and ne, ris2s. #his ,ould define a ne, alue for the consolidated ris2
e7posure. #he Ris2 HurnCdo,n Chart ma! .e created .! plottin$ the consolidated ris2 e7posure across
the num.er of sprints run .! the team. -n short, the .urnCdo,n chart represents the status of the ris2
across the iterations. @rom a pro+ect mana$ement perspectie, this is an e7cellent indicator of ho, the
ris2s are mana$ed and controlled.
#hese all approaches are post identification techni6uesB means li2elihood of the ris2 is identified
dependin$ upon preious e7perience of the team mem.ers or sprint reie, meetin$ or on similar pro+ect
.asis. -n this paper, a pre identification approach has .een proposed that is .ased on desi$n measure of
the indiidual stor! of the sprint. #his determination is done .ased on the lin2 that e7ists amon$ the
stories of the sprint. Dependin$ upon the lin2As ,ei$ht, a desi$n parameter for the sprint is measured and
anal!4ed for ris2 assessment. #his oerall desi$n measurement of the sprint helps in sta.ili4in$ the
s!stem to the $reat e7tent.
III# 2R727$6; =7;6>
#he a$ile desi$ns are d!namic in nature as these desi$ns emer$e oer time. Desi$n is dependent on the
re6uirements framed .! the client at the time of conception of the idea. As re6uirements 2eep on addin$,
desi$n ma! eole. #he re6uirements in the a$ile enironment are 2no,n as epic 1oerie, of the ma+or
tas2 in &C> lines3. @urther, epic are splitted into small stories. #he ne7t step follo,in$ this splittin$ is the
estimation of the stor! points in conte7t ,ith the comple7it! of the user stor!. #he .loc2 dia$ram for the
a.oe process is represented in the fi$ure ). Estimation for user stories ma! .e done in man! ,a!s. 8ne
of the ,a!s is plannin$ po2er in ,hich man! e7perts sit to$ether and dependin$ upon their e7perience
the! assi$n some point to the user stor!.

Fi-ure 1# A-ile 6nvironment

Requirement
s
Story
Splitter
Estimation
Idea
Conception
User story 1
User story 2
User story n
Story Point
2
Story Point
n
Story Point
1
International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

1? + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
#his estimation process ma! comprise of some e7perts in the field of a$ile. #hese e7perts after assi$nin$
stor! point num.er to the user stor! +ustif! that num.er and collectiel! decision is ta2en to reflect the
one +ustifia.le stor! point as per the comple7it! of the user stor!. #he defined series for estimatin$ stor!
point ma! .e an! arithmetic series, $eometric series or @i.onacci series. E7amples ma! include ', ), ), &,
>, 9, ;..................or &, <, :, ;, )'.................etc. @or the sa2e of simplicit!, @i.onacci series has .een chosen to
demonstrate the scenarios discussed in this paper. -n the fi$ure, stor! splitter is the one component that
helps in splittin$ the epic into num.er of stories ,hich are further processed or e7ecuted in the follo,in$
sprints.
-n this paper, an approach has .een proposed that is .ased on the stor! points of the user stor! and
dependin$ upon these stor! points, a ris2 measure matri7 has .een proposed that is the outcome from
complete user stor! matri7As alues. #he complete user stor! matri7 and ris2 measure matri7 ma! .e
ena.led .! considerin$ the further disclosure. #he disclosure of the proposed ,or2 comprises user stor!
dia$ram, complete user stor! matri7, desi$n matri7 and ris2 measure matri7. #hese components are
sho,n .elo, in fi$ure &.
A# <ser $tory ;ia-ram
@irst and foremost component is user stor! dia$ram. User stor! dia$ram sho,s connection amon$
splitted user stories and their dependence in terms of user stor! point. A user stor! point is a measure to
estimate the effort inoled in implementin$ a user stor!. #a.le ) represents the user stor! point
estimation for the splitted stories. Here, in the ta.le sho,n .elo,, four stories are estimated to hae
stor! point ), ;, > and & respectiel!. #hese estimations are .ased on plannin$ po2er $ame. -ts primar!
do,nside has .een that all participants/e7perts had to .e sittin$ in the same room ,ith a ph!sical dec2
of cards in their hands. #his stor! point estimator is an a.stract alue as it is .ased on the .aseline that is
chosen .! the e7perts. #his stor! point alue is further used in dra,in$ user stor! dia$ram. Din2 .et,een
user stories is represented .! some ,ei$ht this ,ei$ht is sum of indiidual stor! point of connected
stor!. @or e7ample, ed$e FI has ,ei$ht *, ,hich is sum of indiidual stor! points, as F stor! has
connection ,ith I and S. F has stor! point ) and I has stor! point ;. -n total, their com.ined effect is *.
Similarl!, other ,ei$hts are calculated.


Fi-ure # 2roposed =odel

User Story Diagram
Risk Measure
Matrix
Complete User
Story Matrix
Dependency
Matrix
International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

! + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
4a8le 1# $tory 6ffort
$#%o# $tory %ame $tory 2oint
). F )
&. I ;
>. R >
<. S &

-n @i$ure >, ,ei$hted $raph is sho,n. Din2s of the $raph are sho,n to .e undirected. #hus, F and I user
stor! are dependent on each other. -t means F and I are re6uired for moin$ to ne7t step. #hatAs ,h!
com.ined effort is calculated.
#he first leel dependence matri7 for @i$ure > is sho,n .elo, in #a.le &. Dia$onal entries in the first
leel dependence matri7 are sho,n to .e ' as no lin2 e7ists .et,een selfCuser stories. Rest of the nonC
4ero entries is as per the user stor! dia$ram. -f no lin2 e7ists .et,een t,o different user stor! then that
cell is represented .! dash. @or e7ample, in @i$ure >, lin2 FR does not e7ist, so, in #a.le &, FR and RF cell
entr! is filled ,ith NA.

Fi-ure &# <ser $tory ;ia-ram

4a8le # First >evel ;ependence 4a8le

2 @ R $
F ' * NA >
I * ' )) )'
R NA )) ' 9
S > )' 9 '

International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

1 + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
1# Complete <ser $tory =atri9
#he complete user stor! matri7 is the matri7 sho,in$ effectie effort for user stories usin$ indirect and
direct lin2s ,ei$hts. @or e7ample, if client has introduced chan$e in user stor! F of user stor! dia$ram
then definitel! user stor! SAs effort ,ould chan$e as there is direct lin2 .et,een these t,o user stories.
8n the other hand, rest of the user stories effort ,ill chan$e as indirect paths e7ist from source F to
destination S. D represents direct lin2 and - represents indirect lin2. Considerin$ the same e7ample, F to S
$ies one direct lin2 and t,o indirect lin2s. #heir ,ei$hts in terms of com.ined user stor! points are
sho,n in #a.le >.
4a8le &# 6ffort ;ata

>in0s ;escription Aei-ht
FCS Direct >
FCICS -ndirect **)'J*'
FCICRCS -ndirect **))*9J<*9

#hese eents are not mutuall! e7clusie, so follo,in$ formula is used to trac2 the cumulatie effect ,hen
there is an! chan$e in the e7istin$ user stor! S.
Fro.JF1K3 L F1Y3 L F1M3 C F1K3.F1Y3 C F1Y3.F1M3 C F1K3.F1M3 L F1K3.F1Y3.F1M3 1)3
5here F1K3CCCCCC5ei$ht of direct path
F1Y3CCCCCCC5ei$ht of indirect path )
F1M3CCCCCCC5ei$ht of indirect path &
Usin$ e6uation ), Fro.a.ilit!, pro., of chan$in$ user stor! S is ;=*>>. Similarl!, other alues ma! .e
calculated usin$ e6uation ). Ela.oration for findin$ indirect paths is sho,n .elo, in @i$ure <.
Dependin$ upon num.er of indirect paths, e6uation num.er one can .e e7tended accordin$l!. -ndirect
paths in an undirected $raph ma! .e determined .! usin$ .readth first search implementation ,hich is
.ased on ad+acenc! list. @or the discussed case, complete user stor! matri7 is represented in #a.le <.
Si7th Column of the matri7 is sum of the column entr! in indiidual ro, of the matri7. #his column is
2no,n as dependenc! matri7. #his alue is determination of the ris2 inoled for the user stor!.
Ma7imum alue in the ta.le is for user stor! F and minimum alue is for user stor! R. Ma7imum alue in
dependenc! matri7 is representation of hi$h ris2 stor!. #herefore, plannin$ and estimation efforts are
more for user stor! F as compared to I, R, and S user stor!. #hats ,h!, focus is more on F user stor! .!
team mem.ers so that ris2 can .e tac2led easil!.
4a8le "# Complete <ser $tory =atri9
2 @ R $ $<= ;ependency
2 ' >;'<* ' ;=*>> )&9*;& =a9imum
@ >;'<* ' C<;* C<;9 >='=9
R ' C<;* ' C<>9 C*&< =inimum
$ ;=*>> C<;9 C<>9 ' ;=')>

International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

+ , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
IV# R6$<>4$ B A%A>C$I$
#he proposed model sho,s that user stor! F is hi$h ris2 stor!. #he cumulatie results sho,n in #a.le 9
predicts that F or an! other stor! is directl! or indirectl! lin2ed to rest of the stories and chan$e in
respect of other stor! ma! affect F or an! other correspondin$ user stor!. No,here, in the prior art it is
accomplished that stor! point is lin2ed ,ith the ris2 factor of an! user stor!. So, sho,n results are useful
for prioriti4ation of e7istin$ user stories after calculatin$ complete user stor! matri7. Ris2 matri7 reeals
the order of user stor! prioriti4ation as F, S, I and R. R user stor! is least ris2! stor!.
4a8le )#Ris0 =atri9
$47RC $<= ;626%;6%CC
2 )&9*;& Most Dependent
@ >='=9
R C*&< Deast Dependent
$ ;=')>

After findin$ the most ris2! user stor!, testin$ efforts ma! .e calculated from the .e$innin$ of the sprint.
A hi$h ris2! stor! of the sprint is al,a!s centre of attraction. Eer! reie,er, of that stor!, has to focus
more for accomplishin$ .u$ free deliera.le as compared to less ris2! stor!. -t does not mean that less
ris2! stor! ma! .e i$nored. Hoth hae e6ual importance. 8nl! difference is hi$h ris2! stor! demands
more reie,s as compared ,ith less ris2! stor!. Also a hi$h ris2! stor! demands e7perienced
professional as compared ,ith a less ris2! stor!. -n li$ht of this, a reie,er ma! concentrate on follo,in$
points .efore releasin$ deliera.le 1soft,are3 to customer from her endB
1a3 Minimum usa$e and ma7imum usa$e of deliera.le is chec2ed.
1.3 8ptimum resources needed for runnin$ soft,are are erified.
1c3 Read! stor! of user stor! is mapped ,ith the deliera.le.
1d3 Freious deliera.le mista2es are not repeated if customer is same.
1e3 Standard set .! customers are fulfilled ,hether she is pilot or non pilot customer.
1f3 Criteria such as 6ualit! and timel! delier! of outcome is fulfilled.
1$3 Manual has to .e released alon$ ,ith the deliera.le for handlin$ e7ceptional conditions.

International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

& + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-


Fi-ure "# 1F$ 8ased Indirect path Implementation


International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

" + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
V# C7%C><$I7%
Frioriti4ation in soft,are deelopment has .een implemented for decades ,ith arious models,
approaches and tools. E7amples ma! include re6uirements prioriti4ation, module prioriti4ation, test case
prioriti4ationCone of the re$ression testin$ method, desi$n preferences such as top do,n or .ottom up,
tools preferences etc. -n this paper, a ris2 .ased stor! prioriti4ation techni6ue is proposed that ma! .e
used to prioriti4e user stor! 1formal representation3 framed .! a$ile team mem.ers .ased on the
customerAs informal re6uirements. Usin$ said techni6ue, correspondin$ ris2 factor is calculated for the
formal user stor!. @urther, ris2 matri7 is used for measurin$ the hi$hl! ris2! stor!. #his measurement is
performed in the sprint plannin$ phase of an! a$ile methodolo$!. Durin$ sprint implementation sta$e,
a$ile team ma! focus more on the identified stor! hain$ more ris2. Su.se6uentl!, stor! prioriti4ation is
performed for rest of the stories.
Application area is not onl! limited to soft,are deelopment, rather this techni6ue ma! .e used to
prioriti4e stories in an! area such as manufacturin$ sector, automo.ile sector ,here a$ile is implemented
in closed colla.oration ,ith customer or client. #his techni6ue ma! also .e used for implementin$
NANHAN concept for controllin$ ,or2flo, and pullin$ 2e! tas2s. NANHAN is one of the pro+ect
mana$ement tool that is used in a$ile implemented sector. -n future, automation ma! .e applied to
implement the proposed techni6ue so as to utili4e resources in optimum manner. #his ma! help in
reducin$ time, cost and effort.
VI# R6F6R6%C6$

/)0 A. Anita and C. Naresh, OA re$ression test selection techni6ue .! optimi4in$ user stories in an
a$ile enironmentP, pu.lished in <
th
-EEE -nternational Adance Computin$ Conference, @e. &)C
&&, &')<

/&0 A. Anita and C. Naresh,PA Simplest A$ile Dife C!cle for All Sta2eholdersP, =
th
international
conference on soft,are en$ineerin$, )9
th
C)=
th
No &')>, Fune, -ndia

/>0 Sin$h Yo$esh and A$$ar,al N.N., Soft,are En$ineerin$, Ne, A$e -nternational Fu.lishers,
Reised second edition &''9

/<0 A. Anita and C. Naresh ,O#estin$ in an A$ile EnironmentB A Fro+ectP, -n )
st
-nternational
conference on ne7t Qeneration Communication " Computin$ S!stems1-CNQC&SC)'3, &<
th
C&9
th
Dec
&')', Chandi$arh, -ndia

/90 A. Anita and C. Naresh,P A Din$uistic Approach for #est Case Frioriti4ation in an A$ile
EnironmentP, pu.lished in )>
th
Annual -nternational soft,are testin$ conference &')>, <
th
C9
th

Decem.er, Han$alore &')>.
/:0 A. Anita and C. Naresh,P A Fattern Hased Approach #o Frioriti4e #est Cases @or User Stories -n An
A$ile EnironmentP, pu.lished in -nternational Gournal of Adance @oundation and Research in
Computer 1-GA@RC3 Feer Reie,ed And Iualit! Article Monthl! Gournal -SSN &><; R <;9> in Ma!
&')<.
/=0 Games Hach, SRis2 Hased #estin$S, S#IE Ma$a4ine,:/)***, ,,,.st6ema$a4ine.com
International Journal of Advance Foundation and Research in Computer (IJAFRC)
Volume 1, Issue 7, July !1"# I$$% &"' ( "')&

) + , !1", IJAFRC All Ri-hts Reserved ...#i/afrc#or-
/;0 Goachim Narlsson " Nein R!an, SA CostCTalue Approach for Frioriti4in$ Re6uirementsS, -EEE
Soft,are, Sept. )**=
/*0 Ris2CHased #estin$, S#Da.s Report, ol. > no. 9 1info(stla.s.com3
/)'0 StUle Amland, SRis2 Hased #estin$ of a Dar$e @inancial ApplicationS, Froceedin$s of the )<th
-nternational Conference and E7position on #ES#-NQ Computer Soft,are, Gune ):C)*, )**=,
5ashin$ton, D.C., USA
/))0 Coc2.urn, A., and Hi$hsmith, G., OA$ile Soft,are DeelopmentB #he Feople @actorP, -n Computer,
No. &''), pp. )>)C)>>.