TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 13, No. 3, March 2
015,
pp. 555 ~ 56
0
DOI: 10.115
9
1
/telkomni
ka.
v
13i3.713
4
555
Re
cei
v
ed
De
cem
ber 3, 20
14; Re
vised Janua
ry 2
7
, 20
15; Accepted
February 12,
2015
Knowledge Work Process: So
ftware Developer’s in
Small Medium Enterprise
Mohd Zairol Yusoff*
1
, Massudi Mahm
uddin
2
, Maz
i
da Ahmad
3
Schoo
l of Com
putin
g, Univ
ersiti Utara Mal
a
ysia
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: zarul1
x@
gm
ail.com
1
, ady
@
uum.edu.my
2
, mazid
a
@u
um.edu.m
y
3
A
b
st
r
a
ct
Mana
gin
g
kno
w
ledge w
o
rk i
n
the w
o
rkpla
c
e is
in
here
n
tly importa
nt a
nd acc
e
ssib
l
e
to th
e
orga
ni
z
a
ti
ons for the lon
g
terms grow
th an
d
perfo
rmanc
e. Softw
are devel
oper is a key s
u
ccessor for th
e
orga
ni
z
a
ti
on su
ccess and kn
o
w
ledge w
o
rk is view
ed as the
hig
hest co
mpl
e
xity of
w
o
rk characteristics. T
h
e
intenti
ons
of s
o
ftw
are dev
elo
pers to
i
m
prov
e the
kn
ow
led
ge w
o
rk
proc
e
ss are
re
ma
in
unco
n
scio
u
sn
e
ss.
This p
aper
w
ill
addr
ess the
iss
ue
of know
l
edg
e w
o
rk
pr
ocess
an
d try to
pro
pose
a
metho
d
how
to
i
m
prov
e
know
led
ge w
o
r
k
process b
a
se
d on
distinct
methods
an
d ap
proac
hes. A lit
erature r
e
view
w
a
s used i
n
or
der
to disting
u
ish t
he metho
d
s a
nd w
ill use dat
a collect
ed 30
0 respo
nde
nts from Sma
ll M
edi
u
m
Enterpri
se
(SMEs) in Mal
a
ysia a
nd val
i
d
a
te the metho
d
s by usin
g structural e
q
u
a
tio
n
mo
de
lli
ng. Our results prov
i
d
e
evid
ence
on th
e i
m
p
o
rtance
o
f
certain
meth
o
d
to i
m
pr
ove k
now
led
ge w
o
rk
on the s
o
ftw
are dev
elo
pers
a
n
d
busi
ness succ
ess, and h
a
ve i
m
p
licati
ons for
both
rese
arch
and pr
actice i
n
the field of SM
Es.
Ke
y
w
ords
:
know
led
ge w
o
r
k
, Know
ledg
e w
o
rk process, know
led
ge w
o
r
k
er, softw
are devel
oper
’
s
an
d
sma
l
l
me
di
um enterp
r
ise
(SMEs)
Copy
right
©
2015 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Managi
ng
kn
owle
dge
in th
e adva
n
ced
o
f
techn
o
logy has
ma
de greater
sen
s
e e
n
tirely
to
the pro
c
e
ss a
nd pro
d
u
c
tivity of knowle
dg
e work. Ac
cording to [1] the most importa
nt contrib
u
tio
n
manag
eme
n
t need
s to m
a
ke i
n
the 2
1
st
centu
r
y is simila
rly to increa
se the
prod
uctivity of
kno
w
le
dge work a
nd the kno
w
le
dge worker an
d the most valua
b
le assets of
a 21
st
centu
r
y
wheth
e
r bu
si
ness or n
o
n
-
b
u
sin
e
ss, will
be its kn
owl
e
dge worke
r
s
and their p
r
o
ductivity.
The term kno
w
led
ge wo
rk deal with the
peopl
e who u
s
ed formal kn
owle
dge a
s
a
major
part of thei
r jo
b. By definition kn
owl
edg
e
works i
s
rel
a
ted to the
any
activities
that requi
re sp
eci
a
l
kno
w
le
dge o
r
skill
or
creat
es n
e
w
kn
owl
edge [2]. It involved a crea
tive work solving un
stru
ctu
r
ed
probl
em
s that
re
quire expl
o
r
ation
and
cre
a
tion of
knowl
edge
[3]. Referen
c
e
[2]
define it as “a
s
e
t
of activities
usin
g in
divid
ual a
n
d
external
kno
w
led
ge
to
produ
ce output
s chara
c
te
rized by
informatio
n content.” Kno
w
led
ge work pre
c
ise more on the
stronge
r comm
unication ne
eds,
assign
multip
le role
to the
perso
n rath
er tha
n
with
the sin
g
le jo
b po
sition a
n
d
incre
a
se
th
e
importa
nce of
team work. It also
com
p
ri
se
s tra
n
sa
cti
ons, inte
ra
ctions
and
de
ci
sion m
a
ki
ng t
hat
requi
re
th
e continuo
us re
vising
a
nd
i
m
provin
g
of the
re
so
urce
kno
w
le
dge. It
means
all the
activities
relat
ed
to
the kno
w
led
ge work stre
ss
the ch
ange
s
in
th
e pro
c
e
ss and pra
c
tice co
m
pare
to traditional
works [4].
It’s never
dou
bt that softwa
r
e d
e
velope
r
as a
kn
owl
e
d
ge worke
r
will
bring
a fun
d
amental
cha
nge in th
e stru
ctu
r
e a
nd process
of kno
w
le
dg
e wo
rk i
n
th
e ICT infrast
r
uctu
re [3].
The
stru
cture an
d
pro
c
e
s
s of
knowl
edge
wo
rk adversely contrast
to p
h
ysical
work whi
c
h
e
m
pha
size
on the a
pplyi
ng existin
g
knowl
edge
or
cre
a
te
ne
w knowl
edge
of softwa
r
e
d
e
velope
r’s ba
se
on
the wo
rk e
n
vironm
ent. It
mean
s the software
dev
el
oper
who ex
pertise in the
particular a
r
ea
need
to u
s
e
their
co
gnitive skill
with
e
ngagi
ng i
n
th
e compl
e
xity of processe
s in the
software
developm
ent [2].
The process
approa
ch allo
ws a
n
end
-to
-
end vi
e
w
of how b
e
st to stru
cture, se
quen
ce,
and
mea
s
u
r
e
wo
rk a
c
tivities to
re
ach targeted
outcom
e
s
.
Ho
wever
the natu
r
e
of
the p
r
o
c
ess o
f
kno
w
le
dge
work i
s
difficult
to describ
e
[5].
The inpu
ts and o
u
tput
s of kn
owl
e
d
ge wo
rk idea
s,
interruption
s
,
inspiratio
ns,
and
so o
n
are often
l
e
ss tan
g
ible
and
di
scret
e
.
The
r
e are
no
pred
etermi
ne
d task se
que
nce
s
that, if exec
ute
d
, gu
arante
e
the
desi
r
ed
outcome. Knowl
e
dge
worke
r
s may
ope
rate by
an intuitive
feel for h
o
w
to a
c
com
p
lish th
eir
work or thro
u
g
h
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 3, March 2
015 : 555 – 5
6
0
556
accumul
a
ted experie
nce. Although
ma
ny of the
re
searche
r
s give
more attenti
on to
wards of
the
pro
c
e
ss in th
e kno
w
le
dge
work such [5] explain the reen
gine
eri
ng and lai
ssez faire meth
ods
empha
si
ze th
e improvem
e
n
t of the process of kn
o
w
le
dge work, [4]
state the pote
n
tial analysi
s
of
kno
w
le
dge
work aim to
en
han
ce the
ab
ility of handli
ng kno
w
led
g
e
and [
3
] stat
e the p
r
o
c
e
s
s in
the modeli
ng
of kno
w
le
dge
wo
rk i
n
fra
s
tructure but
l
a
ck of detail
an
alysis
relate
d
to the soft
wa
re
develop
ers. The cl
assi
cal
model
trie
s to adopt the i
nput and o
u
tput
system f
o
r the creatio
n of
the intellectu
a
l assets, b
u
t still not configure
with
the structu
r
e and pro
c
e
ss an
alysi
s
of
kno
w
le
dge work.
In the case
of software
develope
r, ther
e is a
widely used a
ppro
a
ch to measure
kno
w
le
dge work process orientatio
n.
Software
E
ngi
neeri
ng In
stitute Ca
pabilit
y Maturity Model
(CMM
), ha
s introdu
ce
d a
method that allows the
analysi
s
of the different l
e
vels of pro
c
ess
maturity has
see
n
the two
group
s in a comp
any th
a
t
is in CMM Level 5, the
highe
st level o
f
pro
c
e
ss m
a
tu
rity, and the other two g
r
ou
ps in the
sam
e
comp
any in
Level 3. They found that, for
the most part
,
mature pro
c
e
s
ses expe
rienc
ed by many develope
rs that enabl
e and empo
we
r
rather than
coerc
i
on and is
olation[6].It mean
s
s
o
ftware developers expe
rienc
e
d the increased
pro
c
e
ssi
ng o
r
ientation a
s
p
o
sitive [7].
Definition of SMEs
ha
s al
so ch
ang
ed. North
A
m
e
r
ic
an I
n
d
u
st
ry
Cla
ssif
i
cat
i
on
S
y
st
em
use
s
me
asures such a
s
: the numbe
r of em
ploye
e
s an
d total
revenu
e, depen
ding o
n
the
indu
stry. The
Europ
ean
Union (E
U)
ha
s create
d
a
uniform
defini
t
ion: indepe
n
dent compa
n
i
e
s
with less tha
n
250 emplo
y
ees and h
a
ve either a
turnover of less than 40 million euros o
r
total
assets
of less than
27 mi
llion eu
ro [8].
Most
of the
softwa
r
e
de
velopers
wo
rk at the SM
Es
company [8].
But lack of the requi
re knowledge
and skill
caused
by
less of financial
or access
to funding
an
d wo
rking
ca
pital to create
gre
a
ter
ri
sk i
n
SMEs [9].
Their lack
of
awa
r
en
ess
of the
importa
nce o
f
adopting b
e
st bu
sine
ss pra
c
tice
s an
d quality ma
nagem
ent sy
stem
s, su
ch
as
financi
a
l ma
nagem
ent a
nd cu
stom
er focused a
c
tivities, in order to imp
r
o
v
e a comp
a
n
y's
prod
uctivity and profitabilit
y. Furthermo
re mo
st of
the proj
ect
s
were rel
a
ted
to the failure of
SMEs
Comp
any’s
asso
cia
t
ed with
in
co
rrect
spe
c
if
ica
t
ion
of req
u
irements
[10]. In the meantime
the software
org
ani
zation
s
can
not p
r
o
perly
m
anag
e their software
process
and th
e
sam
e
mistake re
pe
ated after d
e
ca
de [10]. They lack
a
w
arene
ss of the importa
nce of ad
op
ting
busi
n
e
s
s be
st pra
c
tice
s an
d qu
ality man
ageme
n
t sy
stems,
su
ch
as finan
cial m
a
nagem
ent a
n
d
cu
stome
r
focuse
d activitie
s
, in orde
r to enha
nc
e the firms’ p
r
od
ucti
vity and profitability [11].
1.1. The Proposed Me
th
od and Ap
pr
oach
In this pa
rt of
resea
r
ch, we
have id
entif
ie
d an
d m
a
ke a
co
mpa
r
ison
based
on
met
hod
or
approa
ch to
improve the
kno
w
le
dge
work
pro
c
e
s
s. This meth
od
is adapte
d
from kno
w
led
ge
intensity mod
e
l [4] and wo
rk
segm
ents
[7] which
a
ssess the relev
ance of kno
w
ledge
work in
a
major wo
rk p
r
ocess
and
p
r
odu
ct o
r
se
rvices.
T
he
hi
gher of
kno
w
ledge i
n
ten
s
ity, meaning
the
highe
r
of the
effectivene
ss of h
andli
ng t
he
kno
w
le
dg
e
an
d be
com
e
s an impo
rt
ant factor of the
kno
w
le
dge work p
r
o
c
e
s
s.
Acco
rdi
ng to this method t
he level of know
l
edge int
ensity determ
i
nes the rout
e to th
e
kno
w
le
dge
w
o
rk
p
r
o
c
e
ss.
I
n
ca
se
s
of
l
o
w
kn
owl
edg
e int
e
n
s
it
y
is
relat
e
d
t
o
t
h
e
wo
r
k
ef
f
i
ci
en
cy
and fo
cu
s di
rectly ho
w to
use
re
so
urce
s effici
ent
ly (e.g. Shop flo
o
r p
r
od
uctio
n
of bolts). B
u
t
relatively the higher l
e
vel of knowl
edg
e intens
ity m
ean
s the pro
c
e
ss of
kno
w
led
ge work is
compli
cate
d
and
use
full
of kn
owle
dge
as ce
ntral
reso
urce
s a
n
d
will
dete
r
m
i
ne the l
e
vel
of
stand
ard h
a
n
d
ling the info
rmation, com
m
unication a
nd de
cisio
n
makin
g
(e.g.
an investme
nt
ban
k, school,
rese
arch la
b and software develop
ers).
The
se
cond
method i
s
tryi
ng to imp
o
se, pre
c
isely in t
he process
a
nd kno
w
led
g
e
activity.
There a
r
e
so
me key differences in
pro
c
e
s
s ori
entati
ons amo
ng
d
i
fferent types of kno
w
led
g
e
work
and
workers. Ba
se
d
on the
matrix
sho
w
n
in figu
re 2
the
r
e a
r
e
four
key type
s of
kn
owl
e
d
g
e
work ba
se
d on the degre
e
of expertise and the leve
l of coordi
n
a
tion in the work [7]. The first
approa
ch fo
cuse
s o
n
the
work
of the transactio
n
, which i
s
u
s
u
a
ll
y more m
a
lle
able in te
rm
s of
pro
c
e
s
s than
others, b
e
ca
use
the
wo
rk that is
us
u
a
l
ly
r
e
cu
rr
e
n
t
an
d
k
n
ow
le
dge
-
b
as
ed
w
o
rke
r
s
formalitie
s an
d pro
c
ed
ures
pro
c
e
ss. Thi
s
means
le
ss knowl
edge of t
he activities u
nderta
ke
n an
d
pro
c
e
s
s flows into
som
e
form of
co
mputer
-ba
s
e
d
appli
c
atio
n
s
. The
syste
m
s me
asure
the
pro
c
e
s
s and
usu
a
lly brin
g
s
the
wo
rk
an
d all info
rmati
on an
d kno
w
l
edge
req
u
ire
d
to pe
rform i
t
to
the worke
r
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Knowle
dge
Wo
rk Process: Software
Devel
ope
r
’
s in
Sm
all Medium
… (Mohd Z
a
irol Yu
soff)
557
Figure 1. Potential Analysi
s
for Kno
w
led
ge wo
rk Process
Integration
work is
often q
u
ite structu
r
e
d
, ev
en hig
h
e
r
level
s
of co
operation ofte
n lead to
a mo
re
com
p
lex pro
c
e
s
s.
Employee-ori
ented int
egr
a
t
ed quite
po
ssible
to ad
opt
the inte
rventi
o
n
pro
c
e
ss. It is
possibl
e to articulate the p
r
oc
e
s
s to be followed in th
e docume
n
t, and em
ploye
e
s
usu
a
lly have enou
gh time and discretio
n to negotiate
with the docume
n
t. An importa
nt part
of
the process,
but doe
s
not
descri
be
th
e pra
c
tice, but must
follo
w
standard o
perating p
r
o
c
edu
res
within the org
anization.
Expert
wo
rk can be
ma
de
more pro
c
e
ss orie
nted,
but they’re e
x
perts often
hold th
e
process imposed.
Usua
lly,
someone
should give them
the
ab
ility
to overcome
or step out of
the
pro
c
e
ss. Th
e
s
e worke
r
s h
a
ve a high d
egre
e
of auto
nomy and di
scretio
n
in thei
r wo
rk, but m
o
s
t
of the
work a
pplied
techno
logy to
key a
s
pe
cts of
th
e
p
r
oc
es
s
.
Most k
n
ow
led
g
e
w
o
rk
ers
in
vo
lve
d
in the interve
n
tion process of kno
w
led
g
e
creation
in
the intervent
ion proc
ess. The la
st of the
approa
ch is
colla
borative work. It presents a
challe
nge for process orie
nted
manage
rs and
these worke
r
s typically ha
ve a more iterative,
collab
o
rative app
ro
ach to wo
rk f
o
r whi
c
h patt
e
rn
s
are m
o
re
difficult to di
sce
r
n. In this
ap
proa
ch
wo
rk
stru
cture can
be de
nied
a
nd likely they
alway
s
p
r
ovi
de a
n
inte
rvention
app
ro
ach
e
s to
m
e
et their task.
Knowl
edg
e
distrib
u
tion
a
n
d
appli
c
ation is
promi
nently work is thi
s
a
ppro
a
ch.
Figure 2. Approa
ch to Kno
w
led
ge work pro
c
e
s
s
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 3, March 2
015 : 555 – 5
6
0
558
The m
o
st i
m
portant
for th
e softwa
r
e
d
e
velope
r
su
ch a
s
p
r
o
g
ra
mmer is to t
h
ink the
pro
c
e
ss i
n
terms
of kno
w
l
edge a
c
tivities invo
lved. It mean
s the
pro
c
e
ss
ca
n be defe
rre
d
by
wheth
e
r
th
e softwa
r
e dev
elope
rs
can cre
a
te kno
w
l
edge,
distri
bu
te or
apply it.
The
process of
kno
w
le
dge
creation is em
b
edde
d in eve
r
y single
step
of process in
softwa
r
e dev
elopme
n
t. The
evaluation i
s
need
ed to view a
kno
w
le
dge
cre
a
tion
in the ea
ch
stage of proce
s
ses. But in t
h
e
softwa
r
e
dev
elopme
n
t software d
e
velop
e
rs
sometim
e
s n
eed
to a
p
p
ly the
kno
w
l
edge
mo
re th
an
cre
a
ted
it. It's ve
ry impo
rt
ant for prog
rammer to
ap
ply the exi
s
ting
kno
w
le
dg
e to the
p
r
og
ram
appli
c
ation
which
certai
nly exist. It see
m
s li
ke
software
develo
p
e
r
s be
com
e
a
s
inte
gral
whi
c
h
adopte
d
pro
c
ess interventi
on, but
highly
stru
cture
d
work. But in a
nother
situati
on they need
to
colla
borate wi
th their pee
rs
across multipl
e
function to
compl
e
te the task.
3. Process I
m
prov
ement Method for
Kno
w
l
e
dge
Work in SMEs
Software d
e
velope
rs h
a
ve been d
e
scri
b
ed as hi
ghly skilled a
nd cre
a
tive employees [12]
A kno
w
le
dge
work
process much
con
c
e
n
trate o
n
th
in
king
a
c
tivities, colla
borative an
d inte
ra
ctive
whi
c
h ma
ke
s it difficult to structu
r
e. Even t
hough the pro
c
e
s
s is com
p
licate
d
, but in certain
circum
stan
ce
s we can
still view a
kno
w
l
edge
wo
rk proce
s
s
ba
se
d on
the
a
ppro
a
ch and meth
od
given ab
ove.
In this p
a
rt
we try to ma
ke
an
ove
r
view
and
sug
g
e
s
tion ho
w to
im
prove
kn
owl
e
dge
work p
r
o
c
e
s
s in SMEs co
mpany.
All the softwa
r
e devel
ope
rs have different skill
and
kno
w
le
dge o
r
kno
w
led
ge
work a
nd
we have de
scrib
e
there are so
me differen
c
e
s
in
term of pro
c
e
ss o
r
ientatio
n. Based on
the
matrix sho
w
e
d
in Figure
3 we can m
a
ke a
co
n
c
lu
sion, software develope
rs involve in the
integratio
n an
collab
o
rative work fo
cusi
n
g
on
hig
h
kn
owle
dge inte
nsity wo
rk. It mean
s software
developers are more on pr
ocesses oriented, but st
ill highly collaborated.
Figure 3. Co
mbination a
p
p
roa
c
h of kno
w
led
ge work
Software dev
elope
rs have to
wo
rk with a
pa
rt
icipativ
e and
ado
pt any proce
s
s
cha
nge
s if
also
have be
en pa
rty to desig
n the
system [7]. In a particula
r view p
a
rtici
pati
v
e cha
nge al
so
typically yields more incre
m
ental ch
ang
e result
s and
contin
uou
s in
order to ma
ke improvem
e
n
ts
over time.
An
exampl
e of t
h
is type
meth
od
woul
d
b
e
agile develo
p
m
ent
mo
del
s whi
c
h ado
pt and
adapte
d
for the develo
p
m
ent pro
c
e
ss o
f
small com
p
anie
s
. It’s less focu
se
d on
the spe
c
ific
steps
to be followe
d in a proce
s
s, but more o
r
iented to
co
mpositio
n of team work an
d highly iterat
ive
workflo
w
[13]. Pair prog
ra
mming is
su
ch a pra
c
ti
ce t
hat hold
s
pro
m
ise for ove
r
comin
g
so
me
of
the chall
enge
s. In pair pro
g
rammi
ng two softwa
r
e de
velopers such prog
ram
m
e
r
s workin
g si
de-
by-sid
e on o
ne co
mpute
r
colla
borating
on the sa
me
desi
gn, algo
ri
thm, coding
or testin
g. This
will create
a
new
enviro
n
m
ent with th
e nee
d to
promote kno
w
l
edge
sh
arin
g
and
colla
borative
kno
w
le
dge di
scovery acro
ss m
u
ltiple function.
3. Rese
arch
Metho
d
The intentio
n of identifying of kno
w
led
g
e
work
s is to b
u
ild a metho
d
for de
scribin
g
way
s
to evolving
proce
s
s
step
s o
f
a kno
w
led
g
e
work
p
r
o
c
e
s
s so that m
a
nage
rs o
r
kno
w
led
ge
wo
rke
r
s
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Knowle
dge
Wo
rk Process: Software
Devel
ope
r
’
s in
Sm
all Medium
… (Mohd Z
a
irol Yu
soff)
559
can
be
dyna
mically o
r
ga
n
i
zed
an
d
coo
r
dinate
this
method to
su
pport va
rio
u
s pro
c
e
s
s a
c
ti
vities
and gui
de to
the pro
c
e
s
s, in an individ
ual ba
si
s, to
advan
ce p
r
o
c
e
ss
step
s towa
rd
s process
compl
e
tion
with highe
r
efficien
cy an
d q
uality. Hen
c
e
,
it must p
r
ov
ide ways to
d
e
scrib
e
vari
o
u
s
pro
c
e
s
s a
c
tivities. Th
e met
hodol
ogy inv
o
lves fo
ur ph
ase
s
’
namely
theo
retical
studie
s
, empi
ri
cal
study, fra
m
e
w
ork evaluati
on a
n
d
valid
ation a
nd
a comp
arative study.
In
th
e
o
retical study
,
a
literature review i
s
con
d
u
c
ted to u
nde
rstand th
e kno
w
ledge
wo
rk p
r
oce
s
s mo
del
prop
osed
by [4]
and [7]
.
Based on thi
s
review the fa
ctors in
kno
w
le
d
ge work pro
c
e
ss a
r
e
identified. In this
resea
r
ch, we
are fo
cu
sing
on the effecti
v
eness an
d e
fficiency of the method. Th
e se
con
d
pha
se
is a
n
em
piri
cal stu
d
y that
focu
se
s
on
colle
cting
dat
a from
software
develo
p
e
r
s i
n
SME by
distrib
u
ting a
set of que
stio
nnaires. Th
e sampl
e
of
this study is 3
0
0
as regi
ste
r
in
SME Compa
n
y
(SEM). Th
e
data will
be
analyzed u
s
i
ng Stru
ctur
al
Equation
M
odelin
g by u
s
ing P
a
rtial
L
east
Square Tech
nique (P
LS).
The third p
h
a
s
e is mo
dele
d
evaluation
and validatio
n. In this pha
se, the mode
l will be
evaluated a
n
d
validated
usin
g a case study an
d
expert revi
ew. The fou
r
th pha
se i
s
a
comp
arative study with oth
e
r wo
rks o
r
m
e
thod
s to evaluate the kn
o
w
led
ge work prod
uctivity.
The
study t
ook pla
c
e
o
v
er a
peri
o
d
of f
our mo
nths and
fo
urteen
semi-stru
c
tured
interviews,
participation in five
meetin
g
s
, an
d
seve
ral di
re
ct ob
servation
s
we
re carried
out.
In
orde
r to
cl
arif
y themes an
d con
c
eption
s
the
mate
ria
l
ha
s b
een
di
scusse
d
with
the
kno
w
led
ge
workers in several inform
al
meetings
and thus ascert
ains reliability
.
3.1. Data
Col
l
ection
We
co
ndu
cte
d
a
cross-se
ctional
stu
d
y of
SMEs i
n
the
softwa
r
e area i
n
Pe
ninsular
Malaysia. Sin
c
e o
u
r
re
sea
r
ch give
s m
o
re att
ention to
the software
develop
ers
unde
r SMEs,
a
good
pra
c
ti
ca
l rea
s
o
n
in th
e ch
oice of t
he Softwa
r
e
Comp
any, th
e re
se
arche
r
followe
d the
EU
definition of
SMEs, exclu
d
ing
comp
ani
es
with le
ss t
han fou
r
em
p
l
oyees. A ran
dom sampl
e
o
f
the SMEs, the software
de
veloper
wa
s
created f
r
om
company d
a
ta
bases in
Mal
a
ysia. The
s
e
will
phon
e and a
s
ked if they
woul
d c
onfi
r
m that their company u
s
ed
web pag
es,
e-mail, or onl
ine
sy
st
em
s f
o
r
r
e
se
ar
ch p
u
rp
ose;
if
so,
t
h
ey
wer
e
t
hen invited to tak
e
part in the s
u
rvey. In eac
h
SMEs comp
a
n
y, the devel
opers
will
su
rvey by us
in
g
an o
n
line
su
rvey by usin
g
que
stionn
aire
s.
All respon
de
nts ho
pefully
will give
cont
ribution
to thi
s
research. E
a
ch
co
mpa
n
y had
at lea
s
t a
web p
r
e
s
en
ce whe
r
e indiv
i
dual cu
stom
ers o
r
com
p
a
n
ies could fin
d
information
about prod
u
c
ts
and services.
3.2. Data
An
aly
s
is
SEM is
cho
s
en a
s
statisti
cal te
ch
nique
be
cau
s
e it
al
lows the
anal
ysis
of all the
factors
simultan
eou
sl
y. The outco
me is si
gnificant dire
ct effects of qu
ality factors to
ward
s knowl
e
d
ge
work p
r
odu
cti
v
ity. Partial least sq
uares a
nalysi
s
(PLS) was
cho
s
en
as the mo
st appro
p
ri
ate tools
in SEM to an
alyze o
u
r m
o
del. PLS is a
confirma
tory
, second
-ge
n
eration
multivariate a
nalysi
s
techni
que tha
t
is well suite
d
for com
p
lex
predi
ctive model
s.
PLS has several advantages that make it
well suit
ed for our research the ability to
handl
e refle
c
tive and formative indicators and
ro
bustn
ess
with re
spe
c
t to
depa
rture from
multivariate
norm
a
lity as well
as th
e
ability to
h
andle th
e m
u
lticolline
a
rity
found i
n
so
me
comp
eten
cy variable
s
of o
u
r mo
del. Furthermo
re, a
s
with multiple
reg
r
e
ssi
on
s, PLS focu
se
s on
the model’
s
a
b
ility to predi
ct rathe
r
than
just ex
plainin
g
the variabili
ty of the depende
nt variab
le,
makin
g
it most useful in situations whe
r
e
the theory
is still being
developed [
6
]. In PLS th
e
predi
ctive abi
lity of const
r
u
c
ts i
s
optimi
z
ed an
d t
he p
e
rform
a
n
c
e o
f
the individu
al scale item
s i
s
repo
rted. In reportin
g
the result
s of these anal
yses, we start with th
e measureme
n
t models.
Formative ite
m
s re
present
measure
s
th
at affe
c
t
th
e
c
o
ns
tr
uc
t u
nde
r
s
t
ud
y. C
h
an
g
e
s
in
the co
nst
r
u
c
t are th
erefo
r
e not expe
ct
ed to cau
s
e
any ch
ang
es in the indi
ca
tors. As a re
sult,
items
within
a fo
rmative
scale
a
r
e
not expe
ct
ed
to correlate
.
Test
s of
converg
ent a
n
d
discrimi
nant
validity based
on the inte
r
correl
ation
s
b
e
twee
n items are the
r
efo
r
e
not releva
nt for
evaluating th
e psychom
etric p
r
op
ertie
s
of formativ
e items. Inste
ad, item wei
ghts a
r
e u
s
e
d
to
indicate ho
w relevant ea
ch
item is in measu
r
ing its lat
ent con
s
tru
c
t.
The
refle
c
tive item
s a
r
e
believe
d to
be
cau
s
ed
by the l
a
tent
co
nst
r
u
c
ts t
hey a
r
e
intende
d to
measure. Th
e Interco
rrel
a
tions
bet
we
e
n
the item
s
are th
erefore
expe
cted. T
he
psychomet
ric prope
rties
of the reflect
i
ve item
s we
re examine
d
by analyzin
g their internal
con
s
i
s
ten
c
y in terms
of their
conve
r
g
ent and
di
scrimina
n
t validity. Convergent validity wa
s
estimated
ba
sed
on the it
em loadi
ng
s, and a l
oadi
ng of ab
ove
0.70 is
re
co
mmend
ed a
s
this
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 3, March 2
015 : 555 – 5
6
0
560
indicated that
at least half of the variance in
ea
ch i
t
em could b
e
accounte
d
for by the latent
c
o
ns
tr
uc
t.
4. Conclusio
n
Firm
s mu
st b
e
tter ma
nag
e
two
of thei
r
most
pre
c
io
u
s
a
s
set
kno
w
ledge
an
d th
e pe
ople
who
create
a
nd po
sse
s
s i
t. Firms
attempting to m
a
ke th
eir
kn
o
w
led
ge
wo
rk pro
c
e
s
se
s
more
efficient a
nd
effective face
a
choi
ce. T
hey can a
d
o
p
t tran
sa
ctio
n ap
pro
a
che
s
for kno
w
le
dge
work im
prove
m
ent that ha
ve been
emp
l
oyed fo
r th
e
admini
s
trativ
e and
ope
rat
i
onal
work.
Or
they ca
n em
ploy mo
re
co
llaborat
ive ap
proa
ch
es tha
t
rely on
pai
r pro
g
rammin
g
to d
e
si
gn a
n
d
evaluate thei
r own
activities. In
most
cases, however, we believe
t
hat organi
zati
ons
will benefit
by cho
o
si
ng
an inte
rme
d
iate pa
rticip
ative
cou
r
se
betwe
en th
e two
extre
m
es.
Usi
ng
the
strategi
es
we
have di
scussed, co
m
pani
e
s
can sele
ct method
s
an
d ta
ctics that re
flect the type
o
f
kno
w
le
dge work,
they are add
re
ssi
ng,
their
organi
zational
cultu
r
e, an
d the b
u
sin
e
ss
requi
rem
ents for the
cha
n
ge p
r
oje
c
t. Of cou
r
se, imp
r
ovements to
kno
w
le
dge
work are only
one
effort in a bro
ad portfolio of
improveme
n
t and ch
ang
e initiatives that manag
ers mu
st integrate.
Referen
ces
[1]
PF
Drucker. Mana
geme
n
t.
Califor
nia Ma
na
ge
me
nt Revie
w
. 1999; 41(2):
79–9
4.
[2]
JP W
a
re, CE
Grantham. Kno
w
l
e
d
ge W
o
rk a
nd Kno
w
l
e
dg
e W
o
rkers.
Desi
gn
. 200
7; 2(10)
.
[3]
R Maier. Modeling K
n
o
w
ledg
e
W
o
rk for the
Desig
n
of Kno
w
l
e
d
ge Infrastr
uctures.
Univ
er
sal Co
mp
uter
Scienc
e
. 200
5; 11(4): 429
–4
5
1
.
[4]
BS Sebasti
an
Eschen
bac
h, Doris Ri
edl. Kn
o
w
l
e
d
ge W
o
rk P
r
oductiv
i
t
y
: W
h
ere to Start.
Practical Aspet
of Know
led
ge
Mana
ge
me
n 6
conf, PAKM 2006
. 200
6: 49–
6
0
.
[5]
T
Davenp
ort, M Beers. Improvin
g Kno
w
l
e
d
ge W
o
rk Proce
sses. 1995; 1
–
13.
[6]
BD Adler PS,
McGarry
FE, Talbot WB. Enabling pr
ocess d
i
scipl
in
e: lesso
ns from Comp
uter Scie
nce
s
Corp
oratio
n on
Capa
bil
i
t
y
Mat
u
rit
y
Mo
de
l Lev
el 5. 200
3; 26.
[7]
T
H
Davenport
.
Process Manag
ement for
Kno
w
l
edg
e W
o
rk.
Handb
ook on Bus
i
n
e
ss Proces
s
Mana
ge
me
nt 1
,
Internation
a
l
Han
dbo
oks on
Information Sy
stems
. 20
10; 2
(
200
5): 17–
36.
[8]
T
R
Eikebrokk, DH Olsen. An empir
i
cal i
n
vesti
gati
on of
compete
n
c
y
f
a
ct
ors affecting
e-busi
ness
success in Eur
opean SMEs.
Information & Managem
e
nt
. 20
07; 44(4): 3
64–
383.
[9]
AS Saleh, NO
Ndubisi. An
Evalua
tion
of SME Devel
o
p
m
ent in Mal
a
ysia.
Internati
o
n
a
l Rev
i
ew
of
Business Research Papers
. 2006; 2(1): 1
–14
.
[10]
W
A
. Ho
w
W
e
Appro
a
ch So
ft
w
a
re
Devel
opm
ent. 201
3
.
http://
w
w
w
.
se
i.cmu.edu/so
luti
ons/soft
w
are
d
e
v
/? lo
cati
on=
main-
nav&so
u
rc
e=
652
58
8. [Accessed: 2
2
-
Au
g
-
20
14
].
[11]
GC Mohanta,
V Kanna
n, KP
T
hoo
yama
ni. Stra
tegies
F
o
r Improving
Productivit
y
Of Kno
w
l
e
d
g
e
Workers - A
n
Overvi
e
w
.
200
6. Ava
ila
b
l
e:
http://
www.strengthbas
ed
stra
tegies.com/
PAPERS/1
0
Moha
ntaF
orma
tted.pdf.
[12]
Marks, D Sch
o
lari
os. Ch
ore
ogra
phi
ng
a S
ystem: Skill a
n
d
Emplo
y
a
b
i
lit
y i
n
Soft
w
a
re Wo
rk.
Econo
mi
c
and Ind
u
stria
l
De
mocracy
. 2
0
08; 29(1): 9
6–1
24.
[13]
RK Kavith
a, M
S
Irfan Ahme
d.
A Kno
w
l
e
dg
e
M
ana
geme
n
t F
r
ame
w
ork
for
Agil
e Soft
w
a
r
e
Devel
opm
ent
T
eams.
Interna
t
iona
l Conf
eren
ce on Process
Auto
matio
n
, Control a
nd C
o
mp
utin
g
. 201
1: 1–5.
Evaluation Warning : The document was created with Spire.PDF for Python.