Problem Decomposition and Cooperative Learning: an Exploratory Study in Enhancing Software Engineering Projects



Similar documents
3.1. Overview Serial Devices to Ethernet Gateway

Improper Integrals. Dr. Philippe B. laval Kennesaw State University. September 19, f (x) dx over a finite interval [a, b].

Greening Supply Chain Management

Electricity Load Forecasting Science and Practices F. Elakrmi 1,*, N. Abu Shikhah 1

Chapter 13. Network Flow III Applications Edge disjoint paths Edge-disjoint paths in a directed graphs

STRATEGIC PLANNING COMMITTEE Wednesday, February 17, 2010

Circle Geometry (Part 3)

4.8. Solar cells The solar spectrum

USE OF EDUCATION TECHNOLOGY IN ENGLISH CLASSES

Example What is the minimum bandwidth for transmitting data at a rate of 33.6 kbps without ISI?

Solution y y = 0 = = 0.250

Mr. Kepple. Motion at Constant Acceleration 1D Kinematics HW#5. Name: Date: Period: (b) Distance traveled. (a) Acceleration.

Process Modeling for Object Oriented Analysis using BORM Object Behavioral Analysis.

Chapter 7. Response of First-Order RL and RC Circuits

GoRA. For more information on genetics and on Rheumatoid Arthritis: Genetics of Rheumatoid Arthritis. Published work referred to in the results:

Principal components of stock market dynamics. Methodology and applications in brief (to be updated ) Andrei Bouzaev, bouzaev@ya.

Performance Center Overview. Performance Center Overview 1

OPTIMAL BATCH QUANTITY MODELS FOR A LEAN PRODUCTION SYSTEM WITH REWORK AND SCRAP. A Thesis

Phys222 W12 Quiz 2: Chapters 23, 24. Name: = 80 nc, and q = 30 nc in the figure, what is the magnitude of the total electric force on q?

Better Bounds for Online Load Balancing on Unrelated Machines

Heat demand forecasting for concrete district heating system

Dynamic Magnification Factor of SDOF Oscillators under. Harmonic Loading

Acceleration Lab Teacher s Guide

2. The econometric model

Fuzzy and design optimization for an intelligent additional track mechanism of swamp peat vehicle

Age Biased Technical and Organisational Change, Training and Employment Prospects of Older Workers

The Grantor Retained Annuity Trust (GRAT)

Permutations and Combinations

cooking trajectory boiling water B (t) microwave time t (mins)

METHODS OF LEARNING IN STATISTICAL EDUCATION: A RANDOMIZED TRIAL OF PUBLIC HEALTH GRADUATE STUDENTS 3

OPTIMIZING PRODUCTION POLICIES FOR FLEXIBLE MANUFACTURING SYSTEM WITH NON-LINEAR HOLDING COST

Department of Health & Human Services (DHHS) Centers for Medicare & Medicaid Services (CMS) Transmittal 1151 Date: November 16, 2012

Detecting Network Intrusions via Sampling : A Game Theoretic Approach

A Dynamic Model of Health Insurance Choices and Health Care Consumption 1. Jian Ni Johns Hopkins University

Reuse-Based Test Traceability: Automatic Linking of Test Cases and Requirements

Rate-Adjustment Algorithm for Aggregate TCP Congestion-Control

Morningstar Investor Return

µ r of the ferrite amounts to It should be noted that the magnetic length of the + δ

Term-based composition of security protocols

This work is licensed under a Licença Creative Commons Attribution 3.0.

The Greek financial crisis: growing imbalances and sovereign spreads. Heather D. Gibson, Stephan G. Hall and George S. Tavlas

TSG-RAN Working Group 1 (Radio Layer 1) meeting #3 Nynashamn, Sweden 22 nd 26 th March 1999

Identifying Merger Unilateral Effects: HHI or Simulation?

INTRODUCTION TO FORECASTING

Human Body Tracking with Auxiliary Measurements

WHAT ARE OPTION CONTRACTS?

CALCULATION OF OMX TALLINN

RC (Resistor-Capacitor) Circuits. AP Physics C

One Practical Algorithm for Both Stochastic and Adversarial Bandits

Usefulness of the Forward Curve in Forecasting Oil Prices

Individual Health Insurance April 30, 2008 Pages

CHARGE AND DISCHARGE OF A CAPACITOR

Chapter 6: Business Valuation (Income Approach)

Lecture 15 Isolated DC-DC converters

Physical Topology Discovery for Large Multi-Subnet Networks

Policies & Procedures. I.D. Number: 1071

Numerical Simulation and Experimental Verification of Air Flow through a Heated Pipe

Optimal Contracts in a Continuous-Time Delegated Portfolio Management Problem

Mathematics in Pharmacokinetics What and Why (A second attempt to make it clearer)

Optimal Path Routing in Single and Multiple Clock Domain Systems

ON A FAIR VALUE MODEL FOR PARTICIPATING LIFE INSURANCE POLICIES

CULTURAL TRANSMISSION AND THE EVOLUTION OF TRUST AND RECIPROCITY IN THE LABOUR MARKET. Gonzalo Olcina and Vicente Calabuig

Advise on the development of a Learning Technologies Strategy at the Leopold-Franzens-Universität Innsbruck

CLASSIFICATION OF REINSURANCE IN LIFE INSURANCE

How Much Can Taxes Help Selfish Routing?

Project Management Basics

Influence of Network Load on the Performance of Opportunistic Scanning

Fortified financial forecasting models: non-linear searching approaches

Resource allocation in multi-server dynamic PERT networks using multi-objective programming and Markov process.

11/6/2013. Chapter 14: Dynamic AD-AS. Introduction. Introduction. Keeping track of time. The model s elements

PROFITS AND POSITION CONTROL: A WEEK OF FX DEALING

Chapter 1.6 Financial Management

Calculation of variable annuity market sensitivities using a pathwise methodology

Chapter 8 Student Lecture Notes 8-1

Information Technology Investment and Adoption: A Rational Expectations Perspective

A Note on Using the Svensson procedure to estimate the risk free rate in corporate valuation

Course Outline. Course Coordinator: Dr. Tanu Sharma Assistant Professor Dept. of humanities and Social Sciences

Capacitors and inductors

Signal Rectification

Chapter 8: Regression with Lagged Explanatory Variables

Vector Autoregressions (VARs): Operational Perspectives

SELF-EVALUATION FOR VIDEO TRACKING SYSTEMS

Empirical heuristics for improving Intermittent Demand Forecasting

MULTIPLE LIFE INSURANCE PENSION CALCULATION *

9. Capacitor and Resistor Circuits

Transcription:

Problem Decomoiion nd Cooerive Lerning: n Exlorory Sudy in Enhncing Sofwre Engineering Projec Tie Hui Hui nd Irfn Nufl Umr Abrc Comleing workble ofwre wihin iuled imefrme h lwy been chllenge for comuing uden. College educor eching rogrmming coure re fcing difficulie o guide nd monior he uden from he r of he ofwre develomen cycle unil he comleion ge. Likewie, uden find he ofwre develomen roce much more ediou, comliced nd fruring. Problem decomoiion chunking echnique i ued brc conce o brek he rogrmming cenrio ino mller mngeble chunk in relion o fulfilling he yem requiremen. Thi er reen n overview of he difficulie encounered, nd udy on roblem decomoiion echnique wih cooerive lerning o reolve uch fflicion. A ol of 44 econd yer comuing uden were rndomly igned eiher o grou h received combinion of roblem decomoiion nd cooerive lerning (DCL), or o noher grou which received cooerive lerning mehod (CL) in hi 28-week remen. The ricin worked in grou of four, wih mix of uden wih high nd low elf-reguled lerning level. The o-emen w dminiered o meure heir ofwre develomen erformnce h w bed on elf-rogrmming ril, rojec erformnce nd number of rool reviion. The reul reveled h uden in he DCL grou erformed ignificnly beer hn hoe in he CL grou in boh he elf-rogrmming erformnce nd ofwre rojec develomen. Alo, he DCL grou mde fewer mendmen o he rojec rool hn heir counerr. Thu, he roblem decomoiion echnique incorored ino he roblem nd nlyi ge wihin he yem develomen cycle hould be conidered n lernive regy for effecive wy of eching, lerning nd comleing ofwre engineering rojec. Index Term Chunking echnique, cooerive lerning, roblem decomoiion, ofwre develomen rojec. I. INTRODUCTION Wih he dvncemen of echnology, orgnizion re connly incororing uilizing he Informion yem (IS) in heir dily buine rocee in order o chieve ucce, y comeiive, incree rofi nd mrke hre. Wih hi, he execion on informion communicion echnology (ICT) rojec i increing. More ohiiced nd dvnced funcionliie wih uer friendline feure re in demnd. In ddiion o echnologicl chllenge, buine need, uer requiremen, orgnizionl execion iue Mnucri received December 14, 2011; revied Februry 3, 2012 H. H. Tie i wih he Cenre for Pogrdue Sudie, SEGi College Penng, Mlyi, 43 Green Hll, 10200 Penng, Mlyi (e-mil: hhie@egi.edu.my). I. N. Umr i wih he Cenre for Inrucionl Technology & Mulimedi, Univeriy Science Mlyi, 11800 USM, Penng, Mlyi (e-mil: irfn@um.my). rie, exerience nd exerie of IS rojec em re conribuing filure o n exenive number of develomen iniiive [1]. The develomen of IS rojec require mngeril, coordinion, roblem olving, criicl hinking nd nlyicl kill. Thee re he key kill h need o be develoed nd gred in udying rogrmming coure. Converely, rogrmming demnd comlex cogniive kill h uden find hem oo comliced o undernd, inerre nd erform. Likewie, educor involved in eching he ofwre develomen roce o comuing uden re coninully fcing differen chllenge in delivering he yem conce, rogrmming civiie, develomen ool nd echnique; nd uden re hving difficulie o comrehend hem. During he lerning roce, uden re o undernd boh bic requiremen uch (i) he uer requiremen nd (ii) he funcionliie of he ofwre o erform hoe requiremen. Once hee need hve been idenified, deermining he coe nd eblihing buine requiremen re crried ou before he ge of ofwre deign, imlemenion nd minennce. Thi ofwre develomen roce involved roblem olving, lnning, criicl hinking, reoning nd ocil inercion h comuing uden find hee combinion of kill oo difficul o cquire. [2] Indiced h undernding he ofwre develomen roce, nd lerning how, why nd when o cree deliverble ofwre re hrough working wih rel yem develomen rojec. The enggemen in develoing ofwre rojec could foer growh in roblem olving, criicl hinking nd inereronl kill. Bed on roblem decomoiion echnique, uden re o look ll he oible requiremen ed by he uer nd hen o nlyze hem. In hi ce, roblem decomoiion chunking mehod i yemic wy of decomoing he roblem cenrio ino mller mngeble ub-roblem [3]. Wih i emhi on roblem rher hn oluion, hi roch llow he uden o undernd he idelized roblem nd hen o link hem wih heir ecific domin knowledge in order o drive he requiremen nd nlyi engineering roce. Menwhile, [3] ed h decomoiion i wy of nlyzing nd mnging comlexiy. I i roblem olving echnique h o be lied in he requiremen nd nlyi he for idenifying he coe of he yem. Subequenly, i i incorored ino he deign he o o deermine he logicl rucure nd funcionliy for building he yem. Thi roch uor boh he high-level brc view of requiremen nd he lower-level deiled view of rocee. Thu, could hi roch i he uden in idenifying he novel cenrio nd hen in decomoing he cenrio ino ub-cenrio by looking ll he oible develomen 105

conrin? Alo, will hi roch hel he uden o urn roblem ino oluion by creing new ide? II. PROBLEM STATEMENT Sofwre rojec develomen i no ju bou idenifying he bic requiremen ed in he given novel cenrio. I i lo no merely bou lerning ome rogrmming lnguge ynx. Thi develomen roce require he biliie o hndle he ofwre engineering ge nd o urn he roblem cenrio ino vlid workble licion. Emiricl reerch finding dicloed h undergrdue comuing uden wih limied heoreicl rogrmming knowledge, roblem olving nd rcicl kill re fcing chllenge in eblihing rojec objecive, defining ofwre funcionliie nd comleing he yem (ofwre) wihin he given imeline [4]. For ucceful ofwre develomen, hee knowledge nd rcicl kill re needed nd o be lied in ll he of develomen cycle which uden eem o lck. Likewie, he iniil finding in Tble I revel h he roblem fced by undergrdue comuing uden in lerning rogrmming nd develoing yem re in ccordnce wih lierure concerning roblem in comuing rogrmming. The reliminry iicl reul indiced h 20 ercen of he uden rojec coe hd been revied fer heir rojec rool ubmiion. Wih hi, only few uden hd comleed he yem on ime nd chieved he rojec objecive. However, mjoriy of hem were ill working on i unil he l minue wih incomlee funcionliy. Thu, only limied eing w conduced. If rojec durion i no fcor, hen comrehenive eing regy could be imlemened nd undicovered error my erh be deeced. Thi will omehow incree he robune of he imlemened ofwre. In line wih he finding reored by [5] on ofwre rojec filure in he ofwre indury h chievble rojec objecive, execuble funcionliie nd fulfilling he uer requiremen re he key o IT rojec ucce. TABLE I: PROBLEM IN DEVELOPING SOFTWARE ENGINEERING Requiremen nd nlyi ge coe of he yem re oo lrge or oo viionry lck of clerer ofwre coe uncerin wih he rojec objecive lck of keholder riciion in he deign lierure review i oo brief or incomlee ill-defined funcionl module oor lnning nd ime mngemen incomlee eing limiion on rogrmming lnguge knowledge incomibiliy beween echnology nd rogrmming lnguge lck of roblem olving, nlyicl nd reoning kill lck of rogrmming conceul undernding Progrmme develomen begin wih roblem idenificion. Once hee roblem requiremen hve been idenified, he eleced rogrmming lnguge will be lied in he imlemenion ge. Progrmming lnguge uch C#, JAVA, VB.ne nd ASP re commonly ued in develoing he rogrmming oluion. Plnning, logicl reoning, roblem olving nd criicl hinking re he kill required in he roce of lerning nd foremo during ofwre develomen [6]. In order o chieve higher ucce re in ofwre develomen, hee knowledge nd kill re crucil nd needed o be lied hroughou he ofwre develomen he, which he uden re fcing difficuly o cquire hem. Uing he combinion of rogrmming nd roblem decomoiion echnique effecively in coure delivery, [6] reveled h logicl reoning nd roblem olving kill cn be culived while lerning o develo vlid workble licion. The reliminry queion ked rior o he ofwre develomen roce re bed on he nine oic re uch (i) buine coe of orgnizion, (ii) buine requiremen, (iii) uer requiremen, (iv) yem requiremen, (v) rojec coe, (vi) rojec imeline reure, (vii) echnology comibiliy, (viii) funcionliie / echnicl comlexiy, nd (ix) rogrmming lnguge. In ddiion o he informion obined, conrin encounered by uden hndling ofwre develomen civiie re (i) uncler rojec coe uer requiremen hve been frequenly revied, (ii) le uer involvemen in he deigning ge, (iii) lck of mngemen involvemen, nd (iv) dily roceing e hve been ill-defined nd ofen being ignored by ff oering hem, nd (v) nxiey in hndling ofwre rojec. Wih lo of uncerinie he reliminry ge, here i no urrie h he rojec coe, requiremen nd conrin (echnicl nd rogrmming ool) re frequenly revied hroughou he develoing cycle. In urn, he deign nd coding civiie only ke lce owrd he end of he develomen chedule ln. Thi give he uden inufficien ime o imlemen he ecificion ed in he ofwre rool; nd my no hve ufficien durion for evluing he roduc [7]. The develomen of IS rojec require n effecive riciion of keholder (uer), comrehenive reviewing of he curren k flow, undernding of uer requiremen nd roblem encounered. Deie hi iniil inveigion nd involvemen, ofwre develomen conrin in relion o exernl fcor nd develomen fcor hould be emerged. Thee exernl fcor re ill lrgely conrolled by he moun of informion obined from he keholder. I h been een he reonibiliy of he develoer o obin he required informion. The develomen fcor re generlly conceulized n licion roce o incorore boh knowledge nd rogrmming kill well iude in he rojec develomen. By reviewing he develomen fcor, hi llow he uden o e heir rengh, kill nd weknee. Thu, hey will ue hi informion o mke ubequen judgmen on wheher ny ddiionl deil, furher inveigion or ny refinemen in he rojec rool i required. The conrin fcor model (CFM) reened in Fig. 1 encourged he uden o hve helicoer view of he rojec ec nd hen o zoom ino ll oible deil. Wih hi, uden will hve he ooruniie o conruc reled rogrmming civiie h re bed on he nlyi finding. From he informion ghered, uden will cively engge in evluing heir biliie o develo hoe required kill wih n umion h hey inend o underke hi rojec for ucce. I become he uden 106

Inernionl Journl of Socil Science nd Humniy, Vol. 2, No. 2, Mrch 2012 com leion. In hhi ec, Fig. 2 reen robleem deco omoiion oofwre enginneering (PDS SE) model hh ynh heize in-deh evluionn in he develo omen he. reeonibiliy o mnge heeir lerning nd ke invrribly coomlex nd chhllenging roogrmming k. Fig. 1. Conrin Fcor Moddel (CFM) in ofw wre develomenn. Fig. 2. Problem Decoomoiion Sofw wre Engineering g (PDSE) Model in i roduc develomen roce. hee conrrin indiceed in he Conrin By looking Fcor Model (C CFM) he erly e ge of he develomeen, i w hel he uden will o elec he uible develomen m mehodology, roche nnd deign for heir new rojec. A Alo, i encourge he uden o nlyzee he ucce re of hhi rojec well hhe oenil benefi of uch im mlemenionn. Thi furheer i in rengheningg he unndernding of o he new knnowledge cquuired. Thee enire e cciviie indicce cive, reflecive, r ndd exlororyy nd exxerimenory rocee of he lerning roch. A. Problem Decomoiio A D on in Sofwree Develomen Sge Similr o ny rogrm m olving k, develoiing rrogrm involvve e. The five min ge in he of fwre deevelomen roce beginn wih robleem nlyi, hen rroceed o lnnning he oluion, coding, eeing nd endd wih hhe documenion. I requiree he uden o undernnd he rrogrmming roce nd o incorore hhee ge ino he yyem devellomen meehodologie. Likewie, he keholder reequiremen nd yem objjecive re crrucil o IS develom men rojec. A A buine yyem oo he ucce of nnd environmen undergo cooninul chnnge, he IS hhould keee ce. Som mehow hi reevoluion h conribued o he coonn chngge in buinne nd ueer requirem men. H However, i i chllenge o deerminee he coe of o he yyem nd o eblih e he oofwre requirremen ueer do noo lwy connvey heir neeed nd execion. Thi could c bee he fc h hey my no fully undernd reciely wh rre needed in hhe fir lce. Wih W uch unccerinie, uuden finnd he develomen roce ime conum ming nd edioou hhey need longeer ime o hvee olid undernding of he uer reequiremen nnd he yem m funcionliiie. Adding o he chhllenge i whhen he uden dicover hee incomibilliy of hhe echnologyy device wih rogrmmingg lnguge he cooding ge. Alo, he lieerure review w of uch iimilr rroduc w uully no donne ye he beginning b of he IS deevelomen roce unil when necery. By henn he rrojec coe h been firmed u nd finll uer requirem men rre eblihedd nd ny modificion orr ddiion oo he rrojec rool will conribbue o he deely in he rrojec 107 Th he fir he of Fig. 2 i bbed on he fir nd ecoond gee in he yyem develoomen life cycle (SDCL L). Sofw wre conrin, lierurre review off rocee nd echn nologie re illured in he roblem m decomoiiion he. Thi indiice he k h u uden need o inveige, reviiew nd unndernd before definiing uiremen. Theee re he crrucil e off idenifying he requ civ viie in rojeec develom men roce. Deil nlyyi hrou ugh roblem m decomoiion echniqu ue llow he udeen o exminne he uer reqquiremen more m criiclly by chun nking hem inno mngeblle ecion. En ngging wih he civ viy require hhe uden o hve deeer undernding u g of requ uiremen deeermined by hhe uer well o revieew heirr own rogrm mming kill nnd cbiliie in conruciing nd comleing c hhe rooed yyem. A he econdd he (reqquiremen n nd deign), he udeen formule oluion bed on he gherred inforrmion. The roblem form mulion i derived from he geneerl oluion nlyi nd he comrio on of he yeem objeccive gin he requirem men. I hel he uden o nly yze he ideniffied roblem nd uer eccific echniquue, nd hen o reenn he roblem m oluion, nd d hu o vlide he oluion. Thi llow he uden o deermine whehher heee defined robblem nd olluion could be b imlemened uccefully or if ome form off modificion i necery. A w heir exiiing hi he, he uden migh even review rog grm kill ndd knowledge w which in urn could influennce IS develomen d rojec uccee. In he deevelomen nd imllemenion he, he uuden re o ly he moo uible develom men ool nd ecificc rogrmmiing lngu uge while convering he oluio on deign inno execcuble oluioon (yem). T Teing i crriied ou o verify for yncicl errror, nd hen o enure fully f oerionnl yeem h he ouu i in ccordncce wih ueer exeecion defined in he iniil rool. Buine nd uer requiremen evolve hrougghou he yeem uge, whiich mkee minenncee nd uor n ongoing roce. A ucch, he develomen roce bbegin when n chnge in exeecion ree eblihedd. In hi ce, yeem deveelomen i ierive rocce where new n buinee

rocee re idenified wih incree in dvnce funcionliie nd comlexiie of rocee. In fc he rojec ucce deend on in-deh evluion he roblem decomoiion he in he PDSE model h derive reliic execion in order o ccomlih he fully oerionl yem wihin he lloced imeline. Thu, i could omehow i he uden in hndling he yem develomen rojec o rnform heir erformnce. III. RESEARCH QUESTIONS In hi udy, hree rimry queion hve been formuled o ddre he reerch oucome: 1) I here ny ignificn difference in he number of reviion mde o he rojec rool beween uden ugh in he DCL nd CL inrucionl mehod? 2) I here ny ignificn difference in elf-rogrmming erformnce beween uden ugh in he DCL nd CL inrucionl mehod? 3) I here ny ignificn difference in ofwre rojec erformnce beween uden ugh in he DCL nd DCL inrucionl mehod? IV. RESEARCH METHODOLOGY The uroe of hi udy i o inveige he effec of roblem decomoiion wih cooerive lerning (DCL) nd cooerive lerning only (CL) inrucionl mehod on he undergrdue comuing uden in develoing finl roduc (ofwre). I im o exmine wheher he combinion of roblem decomoiion echnique nd cooerive lerning n effecive lernive oluion in rogrmming educion during ofwre develomen. A. Reerch Deign A qui-exerimenl deign w ued o meure he effec of DCL nd CL inrucionl mehod on he uden rogrmming develomen erformnce. The emen of uden rogrmming develomen erformnce w cegorized ino hree comonen h include (i) number of reviion o rojec rool, (2) elf-rogrmming erformnce, nd (3) ofwre rojec erformnce. A uch, heir rogrmming develomen erformnce w meured bed on he core obined from he number of revied rojec rool, he elf-rogrmming ril (SPA) queionnire, nd he ofwre rojec evluion (SPE) mrk hee. In hi ilo udy, grou of 44 uden from he econd yer undergrdue comuing coure were involved. Thee wo cle, ll inc grou, were rndomly igned o he wo remen grou. The exerimenl grou (20 uden) received he DCL remen while he conrol grou (24 uden) received he CL remen. Thoe in he DCL grou received he combinion of roblem decomoiion nd cooerive lerning regy in cquiring he bic ofwre develomen mehodologie. In he CL grou, hey were only exoed o cooerive lerning regy in develoing he grou rojec. To undernd he conce of ofwre engineering, boh grou were ugh o nlyze he ye of ofwre roce echnique (e.g.: irl mehodology, rid rooying model, incremenl model, objec oriened rogrmming, nd Agile ofwre develomen). For hi udy, he coure comried lecure nd rcicl eion. In clroom, he lecurer reviewed nd exloied he vriou ye of yem develomen rocee nd ool. During rcicl eion, he uden lied uible rogrmming lnguge in coding. B. Reerch Inrumen The uden rogrmming erformnce w meured bed on he core obined from he number of revied rojec rool, he elf-rogrmming ril (SPA) queionnire, nd ofwre rojec evluion (SPE) mrk hee. The emen of heir develomen erformnce include: (i) number of reviion o rojec rool, (2) elf-rogrmming erformnce, nd (3) ofwre rojec erformnce. The SPA queionnire w ued o meure he uden rogrmming knowledge nd kill develoed hroughou he coure emeer during he em rojec civiie. The number of reviion mde o he rojec rool w meured o deermine heir undernding on he iniil roblem, uer requiremen nd coe of he yem. Menwhile, he Moived Sregie for Lerning Queionnire (ded from [8]) coni of 23 iem w ued o idenify he uden elf-reguled lerning level (high or low SRL) rior o he remen. In hi udy, he MSLQ men core of he mle w 3.50. Suden who cored 3.50 nd bove he grou men were cegorized high SRL nd hoe who cored below 3.50 were clified low SRL. Immediely fer he remen, he SPE w dminiered o he ricin during heir ofwre reenion eion o meure he uden regic/condiionl knowledge nd rogrmming kill. Prior o i, e of relibiliy e were conduced on he inrumen ued in order o deermine he Cronbch Alh relibiliy coefficien. The relibiliy vlue of inrumen were: (i) 0.90 for he SPE, nd (ii) 0.92 for he SPA. C. The Coure Meril Toic imred o he uden during he 28 week of remen (wo emeer) w ocied wih he ofwre rojec mngemen nd develomen. In he fir emeer (foureen week), oic reled o he bic ofwre rojec mngemen were dicued wih uden. In ddiion o hee oic, he roblem olving nlyi hrough roblem decomoiion i covered. For he econd emeer, uden were minly concenring on develoing he ofwre nd he role of lecurer h been wiched from lecuring o fciliing nd uoring. Thi i o enure h he uden chieve ll objecive e in he rojec rool nd comlee he rojec wihin iuled imefrme. D. D Collecion Procedure The econd yer emeer one comuing uden in he wo inc cle were involved in hi 28 week udy. They were rndomly igned o he wo remen grou (DCL nd CL). In hi udy, he coure comried lecure nd rcicl eion. To undernd ofwre engineering conce, he uden were exoed o he bic ofwre develomen mehodologie (SDM) in he heoreicl cl eion. During he rcicl eion, hey were igned o develo ofwre by uing he idenified rogrmming lnguge nd he eleced ofwre roce echnique while rogreed hroughou he develomen ge. 108

The uden in boh exerimenl grou worked cooerively in em of four member. Ech em conied of wo high nd wo low SRL uden. During he econd week of remen, he uden hd o ubmi heir rojec rool nd he number of chnge mde fer he fir ubmiion w recorded. They were erienly required o cooere on every rogrmming civiy hroughou he develomen ge. Once he coding h been comleed by he individul em member, he inegrion of ll module w crried ou o enure he comibiliie of he funcion wihin he roduc (ofwre). In he econd emeer, hey concenred on develoing he ofwre nd he role of lecurer h been wiched o fcilior mode. I w o enure h he uden chieve ll objecive nd comlee he rojec wihin iuled imefrme. In Week 28, boh grou were igned reenion lo o demonre heir ofwre. Immediely fer he remen, ech grou w given n hour o demonre heir ofwre. Immediely fer heir ofwre reenion, he uden were ked o comlee he SPA queionnire o meure heir elf-rogreion on he rogrmming knowledge nd kill gined hroughou he develomen cycle. Menwhile, he SPE inrumen w ued o meure he uden logicl undernding of he ofwre code, regic/condiionl knowledge nd rogrmming kill. Thee wo inrumen were ued o e he uden rogrmming erformnce in erm of coding kill, rogrmming knowledge nd ofwre roce echnique. Alo, he number of reviion mde o he rojec rool w meured o deermine heir undernding on he iniil roblem, uer requiremen nd coe of he yem. V. RESEARCH FINDINGS The SPSS 17.0 for Window w ued in hi udy. The MANOVA iicl echnique w lied o e he reerch hyohee on he hree deenden vrible: (i) number of reviion o rool, (ii) elf-rogrmming erformnce nd (iii) ofwre develomen erformnce. The nlye reul re hown in Tble II nd Tble III. The MANOVA reul clerly reveled ignificn difference in ll he hree deenden vrible ed in he reerch hyohee. TABLE II: MANOVA FOR THE SCORES OF DEPENDENT VARIABLES BETWEEN THE TWO EXPERIMENTAL GROUPS Deenden vrible df Men qure F Sig. Number of revied rool 1 19.88 13.62 0.001* Self-rogrmming erformnce 1 2.21 13.10 0.001* Sofwre develomen erformnce 1 686.59 14.68 0.000* *ignificn 0.05 level Hyohei 1: There w no ignificn difference in he number of reviion mde o he rojec rool beween he DCL nd CL grou. In hi udy, he MANOVA nlyi reul indiced ignificn difference in he number of reviion mde o he rool beween he DCL nd he CL grou (F: 13.62; : 0.001), wih he DCL grou erformed ignificnly le mendmen o he rojec coe in he rool hn he CL grou ( X DCL: 1.40; X CL: 2.75). Therefore, hi finding h rejeced he fir hyohei. TABLE III: DESCRIPTIVE STATISTICS OF THE DEPENDENT VARIABLES FOR THE TWO TREATMENT GROUPS Men Deenden vrible Grou Men SD N Diff Number of revied rool Self-rogrmming erformnce Sofwre develomen erformnce DCL 1.40 1.00 20 1.35 CL 2.75 1.36 24 TOTAL 2.14 1.37 48 DCL 3.71 0.34 20 0.45 CL 3.26 0.46 24 TOTAL 3.46 0.46 48 DCL 77.60 6.50 20 7.93 CL 69.67 7.12 24 TOTAL 73.27 7.85 48 Hyohei 2: There w no ignificn difference in elf-rogrmming erformnce beween he DCL nd CL grou. The reul indiced h here w ignificn difference in elf-rogrmming erformnce beween boh grou (F: 13.10; : 0.001). The uden who received DCL remen ignificnly ouerformed hoe of CL remen ( X DCL: 3.71; X CL: 3.26) in elf rogrmming knowledge emen. Thu, he econd hyohei h been rejeced. Hyohei 3: There w no ignificn difference in he rojec erformnce beween he DCL nd CL grou. The finding indiced ignificn difference in he ofwre develomen erformnce beween he DCL grou nd he CL grou, wih he former erforming ignificnly beer hn he ler grou (F: 14.68; : 0.000; X DCL: 77.60; X CL: 69.67). Thu, hi finding h lo rejeced he hird hyohei. VI. DISCUSSION Thi udy im o inveige he effecivene of roblem decomoiion echnique ued in lerning ofwre develomen roce on he uden rogrmming develomen erformnce. Thee econd yer comuing uden from wo inc grou were rndomly igned o eiher one of he inrucionl mehod (DCL or CL). One grou received he DCL remen nd he oher w receiving he CL remen. During he weny eigh week of remen, he uden in boh grou worked cooerively in em of four member. There were rndomly igned o em bed on heir level of SRL, nd ech em w conied of wo high nd wo low SRL member. The reerch finding reveled ignificn difference beween uden ugh in he wo inrucionl mehod on ll he hree deenden vrible: (i) he number of reviion mde o he rojec rool, (ii) elf-rogrmming erformnce, nd (iii) he rojec develomen erformnce. The nlye indiced h uden in he DCL grou erformed ignificnly beer in mking leer djumen o he rojec rool hn heir eer in he DL grou. Thee uden in DCL grou hd hown ignificn reul on he cbiliy of elf evluion in erm of heir 109

elf-rogrmming erformnce comred o hoe in CL grou. Similrly, hee uden of DCL grou ignificnly ouerformed hoe in he CL grou on rojec develomen erformnce. A uch, he DCL inrucionl mehod ignificnly influenced he uden rogrmming erformnce hey rogreed. The roblem decomoiion wih cooerive lerning mehod ignificnly ied he uden in idenifying he oenil requiremen nd rik by zooming ll oible ngle while deermining he ofwre coe. Thi echnique heled he uden o focu on he novel roblem nd brek hem ino mller mngeble chunk for furher nlyi. Thi finding i in line wih [3] nd [9] h roblem decomoiion echnique chunking mehod did i he uden ugh in he DCL grou in dividing he k ino ynciclly reled non-overling grou of requiremen (uer or yem). Thi echnique i een o be n imorn e h enble he uden o idenify nd define he ofwre coe before roceeding o develomen. In ccordnce wih [2], hi micro level of idenifying nd nlyzing roblem ech IS develomen roce llow he DCL uden o viulize ofwre requiremen from he uer erecive. Bed on uer requiremen, he evluion on buine rocee, reliic execion nd conrin on ofwre rojec wihin he rojec coe ecrum from he mcro level o micro level hve heled hem in creing ieline of new chllenging k for he ofwre rojec. Thi llow he uden o idenify uible nd mngeble rojec coe o o comlee he rojec wihin he defined imeline. In oher word, he recogniion of requiremen, reource nd conrin, well he dicovering of roblem re crucil o ofwre rojec ucce. The roblem decomoiion roch encourge he uden o look oible comonen nd nlye every oibiliy, hu imroving heir nlyicl hinking nd roblem olving kill. Therefore hi finding demonred h uden who worked on decomoiion hrough roblem-bed lerning hd inveed more effor nd ime on requiremen inveigion nd nlyi nd hown deeer undernding on ofwre comonen comred o hoe ugh in he CL mehod. In urn, hey hve demonred beer biliy o comlee he ofwre rojec wihin he eriod of ime. Likewie, uden in he DCL grou mde ignificnly fewer chnge o he rojec coe in heir rojec rool. Thi roblem idenificion roce lied in ofwre develomen ge llowed uden o e reonble chllenging rojec coe h i wihin heir cbiliie for comleion wihou overly mbiiou in building uch fully oerionl yem. Moreover, hi roblem decomoiion echnique llow he uden o brinorm on he novel roblem h imule higher cogniive hinking nd in urn culive elf roublehooing biliie [3][10]. Through em inercion, he dicuion heled ech member o k why nd how on rogrmming emen h omehow reduce mike, incree quliy in code nd horer develomen cycle well incree elf confiden level. Thu he finding on ofwre develomen erformnce hve uored reviou emiricl udie h he doion of roblem decomoiion nd cooerive lerning yielded beer quliy, fewer defec of code nd horer develomen cycle [7][9][11]. Thi furher incree he ofwre rojec ucce re. VII. CONCLUSION The udy h emhized he imornce of roblem decomoiion nd cooerive lerning echnique in lerning ofwre develomen roce hrough hndling em rojec for clroom cdemic erformnce. The inegrion of roblem decomoiion in informion yem develomen cycle h hown ignificn imrovemen in comleing ofwre rojec. Thi echnique llow uden o chunk roblem cenrio nd uer requiremen ino mngeble funcion. Wih hi, uden hve deeer undernding of iniil roblem nd requiremen during nlyi nd deign ge h evenully o fulfill key ec owrd he ofwre rojec ucce. I emhize on roblem olving hrough decomoiion, yem requiremen, licion develomen nd imlemenion, which in urn hve hown oiive influenced in uden rogrmming erformnce. Likewie, i enble he uden o culive he eenil cogniive kill uch nlyicl, logicl nd roblem olving kill in he ofwre rojec environmen. Thu, hi regy h heled o cloe he g beween boh ge in he rcicl e nd generl reerch nlyi rincile nd o reduce he ofwre rojec filure re. Thu, hi echnique when lied in ofwre engineering could be of gre vlue in erm of guiding he uden o idenify, undernd nd mke rom deciion hroughou he rojec develomen roce. Subequenly, i romoe he develomen of rogrmming knowledge nd comeency wihin elf hrough breking brc cenrio ino mller chunk. In eching nd lerning, educor hould romoe nd enforce roblem decomoiion echnique hroughou he develomen civiie h in urn will culive innovive ofwre-building nd revel higher rogrmming chievemen. REFERENCES [1] B. Hughe nd M. Coerell, Sofwre rojec mngemen. Midenhed: McGrw Hill. 2006. [2] J. Helwig, Uing rel yem develomen rojec o enrich yem nlyi nd deign, in Proc. of ISE Conference, Columbu, 2005,. 1-6. [3] J. P. Poved nd J. T. Borr, Inducive logic rogrmming nd i licion o he emorl exreion chunking roblem. Ph.D. Progrmme on Arificil Inelligence (UPC). LSI Dermen Technicl Reor Jnury. 2007. [4] S. Shhid, T. K. Ahmd, Z. Zurinhni, nd S. Srin, Syem develomen: Wh, why, when nd how CASE ool hould uor novice ofwre engineering, The 3rd Mlyin Sofwre Engineering Conference. 2007,. 256-260. [5] L. A. Kelmn, R. McKeemn, nd L. Zhng. Erly wrning ign of IT rojec filure: The dominn dozen. Informion Mngemen Journl. 2006.. 31-36. Rerieved My 2, 2011. Avilble: h://www.im-journl.com/ittody/rojecfilure.df [6] I. Milizewk nd G. Tn, Befriending comuer rogrmming: A rooed roch o eching inroducory rogrmming, Iue in Informing Science nd Informion Technology, 2007, 4: 277-289. [7] R. Sencer. The remlined cogniive wlkhrough mehod, working round ocil conrin encounered in ofwre develomen comny. CHI Leer. 2000, 2(1): 353-359. [8] R. R. Pinrich nd E. V. DeGroo, Moivionl nd elf-reguled lerning comonen of clroom cdemic erformnce, Journl of Educionl Pychology, 82: 33-40, 1990. 110

[9] K. Tku nd M. Yuji, Ue of uor vecor lerning for chunk idenificion, In Proceeding of CONLL-2000 nd LLL-2000,. 142-144, 2000. [10] K. A. Rie nd Z.Tong. A high erformnce emi-uervied lerning mehod for ex chunking. 2005. Rerieved July 10, 2011. Avilble: h://riejohnon.com/rie/ndo_zhng_c205.df [11] C. A. Meek, R. Nfziger, nd J. K. Meek, Suden iude, ifcion, nd lerning in collborive eing environmen, The Journl of Chirorcic Educion, vol. 24, no. 1,. 19-29, 2010. Tie Hui Hui received her Docor of Philoohy (PhD) degree in Inrucionl Technology from he Univeriy Science Mlyi (USM), Mlyi in 2011 nd hold Mer in IT nd Buine (MSc in ITBu) degree from he Univeriy of Lincolnhire nd Humberide, UK. She i enior fculy member he Cenre for Pogrdue Sudie, SEGi College Penng. Her reerch inere include effecive mehod of eching rogrmming, ofwre develomen nd reerch mehodologie. Irfn N. Umr received hi Docor of Educion (Ed.D) degree in Inrucionl Deign nd Technology from he Univeriy of Piburgh, USA in 1999. He i n ocie rofeor he Cenre for Inrucionl Technology nd Mulimedi, Univerii Sin Mlyi (CITM USM), Mlyi. Hi reerch inere i minly in inrucionl deign, inrucionl regie nd e-lerning. 111