A practical application of the IDEAL model

SOFTWARE PROCESS IMPROVEMENT AND PRACTICESoftw. Process Improve. Pract. 2004; 9: 123–132Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/spip.201

A Practical Application ofthe IDEALSM Model

Research Section

Valentine Casey*,† and Ita RichardsonDepartment of Computer Science and Information Systems, University ofLimerick, Limerick, Ireland

The focus of this research is to outline the experience of a small-to-medium-sized European-based software development organization, utilizing the IDEALSM model while implementinga tailored Capability Maturity Model (CMM) software process improvement (SPI) program.The goal of the approach undertaken was to achieve process improvement rather than aspecific CMM maturity level. In doing this, the IDEALSM model was extensively researchedand employed. The benefits and limitations of the use of the IDEALSM model are presented asexperienced.

Research was carried out on a number of software process improvement paradigms priorto the final selection of the CMM. The approach employed as far as possible remained trueto the spirit of the CMM. A key element of this strategy was to see the requirements of theorganization as paramount and immediate. It was deemed important for the organization toachieve specific Key Process Areas regardless of their position in the CMM. The approachprovided the organization with the flexibility to invest in the achievement of specific maturitylevels at some future date and thereby capitalize on their current process improvement work.Copyright 2004 John Wiley & Sons, Ltd.

KEY WORDS: IDEALSM Model; capability maturity model; CMM; small-to-medium-sized enterprises; SME; process improvement;process assessment; effective process change

1. INTRODUCTION

Software process improvement (SPI) is a complexand expensive exercise that should not be enteredinto lightly or without due preparation. The cor-rect implementation of any improvement initiativeis an important undertaking. Prior to implement-ing an improvement plan, an organization mustgive serious consideration to the approach to be

∗ Correspondence to: Valentine Casey, Department of ComputerScience and Information Systems, University of Limerick,Limerick, Ireland†E-mail: [email protected] CMM is registered in the US Patent and Trademark Office byCarnegie Mellon University.SM IDEAL & SEPG are service marks of Carnegie Mel-lon University.

Copyright 2004 John Wiley & Sons, Ltd.

employed. This is particularly important for thesmall to medium sized enterprises (SME), wherethe company needs to have fast return on invest-ment (Richardson 2002). No company, regardlessof whether they are large or small, is willing toundertake any project without being assured thatthe resources expended will in fact give maxi-mum value for money. The research project pre-sented here demonstrates an implementation of theIDEALSM (Paulk et al. 1997) model in a small tomedium sized enterprise, ensuring that the organi-zation’s business requirements are top priority.

2. RESEARCH PROJECT

The methodology employed for this research wasthe action research paradigm. Action research

Research Section V. Casey and I. Richardson

entails the analysis of direct intervention of theresearcher (Gill and Johnson 1997). The traditionalaction research five-phase cyclical process–basedapproach as defined by Susman and Evered (1978)and Baskerville (Baskerville 1997) was utilized forthis research project. Using this approach allowedone of the authors perform the role of assessorin the process-improvement program detailed inthis paper. It also provided both authors with theobjectivity and structure to effectively perform theirwork.

The action research five-phase cyclical-basedapproach is coincidentally similar in structure, buttotally unrelated to the IDEALSM model. Initially, theclient-system infrastructure is established and theremaining steps are defined as follows (Baskerville1999):

1. Diagnosing2. Action planning3. Action taking4. Evaluation5. Specify learning

The client-system infrastructure is the collabora-tive agreement reached between the client (orga-nization) with whom the research is being carriedout and the researchers. This agreement defines theboundaries of the research area. The initial diag-nosing phase is where the underlying causes andreasons for change are collaboratively defined. Inthe ‘action planning phase’, the activities required tobring about that change are collaboratively planned.This is followed by the ‘action taking phase’ wherethose activities are carried out. A collaborative eval-uation is carried out and the effectiveness of theactions employed are researched. This is followedby the specifying learning phase, where the lessonslearned from the activities as a whole are directedto the client. The client may utilize this informationto amend the organizational norms to reflect thisknowledge. It is also utilized by the researchers aspart of the research program, where it is used toevaluate and validate the research carried out. Theaction research five-phase cyclical based approachwas selected prior to the selection of the IDEALSM

process improvement model. However, given thesimilarity between the two approaches, this proveda very effective research methodology to employ.

2.1. Research Questions

The SPI program was carried out within a subsidiaryof a software company employing 120 people. Itfocused on two teams with a total staff of 20, eachrunning an independent process. It had its ownProject Manager who reported directly to seniormanagement and was financed as an independentprofit center. This subsidiary, given its independentmode of operation, meets the criteria necessary tobe considered a small to medium sized enterprise.

The questions researched in this project included:

• Can standard models work within SMEs?• Do process improvement models make a mean-

ingful contribution, or do they add unnecessarylevels of bureaucracy?

• Does the IDEALSM model work?• How does the research undertaken in this project

compare with the findings on the IDEALSM

model presented by Bill Curtis at ICSE 2000(Curtis 2000a)?

3. THE ORGANIZATION AND SELECTIONOF THE CMM

The organization, Software Future Technologies(pseudonym) is based in the Republic of Irelandand while having an Irish management team, is partof a multinational company whose parent is in theUnited States. The parent organization established amainframe software application development andmaintenance company in Ireland during the lastdecade. While mainframe application developmentand maintenance continued to be a significant partof the organization’s business, by 1999 the techno-logical focus of the Irish operation had evolved toinclude the development and maintenance of appli-cations on a number of diverse platforms, includingclient server, web-based and CASE tool technolo-gies. Software Future Technologies developed andexpanded its market share in Europe and the UnitedStates, and while this research was being carried out,employed 120 staff.

When Software Future Technologies was estab-lished, a defined process had been documentedfor mainframe software application developmentand maintenance. The documented process wasstored on the company’s intranet. This proved verysuccessful. Despite the introduction of projects forother platforms, no new defined processes weredeveloped.

Copyright 2004 John Wiley & Sons, Ltd. Softw. Process Improve. Pract., 2004; 9: 123–132

124

Research Section IDEAL Model

3.1. The Selection of the CMM

Having reviewed customer and prospective cus-tomer quality expectations, it was determined by theMarketing department that there was no demandfor the organization to achieve any specific externalcertified quality standard, or the achievement of anyspecific maturity level. This allowed the assessor toreview and evaluate ISO 9001 (International Orga-nization for Standardization 1994), ISO/IEC 15 504(SPICE) (ISO/IEC 1998a, 1998b) and CMM (Paulket al. 1997), prior to the selection of the CMM asthe basis for the process improvement initiative.The goal of this review was to determine the mostsuitable and flexible model for the organization toimplement. Each model was extensively comparedand their strengths and weaknesses explored.

On the basis of the organization’s goals, an exten-sive comparison was made between the CMM andISO 9001. This included the examination of a corre-lation between ISO 9001 and the CMM (Paulk1994), and differences between them (O’Tinney1997), (Jalote 2000). The final decision to choosethe CMM was reached because ISO 9001 had spe-cific limitations from the organization’s perspective.The major limitation was that the ISO 9001 standardprovided only a minimum quality baseline for soft-ware organizations (Paulk 1994). ISO 9001 placedan emphasis on meeting minimum requirementsrather than promoting continued process improve-ment.

ISO/IEC 15 504 (SPICE) was also considered. Thisstandard was in beta test at the time the initiativewas undertaken. It did offer advantages; theseincluded the separation of process and capabilityinto two dimensions and the consistency of resultsbetween ISO/IEC 15 504 and CMM (Varkoi andMakinen 1998). The problem identified with thisoption was that there was only an academicknowledge of the model within the organizationand the cost of further training and outside supportwould be high.

On the other hand, the CMM offered a compre-hensive approach to process improvement. Therewas also the advantage of the opportunity to lever-age any formal CMM assessment that might takeplace in the medium to long term, utilizing ISO/IEC15 504 to generate an internationally standardizedrating. The compatibility between ISO/IEC 15 504and the CMM was seen as a further advantageoffered for utilizing the CMM. A commitment to

continued process improvement was a key elementin the initiative undertaken. It was also determinedthat given the location of a number of the organiza-tion’s customers in the United States, a US centricimprovement paradigm with acceptance in Europewould be the perfect model to apply.

The process improvement initiative had to beperformed under a number of constraints andwithin a defined scope to achieve goals andobjectives. The constraints included the availabilityof only a limited budget to support the effort. Therewas only local senior management support for theinitiative and only one person was available tocarry out the assessment and be the day-to-daydriving force behind it. The scope of the projectbeing undertaken was also restricted to only onepart of the organization. As a result, the goalsand objectives of increased efficiency, customersatisfaction, market share and effective processimprovement would all be impacted. At this stage,the scope, constraints, objectives and goals had onlybeen verbally discussed and their full implicationson the process improvement effort were not clearlyanalyzed.

It was understood that there was need for clearguidance on the implementation of the processimprovement initiative. That guidance appearedto be provided by the IDEALSM model (Dunaway1996). The initial selection of the IDEALSM model asthe life cycle approach to the process improvementinitiative went hand in hand with the selection ofthe CMM (McFeeley 1996).

4. USING THE IDEALSM MODEL

Having accepted the IDEALSM model (Figure 1.)as the life cycle paradigm for the initiative at apreliminary stage, its suitability was reviewed indetail. It was determined that in general it providedcomprehensive guidelines and direction with whichto manage the change process. It was noted that itparticularly addressed initial areas of concern in thekey stages of the improvement project. Managementin Software Future Technologies agreed that itshould be utilized in the development of theinitiative’s mission statement, the initial project planand as a framework for reference as the initiativedeveloped. It was also agreed that it provideda good basis for continued improvement duringfuture iterations of the process-improvement cycle.


125


Leveraging

Reviseorganisational

approach

Document andanalyselessons

Defineprocedures

and measures

Plan andexecutepilots

Plan,execute,and track

installation

Establish processaction teams

Plan actionsSet strategyand priorities

Developrecommendations

and documentphase results

Stimulus forimprovement

Set contextand establishsponsorship

Establishimprovementinfrastructure

Appraise andcharacterise

currentpractice

Acting

Diagnosing

Initiating

Establishing

Figure 1. The IDEALSM Model (Paulk et al. 1997)

However, it was recognized that the IDEALSM

model had limitations and would have to be tailoredfor use within the organization. The reason for thistailoring was that any approach taken by a small tomedium sized software development organizationundertaking effective process improvement shouldinclude the following characteristics (Richardson2002):

• Give maximum value for money.• Propose improvements that have a maximum

effect in as short a time as possible.• Provide fast return on investment.

While these characteristics are important to anyorganization undertaking such an endeavor, theyare critical to any small-to-medium-sized enterpriseas they have limited financial resources and requiretimely and effective results from any processimprovement efforts undertaken. This dictates anapproach that minimizes delay and maximizesquantifiable returns.

The Acting phase was identified from the detailedreview of the IDEALSM as the area where tailoringwas required. The organization did not have thetime nor the resources to get involved in such

a protracted exercise. Once the initiative wasunderway, prompt positive results were essentialto maintaining management and staff support. Ifthe pilot test solution and refining solution stages asoutlined in the Acting phase had been implemented,it would have had a negative impact on the processimprovement effort. The approach that would betaken was that Key Process Areas (KPA) to beimplemented would be simply tailored to the needsof the organization and based on achieving the goalsand objectives defined in the CMM.

Using a software process improvement modellike the CMM, which outlines and addresses theachievement of the goals of a mature softwareorganization, ensures that a cyclical best guessapproach is not required. Tailoring of the IDEALSM

model to meet the needs of the organizationimplementing process improvement is stronglyrecommended (McFeeley 1996).

4.1. The Initiating Phase

Utilizing the IDEALSM model as the basis for theprocess improvement initiative, the assessor and asenior manager held their first meeting. The first


126


items to be addressed were the stimulus for changeand the business reasons for change. These had to bedefined and articulated. The importance of this stepcannot be understated. Without a clear understand-ing of why an improvement initiative is undertaken,it will in most circumstances be doomed to fail-ure, given the level of commitment required by allconcerned for its successful conclusion.

The stimulus for change and the business reasonsidentified included increased profitability, betterlevels of service to existing customers and thedevelopment of new business. This initiative alsooffered the opportunity for the development of aCMM based development process, which wouldtie in with any future quality initiatives that mightbe undertaken by the parent organization. Toaddress each business reason, clear objectives weredefined and agreed upon.

One of the business reasons identified wasSoftware Future Technologies requirement for thedevelopment of new business. The objectives toachieve this were outlined as follows:

• Leverage the organizations commitment to qual-ity as a marketing asset.

• Utilize the organizations reputation within theindustry to attract new customers.

• Meet potential customer requirements so thatprior to awarding contracts an effective com-mitment to quality and process improvement isevident within the organization.

Having determined the stimulus for change, itwas now time to set the SPI initiative in thecontext of Software Futures Technology’s businessstrategy. It was determined that the initiative wouldassist with future development and expansionof business with new and existing customers byincreasing customer satisfaction and enhancingthe organization’s reputation. It would allow thedevelopment of a repeatable, effective and efficientproduction process, which would underscore thatexpansion. It would help to ensure that productdelivery was on time and within budget. It wouldalso allow management more visibility into theprocess at all stages of development, which wasrecognized as being essential for the successfulmaintenance and development of quality software.

Once business reasons and how they fitted intothe company’s overall business strategy were estab-lished, management sponsorship was sought. The

level of commitment required through manage-rial support and the availability of resources wereoutlined. The constraints that would have to beimposed on the effort were also formally definedand fully discussed. This was very important, asmanagement now clearly understood the full impli-cations of these constraints and how they wouldimpact on the level of achievable goals and objec-tives. This provided an excellent first step, whichgreatly assisted with the building of managementsponsorship. A clear understanding of why theinitiative was being undertaken, what goals it wasgoing to achieve and how they fitted into the overallbusiness strategy of the organization was invalu-able in presenting and winning the required levelof approval and commitment from senior man-agement. As a result, a local senior manager wasappointed as sponsor and a long-term commitmentto the initiative was made.

Sponsorship being established, it was essentialto develop a mechanism to manage the implemen-tation of the initiative. The need for the establish-ment of a Management Steering Group (MSG) wasaddressed and its role and responsibilities weredefined. They included the strategic and tacticaldirection of the initiative. This necessitated theestablishment of clear goals for the initiative toachieve, setting direction, priorities and monitoringthe effort.

A Software Engineering Process Group (SEPGSM)was established, and the members consisted of theassessor and the local outsourcing center ProjectManager. The role of the SEPGSM was to maintainthe motivation and the enthusiasm for processimprovement within the organization.

To complete chartering the infrastructure for theSPI initiative, the need for the establishment of aTechnical Working Group (TWG) was identified.The role and responsibility of the TWG wasdefined and included dealing with the specificelements of process improvement as the initiativeprogressed. This included the documenting andassessing of current processes as well as making ameaningful contribution in implementing processimprovement.

It was decided after discussions with the sponsorthat a mission statement would not be formulated.It was also agreed that the assessor would developa project plan for the initiative in consultation withthe MSG and the SEPGSM.


127


The project plan formally recognized the currentand desired states for process improvement in thecompany. It set priorities for change and it formallyoutlined the business reasons for the initiative. Itlaid out a time frame and highlighted key activitiesthat would have to be carried out to implementthe tailored process in keeping with the IDEALSM

model and the CMM.

4.2. The Diagnosing Phase

Before meaningful process improvement can takeplace, a company must determine its current levelof maturity. ‘If you don’t know where you are, a mapwon’t help’ (Humphrey 1989). To that end, a CMM

based process assessment was undertaken. As aninitial step, information was provided to the teamswhose process was being assessed describing whatthe CMM was about and what was involved in aCMM assessment. There was no official rolloutmeeting for the initiative as the sponsor felt itwas unnecessary. This was something the assessordisagreed with, as this offered an ideal opportunityto demonstrate visible management support for theeffort.

A Project Manager, two team leaders and threesoftware engineers completed the full CMM matu-rity questionnaire. Given the complexity of thequestionnaire, full support was provided to thosewho required clarification on any of the contentof the questionnaires. On completion of all ques-tionnaires, a comprehensive report was compileddetailing the results. This report was analyzed andareas highlighted for further research. A documentreview was undertaken and some queries wereanswered while others were raised.

Interviews were held to clarify outstandingissues. Seven people were interviewed, a ProjectManager, two team leaders and four softwareengineers. Four of the people interviewed hadcompleted the questionnaire. The interviews fol-lowed a structured approach and while dealingwith outstanding issues also endeavored to definethe existing process culture. This broadened theassessment to include areas outside the definitionof the CMM, but which directly impacted on theexisting process. On completion of the interviews, areport was compiled

Each CMM Key Process Area was addressedin the light of the questionnaire, document reviewand interview report. The results were as follows:

basic Requirements Management (KPA at Level 2of the CMM) was in place as was limited ProjectPlanning (KPA at Level 2) and Project Tracking andOversight (KPA at Level 2). A number of methods ofConfiguration Management (KPA at Level 2) wereemployed, but this did not extend to documentationor specifications and was not used on all projects.An extensive Training Program (KPA at Level 3)was in operation and Peer Reviews (KPA at Level 3)were carried out from the establishment of SoftwareFuture Technologies in Ireland. While some basicSoftware Quality Assurance practices were in place,defect prevention had been identified as a majorproblem area. In the CMM, this is a Level 5 KPA.Any process improvement initiative in SoftwareFuture Technologies would have to address thisproblem area

A maturity audit report was prepared and amaturity level determined. The outsourcing processwas rated at a CMM maturity Level 1. It was clearthat two process cultures existed side by side inthe teams reviewed. These cultures, which weredue to the background of the teams involved, werea formal disciplined approach to the process anda more ad hoc Rapid Application Development(RAD) approach. The maturity audit report withits specific recommendations was incorporated intothe assessment report. The findings were presentedto management and the teams to ensure that theywere aware of what had been achieved and toprepare them for the next phase.

4.3. The Establishing Phase

Having reviewed the assessment report and ana-lyzed pressing business requirements, the Manage-ment Steering Group prioritized the establishmentof a defect tracking system as an urgent require-ment. The template provided by the CMM Level5 Defect Prevention KPA provided guidance onwhat was required. While the researchers wereaware that it was contrary to CMM recommen-dations to implement a Level 5 process in a Level1 organization, their knowledge of the continu-ous SPICE model (Zahran 1998) and the businessrequirement gave them the confidence to proceedin this manner. Indeed, the recent publication of theCMMI (Phillips 2002) would suggest that this isnow an acceptable approach. An implementationplan was prepared, which outlined the schedule,


128


roles, milestones, and metrics, tracking and report-ing procedures for the initiative.

The Software Engineering Process Group metand undertook the management of the implementa-tion of the Defect Prevention KPA. This was purelya supportive role. The assessor prepared a non-technical translation of the KPA’s key practices forpresentation to the TWG. The SEPGSM also tookresponsibility for informing management and staffon a regular basis on the status of the initiative.

The next step was the selection of the TWG. Itwas agreed that the initial group would be made upof four members, the assessor and three softwareengineers. Membership of the TWG was to be apart time position. This would ensure that the TWGmembers were not divorced from the rest of theorganization.

The initial role of the TWG would be to examinethe requirements of the KPA. When they had estab-lished a clear understanding of them, they wouldtailor the KPA to meet the specific requirements ofthe company. When this was complete, the groupwould take responsibility for introducing new pro-cedures, roles, documentation, organize trainingand ensure the availability of essential infrastruc-ture and resources. Feedback on the progress of theinitiative would be presented to the rest of the teamson a regular basis through the SEPGSM.

4.4. The Acting Phase

At the first meeting of the TWG, the goals of theinitiative were outlined and the need for the estab-lishment of a Defect Prevention Group presented.A presentation was made on the Defect Preven-tion KPA (CMM Level 5) and the introduction ofa Defect Tracking System. Using the Defect Pre-vention KPA as a guide, a detailed action planwas drawn up. This included the identificationof resources, responsibilities, tasks and milestones.Measurements to assess the success of the initiativewere discussed and agreed upon; these incorpo-rated the metrics outlined in the implementationplan.

Having a clear understanding of what wasrequired and utilizing the CMM as a template,an effective solution was created. As outlined inSection 4, there was no benefit to be gained throughthe development of a best-guess solution, cyclicalpilot testing and refining. Indeed, it would havehad a negative impact on the project. The CMM

outlined an effective approach and the knowledgeand experience of the TWG confirmed that prior toits implementation. The tailored key practices of theKey Process Area were implemented on a projectby project basis over a three month period.

A problem-centered approach was applied inthis initiative. If all the activities outlined for theActing phase had been carried out, the time scalerequired would have had a serious impact onthe success of the program. Curtis (June 2000b)states, ‘Most successful improvement programs beginworking with projects to make improvements veryearly’. This ensures that preliminary positive resultsare available to management and staff and thisencourages support and enthusiasm for continuedprocess improvement. This is particularly importantin an SME with limited resources.

4.5. The Leveraging Phase

Having successfully introduced a Defect TrackingSystem based on the CMM Defect Prevention KPA,the team validated and analyzed what had beendone. The criteria for monitoring the performanceof the Defect Tracking System was evaluated andreviewed. The level of success achieved in all aspectsof the improvement program was assessed. It wasdetermined that the effort had been an overallsuccess, which led to a quantifiable improvement inthe operation of the software development process.The MSG, SEPGSM and TWG had all worked welltogether and having achieved an initial success ithelped to reinforce the value of team effort andsupport.

While the Sponsor had been committed tothe initiative, it became clear that there wasa lack of overall senior management supportfor further process improvement in general. Thiswas hard to understand given the success ofthe initiative and the minimal cost it incurred.There had been a positive effect on all aspectsof the process as a result of the undertaking.The goals of the effort had been closely tied tothe overall business strategy and objectives of theorganization. Better software was being producedas defects were reduced and tracked back to source,allowing preventive and effective action to betaken to stop reoccurrence. As a result of thereevaluation of management sponsorship, furtherprocess improvement was put on hold. To leveragefrom the exercise undertaken, as much material


129


and experience gained was documented and stored.This valuable resource is available for utilization infuture process improvement activities.

5. ANALYSIS AND EVALUATION

5.1. Use of IDEAL SM Model

The use of the IDEALSM model made a substantialcontribution to the success of the process improve-ment initiative outlined in this research. The onlyfull-time person employed on the initiative wasthe assessor. The Management Steering Group con-sisted of only two members, as did the SoftwareEngineering Process Group. Membership of bothgroups was on a part-time basis. The roles that bothgroups fulfilled provided a substantial contributionto the overall success achieved. When applying theIDEALSM model in a small-to-medium-sized com-pany, the temptation is to ignore roles like theManagement Steering Group. Our research demon-strates that this group is vital and that its inclusionhas direct benefits.

The Initiating phase addressed the need for aclear understanding of the initiative. Once that hadbeen determined, goals and objectives were definedand put in context with the objectives and businessstrategy of the organization. The clarity provided bythis exercise allowed senior management to makean informed commitment to provide the necessaryleadership and resources required for effectiveprocess improvement to take place. It also providedsenior management with the opportunity to decideat an early stage if the effort would be worthwhile,or if it should be abandoned. The appointmentof a committed sponsor ensured that the requiredresources would be provided and the resistanceto change that is encountered in all organizationscould be addressed and successfully overcome.The commitment of senior management is key toany successful process improvement initiative andcannot be underestimated.

The Diagnosing phase allowed a baseline of theexisting process to be established. It allowed amaturity level to be determined and the strengthsand weaknesses of the process were highlighted.This facilitated the development of clear recom-mendations for current and future improvements.During the Establishing phase, changes were out-lined and the need for a Defect Tracking System

prioritized. The Defect Prevention KPA was uti-lized as a template for process improvement basedon the business needs of the organization. An imple-mentation plan was prepared and the TWG wasestablished.

In the Acting phase, the TWG’s approach andavailable time scale did not allow or require thedevelopment of a best-guess solution. Neither didit require the use of pilot testing and furtherrefinement prior to implementing the improvementstrategy. The CMM provided the required solu-tion. The TWG did not divorce itself from the rest ofthe organization. Membership was a part-time roleand change was introduced on a project-by-projectbasis. On the basis of the Defect Prevention KPA, theTWG determined the requirements of the organiza-tion. The implementation of the KPA was tailoredto meet those defined requirements. If the pilot testsolution and refining solution stages as outlined inthe Acting phase had been implemented, it wouldhave had a negative impact on the process improve-ment effort. The organization did not have the timeor the resources to get involved in such an exer-cise. This fact was identified at an early stage andthe implementation of IDEALSM model modifiedaccordingly.

The Leveraging phase provides an excellentopportunity to evaluate what has been achievedand opportunity to learn from work that hasbeen completed. The activities of the initiative,which has been undertaken, are reviewed andanalyzed. The achievement of goals and objectivesare assessed. Lessons learned are recorded for futurereference. The evaluation of continued sponsorshipis very important. The IDEALSM model quite rightlystresses the importance of continued sponsorshipand support of senior management for effectiveprocess improvement to be undertaken in furtheriterations of the model.

5.1. Response to Research Questions

In summary, the IDEALSM model provided a goodframework for process improvements to take placewithin the SME we researched. That stated, wewould stress that the model should be tailoredto fit the needs of the organization utilizing it(Karlheinz et al. 2000). The need for tailoring isclearly understood and recommended by McFeeley(1996) who states, when referring to the IDEALSM


130


model ‘One Size Does Not Fit All!’ and ‘SPI managersmust tailor the guide to their particular situation’.

The layers of management, supervision and sup-port provided by the IDEALSM model encouragedeffective process change and improvement. To beeffective, the model did not require the establish-ment of a bureaucratic system. It offered a clearpath for continuous process improvement to takeplace. We conclude that the tailored IDEALSM modelworked.

In his workshop presentation at the ICSE 2000,Curtis (2000a) outlined the problems encounteredutilizing the IDEALSM model. The problem as hedefined it is what he termed ‘the action teams separa-tion from development work’. It takes too long to setpriorities, develop and pilot its approach and imple-ment process improvements. When improvementsare finally implemented, it is too late, both manage-ment and staff have lost faith in the improvementprogram. This paper concurs with those findingsand has been researched independently of his work.We would not suggest that there should be a quickfix approach, but the CMM offers clear guide-lines with regard to where and how improvementscan be made. The CMM template should be uti-lized to ensure that timely and effective processimprovement takes place.

6. CONCLUSION

The Defect Tracking System introduced as a resultof this effort led to a large reduction in the numberof defects produced. Defects that arise are logged,corrected, tracked back to source and discussed.Preventive action is taken to ensure that wherepossible such defects do not arise again, or areidentified at an earlier stage in the process. Thelevel of rework has been reduced; milestone anddeadlines have been met on a more consistent basis.Both teams are much more aware of quality as eachteam member spends time working as part of theDefect Tracking Team. These improvements havebeen achieved with the support and guidance of theIDEALSM model. The CMM provided the template,while the tailored IDEALSM model provided theframework to implement it.

As demonstrated in our research, the tailoredIDEALSM model is a useful framework for soft-ware companies who wish to implement process

improvement. The Initiating, Diagnosing, Establish-ing and Learning phases are of particular value. Ifthe activities of the Acting phase are adapted to theneeds of the organization then required results canbe achieved.

ACKNOWLEDGEMENTS

This research has been supported by the ScienceFoundation Ireland Investigator Programme B4-STEP (Building a Bi-Directional Bridge BetweenSoftware Theory and Practice).

REFERENCES

Baskerville RL. 1997. Distinguishing action researchfrom participative case studies. Journal of Systems andInformation Technology 1(1): 25–45.

Baskerville RL. 1999. Investigating information systemswith action research. Communications of the Association forInformation Systems 2(19) 1–31.

Curtis B. 2000a. T11 software process improvement: bestpractices and lesson learned. Tutorial notes. ICSE 2000,Limerick, Ireland, June 2000a, 4–11.

Curtis B. 2000b. From MCC to CMM: technology transferbright and dim. Proceedings of the 22nd InternationalConference on Software Engineering, Limerick, Ireland,June 2000b, 521–530.

Dunaway DK. 1996. CMM Based Appraisal forInternal Process Improvement (CBA-IPI) Team MembersHandbook. CUM/SEI-96-HB-005, Carnegie MellonUniversity, PA, USA.

Gill J, Johnson P. 1997. Research Methods for Managers, 2ndedn. Paul Chapman Publishing, London, UK.

Humphrey WS. 1989. Managing the Software Process.Addison Wesley, Boston, USA.

International Organization for Standardization. 1994.ISO9000-1 quality management and quality assurancestandards, Part 1: Guidelines for selection and use.Geneva, Switzerland.

ISO/IEC. 1998a. Information technology – Software processassessment – Part 1: Concepts and introductory guide,ISO/IEC TR 15504-1.

ISO/IEC. 1998b. Information technology – Process assess-ment – Part 2: Performing an assessment, ISO/IEC TR 15504-2.


131


Jalote P. 2000. CMM in Practice. Addison-Wesley, Boston,USA.

Karlheinz K, Hansen HW, Thaysen K. 2000. Applyingand adjusting a software process improvement modelin practice: the use of the IDEALSM Model in smallsoftware enterprise. ICSE 2000, Limerick, Ireland, June2000, 626–633.

McFeeley B. 1996. IDEALSM A Users Guide for SoftwareProcess Improvement, Handbook, CMU/SEI-96-HB-001,Carnegie Mellon University, PA, USA.

O’Tinney M. 1997. Comparing ISO 9000, Malcom Baldrigeand the SEI CMM for Software. Prentice Hall, UpperSaddle River, New Jersey, USA.

Paulk MC. 1994. A comparison of ISO 9001 and thecapability maturity model for software. Technical ReportCMU/SEI-94-TR-12 ESC-TR-94-12, SEI, Carnegie MellonUniversity, PA, USA.

Paulk MC, Charles Webber V, Curtis Bill Chrissis MaryBeth. 1995. The Capability Maturity Model, Guidelines forImproving the Software Process. Addison-Wesley, Boston,USA: (8th Reprint November 1997).

Phillips M. 2002. CMMI: improving process for betterproducts, key note address. Profes 2002, Rovaniemi,Finland, December 2002, 9–11.

Richardson I. 2002. SPI models: What characteristics arerequired for small software development companies?Software Quality Journal 10(2): 101–114.

Susman G, Evered R. 1978. An assessment of the scientificmerits of action research. The Administrative ScienceQuarterly 23(4): 582–603.

Varkoi TK, Makinen TK. 1998. Case study of CMM

and SPICE comparison in software process assessment,engineering and technology management, 1998.Pioneering new technologies: management issues andchallenges in the third millennium. IEMC ‘98 Proceedings,International Conference, Puerto Rico, USA, October1998, 11–13, 477–482.

Zahran S. 1998. Software Process Improvement, PracticalGuidelines for Business Success. Addison-Wesley, Esser,UK.


132

A practical application of the IDEAL model

Documents

Transcript of A practical application of the IDEAL model