Application of a Systemic Soft Domain-Driven Design Framework

Applicatio of a Systemic Soft Domai-Drive Desig Framework Mohammed Salahat, Steve Wade, Izhar Ul-Haq Abstract This paper proposes a soft systems approach to domai-drive desig of computer-based iformatio systems. We propose a systemic framework combiig techiques from Soft Systems Methodology (SSM), the Uified Modellig Laguage (UML), ad a implemetatio patter kow as Naked Objects. We have used this framework i actio research projects that have ivolved the ivestigatio ad modellig of busiess processes usig object-orieted domai models ad the implemetatio of software systems based o those domai models. Withi the proposed framework, Soft Systems Methodology (SSM) is used as a guidig methodology to explore the problem situatio ad to geerate a ubiquitous laguage (soft laguage) which ca be used as the basis for developig a object-orieted domai model. The domai model is further developed usig techiques based o the UML ad is implemeted i software followig the Naked Objects implemetatio patter. We argue that there are advatages from combiig ad usig techiques from differet methodologies i this way. The proposed systemic framework is overviewed ad justified as multimethodologyusig Migers multimethodology ideas. This multimethodology approach is beig evaluated through a series of actio research projects based o real-world case studies. A Peer-Tutorig case study is preseted here as a sample of the framework evaluatio process Keywords SSM, UML, Domai-Drive Desig, Soft Domai- Drive Desig, Naked Objects, Soft Laguage. I. INTRODUCTION HE failure of software support systems has bee well T documeted over the years, ad may of these failures have bee attributed to poor busiess process modellig (Joseph Barjis, (2008)). The systems failed because the busiess process model developed did ot adequately support the process of desigig ad implemetig the software support system. Oe of the mai reasos for iformatio systems failure is a tedecy to cocetrate o the techical aspects of desig rather tha uderstadig the busiess eeds [2]. Mohammed Salahat is with the Iformatics departmet, School of Computig ad Egieerig, Uiversity of Huddersfield, UK, as a part-time- PhD studet, ad Lecturer with Ajma Uiversity i UAE. (e-mail: m.salahat@hud.ac.uk & abac.hasa.m@ajma.ac.ae) Steve Wade is with the Iformatics departmet, School of Computig ad Egieerig, Uiversity of Huddersfield as a Seior Lecturer (e-mail: s.j.wade@hud.ac.uk) Izhar Ul Haq is with New York Istitute of Techology, Abu Dhabi Campus, UAE, as Associate Professor (e-mail: ihaq31@yahoo.com) There is a eed for a systematic approach for capturig the iformatio required by busiess processes [1]. This suggests a eed to bridge the gap betwee busiess process modellig, iformatio systems modellig, ad implemetatio. Our previous work [4, 5] proposed ad evaluated a developmet framework to deal with soft ad techical systems aspects with a emphasis o modellig workflow. The evaluatio results guided us to modify the framework i a ew directio i which the cocept of workflow is less domiat. The ew modified framework focuses o Domai-Drive Busiess Process Modellig (DDBPM) as a approach to modellig busiess processes i a object-orieted domai model. This approach is amed SDDD (Soft Domai-Drive Desig). SDDD combies Soft systems Methodology (SSM), the Uified Modellig Laguage (UML), ad the Naked Objects implemetatio patter. SDDD aims to ivestigate, aalyze ad model a busiess domai so that we ca implemet it as a software support system. SDDD is a multimethodology systemic framework cosistig of four phases with guidig procedures to steer the developer betwee the various compromises that eed to be made throughout the developmet process. Sectio 2 reviews related work. Sectio 3 itroduces SDDD ad explais the basic structure of the framework. Sectio 4 ad 5 the discuss the framework i more detail as a multimethodology approach. Sectio 6 is a brief descriptio of a practical case study i which the method has bee applied. Sectio 7 presets a reflectio o the framework ad the learig process of applyig it suggestig further research. II. RELATED WORKS Domai-Drive desig is a approach that seeks to model the system processes as a domai model ad develop a software support system based o it. The first step of the DDD approach is to develop a Ubiquitous Laguage which cosists of differet cocepts, diagrams, ad documets to facilitate the commuicatios betwee the developers ad domai experts. The Ubiquitous Laguage will be used to create the domai model by the developers ad domai experts [6]. UML defies a umber of diagrams that ca be used to model the busiess process [7] but lacks the ability to explore the soft issues related to the problematic situatio which ca be hadled usig Soft System Methodology. SSM [8, 9, ad 9] is a established meas of problem solvig that focuses o the developmet of idealised models of relevat systems that ca the be compared with real world couterparts. Some re- 2220

searchers have explored the relatioship betwee SSM ad object orieted aalysis ad desig techiques i geeral [11] but less has bee writte about the applicatio of these techiques i the cotext of the UML. UML is cosidered by DDD to model the busiess domai as a Domai Model. Recet works [12]-[13] cosider the SSM coceptual model as a focal poit for likig SSM ad UML by mappig the activities of a SSM coceptual model ito UML use-cases. Recet examples of this approach ca be foud i SWfM [7] ad our previous works [4]-[5]. The SDDD framework guides the developer ito creatig a Soft Laguage which cosists of the output of the SSM stage to deal with the soft aspects which are ot hadled explicitly by Domai Drive Desig. The SSM Coceptual Primary task Model (CPTM) is used to map huma activity to a UML use-case model usig a ew elaboratio techique. Use-cases, as abstractios of busiess activities, are used to model the busiess process i a domai model usig UML diagrams ad based o the philosophy of DDD which employs the idea of Kowledge Cruchig durig the differet stages. SDDD employs the same philosophy durig its four stages as explaied i later sectios. Other researchers have made use of various extesios to the UML. For example [3] employed a systemic framework combiig SSM ad UML extesios proposed by [14] to model the busiess process of a maufacturig factory. Their framework is based o Migers Multimethodology ideas [15] but does ot ecompass the software implemetatio phase of developmet. Our previous works [4]-[5] preseted a systemic framework for busiess process modelig ad implemetatio as a workflow system, that framework was described as a multimethodology based o Migers Multimethodology [15] ad it compassed the software implemetatio phases of developmet. This paper aims to preset a updated framework for modelig the system busiess processes as a domai model ad implemetig it as a software support system. The SDDD framework combies SSM, UML techiques, ad the Naked Objects implemetatio patter. To the best of our kowledge, this combiatio has ot bee applied i a itervetio before. III. DOMAIN-DRIVEN BUSINESS PROCESS MODELLING The orgaizatio busiess process must be well defied ad modelled for the implemetatio. A busiess process ca be defied as the trasformatio of somethig from oe state to aother state through partially coordiated agets, with the purpose of achievig certai goals that are derived from the resposibility of the process ower [16]. There are may defiitios of busiess process. Most of these defiitios are based o the idea of a busiess process as a determiistic system that receives iputs ad trasforms ito outputs followig a series of activities. For example [17] defies busiess processes as structured sets of activities desiged to produce a specified output for a particular customer or market. Good iformatio systems software will support the orgaizatio work by hadlig the iteral busiess process ad cotrol all aspects affectig the executio of the process. The busiess process must be supported with good busiess process modellig ad implemetatio techiques that ca aalyze, model, ad implemet the busiess process i a professioal way to achieve the orgaizatioal goals [18]. Domai-Drive Desig ca be used to model the busiess process as a busiess domai model [6]. A Ubiquitous Laguage (UL) is geerated first as a commuicatio tool betwee differet stakeholders ad the domai model will be geerated ad implemeted based o UL. We have adapted the idea of a UL ito a Soft Laguage which icorporate certai artefacts of a SSM aalysis ito the model. A objectorieted domai model ca be extracted from this Soft Laguage through a trasitio process which will be explaied i the ext sectio. We argue here that SSM helps the developer to gai a deep uderstadig of differet stakeholders perspectives which will eed to be represeted i the Soft Laguage. UML diagrams are sufficiet tools for requiremet modellig to support busiess process modellig i a objectorieted domai model [19]. Whe it comes to implemetig the system we have made use of the DDD implemetatio patter (i.e. Naked Objects) to reflect the system iterface directly from the domai model. Naked Objects is described as a ope-source java based framework desiged to ecourage the creatio of busiess systems from busiess objects [20]. The latest versio makes use of the Microsoft.Net framework. V. SOFT DOMAIN-DRIVEN DESIGN FOR BUSINESS PROCESS MODELLING AND IMPLEMENTATION The proposed framework is based o research ito multimethodology, which justifies combiig methods for the same busiess itervetio [15]. It is a multi-method framework which iteded to guide the developer through a ivestigatio of a problematic situatio. The purpose here is to isure that a comprehesive uderstadig is achieved i order to facilitate the modellig ad implemetatio of the domai-drive busiess processes as a software support system. The framework is beig developed through a series of actio research case studies. Actio research requires the participatio of the researchers i the developmet process. Accordigly our case studies have ivolved developmet projects withi our ow school. Our first two case studies have focussed o the developmet of a peer-tutorig-system ad a support system for the school s combied studies programme. The researchers are part of the school ad they are participatig i daily activities related to the case studies. The proposed framework SSDDDF (Figure 1) is focused o modellig ad implemetig of the domai-drive busiess process as a software support system. SSM is used as a guidig ad learig methodology with techiques icludig UML ad implemetatio patter (Naked Objects)) embedded 2221

withi it. The DDD philosophy is adapted to geerate a SL istead of UL ad it will be a iput to the ext stages. The implemetatio patter is used after the geeratio of the fial refied chage report which is a iput to the implemetatio process. The research ca t be a discrete evet but a process that has phases with activities to be performed; the research process cosist of four geeric phases [15]: 1- Appreciatio of the problematic situatio ad uderstadig why the problem exists as experieced by the ivolved actors. 2- Aalysis of the methods ad the data produced durig the appreciatio stage to uderstad how ad why they are geerated. 3- Assessmet of alteratives that may be improve the curret situatio to better tha it is, it icludes iterpretatio of the results. 3- Actio icludes reportig about the results i order to recommed chages for improvig the situatio. Usig this geeric model, the proposed framework cosists of four phases ad each phase cosists of a group of activities. I the ext sectio, the framework explaied i details ad Evaluatig the problem usig SSM stage cosists of three activities represeted i three steps [3]. The three steps equate to the appreciatio, aalysis, ad assessmet steps of Migers geeric model. Domai model geeratio takes place usig UML modellig techiques because SSM lacks to techiques for takig actios [3], ad this is equivalet to actio step i Migers geeric model. I our framework, domai modellig ad implemetatio is equivalet to actio step i Migers geeric model. So, the proposed framework satisfies the geeric process of coductig a actio research i the busiess itervetio. SSDDDF represeted i Figure1, Figure 2 represets the coceptualizatio of the framework, ad Figure 3 represets the logical processes embedded i it. VI. THE FRAMEWORK OVERVIEW The proposed framework cosists of four phases ad each phase cosists of a group of activities. The details of these phases are as follows: A..Pre-SSM Phase This phase cosists of the followig activities: 1- Iitial problem idetificatio The problem i a specific area will be determied iitially before startig the process of ivestigatio. 2- Stakeholder roles aalysis This step to clarify the roles of all parities ivolved i the problem ivestigatio to avoid ay coflicts ad to facilitate further proceedig ito the other steps. 1. Iitial problem idetificatio 2. Stakeholder roles aalysis 3. Evaluatig the problem usig SSM 6. Rethik 2-5 4. Geerate SDDD Soft Laguage ad use it to geerate Domai_Drive Busiess Process Model usig UML 5. Geerate a proposal about the DDBPM geerated durig this phase. This will be used i the implemetatio phase, ad it will iclude the whole models developed durig the previous phase ad how to use them i the implemetatio phase. The report will be refied by matchig it with previous stages output util cosidered adequate for implemetatio 7. Domai Model Implemetatio usig DDD implemetatio Patter (i.e. Naked 9. Exit Objects) 8. Rethik 10. (6-7) Reflect o the process ad record learig 9. Exit 10. Reflect o the process ad record learig Fig. 1 A Systemic Soft Domai-Drive Desig (SSDDDF) 2222

Pre-SSM Phase 1-Iitial problem idetificatio (Output: Problem statemet) 2-Stakeholder roles aalysis (Output: Differet views) SSM Phase -Evaluatig the problem usig SSM. (The output: Rich Picture, Root - Defiitio, Coceptual Model, CATWOE) Rethik Rethik Fig. 2 The coceptualizatio of SSDDDF Post1- SSM Phase Geerate SDDD Soft Laguage (The output of SSM will be iput to this laguage) Geerate the Domai- Drive Busiess Process Models usig UML (Use case, class diagram, etc) The fial report icludes chages required to the busiess domai ivestigated based o SSM philosophy (Domai-Drive Busiess Model-> a group of UML diagrams) Post 2- SSM Phase Implemet the software support system based o the geerated Domai-Drive Busiess Process Model usig DDD Implemetatio Patter (i.e Naked Objects) EXIT Reflect o the framework Applicatio ad record learig 2223

b) SSM Applicatio Phase 1- Evaluatig the problem usig SSM SSM is a guidig methodology of the research ad as show i figure (1), there is a rethik about the steps (2-5) which icludes the applicatio of SSM to evaluate the problem; SDDD techiques are used to model the domai busiess processes ad a chage report will be geerated which icludes the modelled domai ad how to implemet it. The output of SSM stage will be a iput to Soft Laguage of SDDD. This laguage is a importat part of SDDD ad represets the commuicatio tool betwee the differet stakeholders. SSM applicatio cosists of the followig steps: 1.1 Ivestigatig the problem situatio usig rich picture model Aythig ca be icluded i rich picture ad it is used to support the overall uderstadig of the orgaisatio situatio, goals, structure, ad issues affectig the problem situatio. 1.2 Modellig the relevat system usig root defiitio ad coceptual model Root defiitio is used to determie the purpose of the system ad the iterested parties. Root defiitio costructs from the differet views of parties cocered, ad these views represet the expected fuctios of the system. Root defiitio represets the missio of the target system ad look at the orgaizatio or the problem situatio from differet poits of view. Root defiitio is oe setece ad over all structure should be tested usig CATWOE. (For details, see 8, 9, ad 10). RD will be used to costruct the coceptual model (CM) or cosesus primary task model (CPTM) ad it represets the huma activity model. 1.3 Compare the (CM) with the real world The coceptual model, as a abstract represetatio, will be compared to the real world (the curret orgaizatioal process) for validatio. If the orgaizatio busiess process model does ot exist, the the coceptual model will be used as a basis to model it as a domai model [11]. The compariso will use the activities, orgaizatioal goals, objectives, ad the structure usig rich picture, root defiitio, ad coceptual model. 2- SDDD Soft Laguage Soft Laguage is the first product of SDDD. It cosists of all documets ad diagrams represetig the busiess domai as commuicatio tool betwee the differet stakeholders. The proposed framework suggested that models developed followig the Pre- Soft N Iitial Problem Idetificatio Stakeholders Aalysis Create rich picture, root defiitio, coceptual model, ad compare CM with existece model Geerate chages Report N -Geerate SDDD Soft Laguage ad Object-Orieted Domai Model usig UML Is the Report adequate for implemetatio The fial refied chages report Implemet the software support system based o the fial refied chages report usig DDD implemetatio patter (i.e. Naked Objects) Is it Adequate Implemetatio? EXIT Reflect o the framework Applicatio Fig. 3 The embedded logic i SSDDDF Pre-SSM Phase SSM Phase Systems Methodology (SSM) ad SSM Phases could provide useful iput to the developmet of a soft laguage (SL). SSM helps the developer to gai a deep uderstadig of differet stakeholders perspectives which will eed to be represeted i the ubiquitous laguage (Soft Laguage). C. Post1-SSM Phase 1- Object-Orieted Domai modellig usig UML The coceptual model (CM) or cosesus primary task model (CPTM) is represets a geeral view of the domai Y Y Post 1- SSM Phase Post 2- SSM Phase 2224

fuctioal perspective. The decompositio of CM ito subsystems will take place usig a subsystem descriptio table [11] ad each subsystem activity will be represeted i a activity descriptio table. There is a idetical similarity betwee coceptual model activities ad use cases which make the coversio process possible ad straightforward. A ew elaboratig techique is used to elaborate about ay activity to be coverted to a use case. This techique represeted i Figure 6 ad demostrated through the case study. 1.1 Buildig a subsystem descriptio ad activity descriptio tables Subsystem descriptio table will be prepared for each subsystem which icludes subsystem umber, ame, head, ad activities. The, a activity descriptio table will be prepared for each activity ad it icludes subsystem umber ad ame, activity ame, precedig ad followig activities, precoditio, iput ad output data, tasks, busiess rules ad costraits, post coditios, required skills ad capabilities, role ame, ad performace criteria. 1.2 Covertig the activities of the coceptual model ito use cases Activities will be tested to determie their goals, ad some of the activities will be combied ad some of them will be decomposed. The activities ad their goals will be tested ad mapped to UML use cases as oe-to-oe relatioship. All use cases will be combied i the use case diagram which cosists of use cases ad their actors. The use case diagram is part of the use case model which is represetig the orgaizatioal busiess process ad it will be the basis for modellig the object-orieted domai model. 1.3 Use cases aalysis ad modellig Each use case will be described usig a textual format template. Each use case will be modelled usig UML activity diagram, sequece diagram, ad class diagram. The activity diagram is used to model the fuctioal, iformatioal, behavioural, ad orgaizatioal work flow perspectives. The sequece diagram is used to model the iteractio betwee the use case objects (the dyamic aspects of the workflow system). Fially, class diagrams for the static ad orgaizatioal structures for each use case will be developed. 1.4 Developig the class diagrams Class diagrams developed to model the behaviour of all use cases will be combied together i oe class diagram called the aalysis model. This model will be coverted to a desig model, by addig to it the desig aspects required to desig the object-orieted domai model. 2- Geerate the chages a proposal Chage proposal to improve the domai model will be produced ad it icludes the whole models developed durig the previous stages ad guidelies for usig them i the implemetatio stage. 3- Geerate the fial refied chages report The report cotets will be matched agaist previous stages results util a adequate report is a achieved. D. Post2-SSM Phase 1- The domai model Implemetatio DDD implemetatio patter (i.e. Naked Objects) will be used i this stage because it s critical to start the implemetatio before refiig the proposed modellig report. The domai model (maily class diagrams) will be used to prototype the software system required. 2- Refiig the implemeted software support system The implemetatio results will be matched to the refied modellig report ad if ay deviatio available must be maaged. This step represeted i figure (1) as rethik (6-7). 3- Exit ad reflect o the framework applicatio Exit implemetatio refiemet step whe a adequate software system reached. The a reflectio o the role of each compoet of the framework will take place. Fially, lessos leared from combiig SSM, UML, ad DDD implemetatio patter will be recorded to guide further applicatios. VI. THE CASE STUDY We have bee egaged i a iformatio systems developmet project usig SSM ad UML techiques withi a agile framework to make recommedatios about the developmet of a itraet for the academic school i which we are employed. At the begiig of the project the departmet had a operatioal itraet but this was ot widely used. A iformatio system strategy was iitiated to ivestigate ways i which the itraet could be developed to support the uiversity missio ad departmetal goals. Iitially we used use cases as the primary fact-gatherig techique but certai limitatios i this approach led us to a more thorough SSM-based aalysis of the situatio. We argue that the techiques of SSM ca help the developer to idetify a richer set of use cases tha would otherwise be possible but developers with a full use case model still have may challeges ahead of them. We are iterested i object orieted desig ad the view that all busiess behaviour idetified i the use case model should be ecapsulated as methods o domai objects. Thus, a Studet object should ot just be a collectio of data about the Studet; it should ecapsulate all the behaviours that we eed to apply to a studet. I Domai-Drive Desig these are ofte referred to as 'behaviourally-rich' domai objects. A umber of software frameworks have bee developed to allow programmers to build prototype applicatios directly from a behaviourally rich domai model implemeted i a object orieted programmig laguage. Promiet amogst these is the Naked Objects implemetatio Patter. This is the oe that we have chose to use to implemet our prototype applicatios. I the ext sectio we preset a quick superficial descriptio of how the method might be applied to a relatively simple project, the desig ad implemetatio of a peertutorig system. 2225

A. Peer-Tutorig System Developmet It aims to desig ad implemet peer-tutorig system for itroductory programmig uit i the departmet of iformatics to support the studets ad reduce umber of failures. Oe of the curret problems facig studets ad lecturers i uiversity is the difficulty of uderstadig ad masterig the skills required to write ad ru computer programs successfully. A umber of researchers have suggested that peer tutorig ca be particularly useful to support this type of learig because it allows learers to lear ad support each other [21], ad it is beeficial to help studets lear ad practise the required skills more actively i a settig that ecourages them to be more active ad itellectually egaged [22]. Other researchers [23] reported about the problems of teachig programmig course at Victoria Uiversity i Australia ad they proposed a approach to ehace the delivery of this module. [24] Raised the difficulties of teachig programmig course i Chiese uiversities ad discussed differet moder icorporatig strategies, to solve this problem, which icludes Cocept Mappig, Peer-learig ad E-learig methods. The proposed solutios to recap the difficulties of teachig programmig uit by the metioed researchers cocetratig o the delivery methods oly without ivestigatig all soft ad hard systems issues that ca cause such a problem [23]-[24]. I this work, we proposed Peer- tutorig system as a improvemet of the teachig process ad to ehace the studets uderstadig which may be reduce the percetage of failures. I the ext sectios we will show how the method is applied. 1- Pre-SSM Phase 1.1 The problem idetificatio The Departmet of Iformatics i the School of Computig ad Egieerig at the Uiversity of Huddersfield i UK ad Iformatio Techology College at Ajma Uiversity of Sciece ad Techology i UAE both offer itroductory programmig modules for their first year computig studets. These modules focus o Java programmig; lecturers face certai difficulties related to studets uderstadig of the subject because of the ature of the required problem-solvig skills. Studets require more tutorig ad practical sessios to help them practise differet exercises i order to ehace their uderstadig ad practical skills. Both Uiversities expect that implemetig a peer-tutorig system will reduce the failure rate. The departmets wat to kow how to select tutors amog good studets ad how to reward them. 1.2 Stakeholder Determiatios The stakeholders of the required system were determied to be peer tutor, peer tutee, lecturer, ad maagemet. The stakeholders have differet expectatios of the system. Peer tutors are geerally lookig for teachig experiece to be added to their CVs. Peer tutees are lookig for extra help. Lecturers are lookig to reduce their workload, ad to determie which studets most require tutorig sessios. Maagemet look to reduce the umber of failures o programmig modules. 2- SSM Phase 2.1 Ivestigatig the problem situatio usig a rich picture I order to develop a rich picture of the situatio uder study, a umber of iformatio sources were used to capture views of the itroductory programmig uit from the perspective of the maagemet (the school & the college i both uiversities), lecturers, ad studets. Iterviews with the school (or college) admiistratio ad groups of studets were coducted to uderstad the problematic situatio of teachig itroductory programmig course ad set out suggestios to solve the problems. Rich pictures were used as a tool used i this ivestigatio. A umber of differet pictures were draw the followig is a simple early example. Fig. 4 Peer-Tutorig System Rich Picture 2.2 Modellig the relevat system usig SSM The relevat system was modelled usig a root defiitio ad coceptual models. Our iitial root defiitio was as follows: a peer-tutorig system for the iformatics departmet will help i the selectio of peer- tutees ad peer-tutors, the schedulig of tutorig sessios based o the availability of rooms, tutors, ad tutees. The system will also moitor the 2226

perceived beefit to tutors ad the progress of tutees i icreased self-cofidece as well as measure the impact o failure rates. A variety of coceptual models were the developed to model the key activities i the system. From these a simple Cosesus Primary Task model (CPTM) was developed idetifyig the core activities that the first versio of the system would eed to support. This preseted i figure 5. Stakeholder Name Descriptio Goal Priority (Low, Medium, High) Descriptio Busiess Activity Name Descriptio Coceptual Model (image) Use Case Idetify Tutors Idetify Room Reward Tutors Idetify Tutees Schedul Ru Tutori Fig. 5 CPTM of Peer-tutorig System 2.3 Compare the coceptual model to the real world SSM required the ivestigator to compare the produced coceptual model with the actual real life work. There is o real life PTS available to be compared with the developed coceptual model. I this case, the coceptual model will be cosidered the base to model the PTS system as a domai model. The CPTM, as a combiatio of all coceptual models, ad by cosiderig the other compoets of SL will be used i the ext phase for to geerate the domai model as stated i the begiig. 3- Post1- SSM Phase 3.1 Movig from SL to domai model usig UML This sectio cosists of three parts: covertig CPTM ito use cases, use case modellig usig UML, ad Class diagram developmet. 3.2 Covertig CPTM ito use case Ay activity required software support will be selected as a use case. The stage of movig from a SSM coceptual model to a use case model is ot as straightforward as this high-level discussio would suggest. I thikig this through we have bee pushed towards makig a clear distictio betwee stakeholder goals, busiess activities ad use cases. The followig model (Fig. 6) shows the relatioship betwee these key abstractios. Fig. 6 Movig from a SSM to use case diagram The model suggests a hierarchy of busiess activities related to stakeholder goals that are take to be the primary reasos for developig the system. The busiess activities would be represeted i a hierarchy of coceptual models with the lowest models cotaiig more primitive, elemetary busiess activities tha the higher oes. A idividual busiess activity is represeted i cotext i the image of the coceptual model of which it is a part. Some of the determied use cases are preseted i the followig Use Cases Diagram (Fig. 7). Fig. 7 Use case diagram 3.3 Developig the class diagram of PTS Each use case preseted usig textual template, activity diagram, sequece diagram, ad all use cases are combied i a use case diagram. The ext step i the process is to take the busiess logic idetified i the use cases ad associate it with classes i a class diagram. We have followed the guidelie that all importat busiess logic must be implemeted i classes i the domai model. A iitial class diagram is preseted below. (Fig. 8) 3.4 Chage report geeratio ad refiemet As show i the framework (SSDDDF), there is a draw back to the previous stages to refie what s doe durig Pre- SSM, SSM, ad Post1-SSM. This refiemet is essetial to be sure that the exact chages required already modelled well as a domai model. As a guidig methodology, SSM focus o 2227

the geeratio of the required chage report as a result to be recommeded for the maagemet actios [8]-[9]-[10]. used to geerate a iitial prototype where the iterface allows users to iteract directly with the domai objects. A screeshot is provided below to give a idea of what the iitial prototypes looked like: (Fig. 9) Fig. 8 Class Diagram of PTS SSDDDF exteded SSM further steps to iclude implemetatio as a major actio to be take as part of the improvemet chage to ehace the ivestigated situatio. This idicate that the implemetatio will be started after the completio ad the refiemet of the chage report (icludes the domai model) to facilitate the implemetatio process ad elimiate the possibility of system failure sice all soft ad hard system cocers are ivestigated, modelled, refied, ad icluded i the object-orieted domai model for implemetatio. 4- Post2-SSM Phase 4.1 Prototype Desig, Implemetatio, Refiemet The class diagram is used to extract the domai objects which lead to a domai model which was implemeted i VB ad the Naked Objects implemetatio patter. This process is Fig. 9 Naked Object Screeshot from PTS Prototype More improvemet ad work is goig o to ehace the productivity of the prototype to be a real system. Curretly, we are Naked Objects.Net to get a real live software product, ad may domai-drive desig features added to this versio. The ew output of the curret work ad further ehacemet o the proposed framework will be a target of a ew publicatio. B. Other case studies This research is part of ogoig research aims to evaluate the proposed multimethodology framework usig differet case studies. Combied studies programs, work placemet maagemet systems, studet associatios systems, ad others are a group of case studies which allow the actio research approach to be applied by the researchers. We aim from this to fid differet importat issues related to the framework i order to evolve as a ISD framework. VII. REFLECTIONS ON THE FRAMEWORK Our work i applyig the framework to a series of realworld developmet projects has focussed our attetio o a umber of issues that we had ot cosidered at the outset. Some of these preset difficulties for the further developmet of the framework which preset opportuities for further research. Some of these will be briefly discussed i this 2228

sectio. a) Role of Re-Use ad Desig Patters i Domai Modellig Our approach tries to preserve as much soft iformatio as possible i the evolvig domai models. Ievitably some of this iformatio is lost as we move from approaches that try to model what people are doig (icludig activities that do ot require software support) through to program code. At preset our framework leads to developmet of a bespoke software system based o a rich object-orieted domai model. I practice may software developers make use of reusable software compoets or wish to desig software with a eye to future reuse. There is clearly a tesio betwee our emphasis o a bespoke solutio ad the software developers objective of developig geeric, reusable software solutios. b) Represetatio of Implicit Iformatio i the Domai Model The coceptual models i SSM do ot have rigorous sytax. We have discovered that whe developig the coceptual models people ofte iclude iformatio i, for example, the sequece of activities or the kowledge required to carry out certai activities which is lost whe we move ito the use case ad object models. We are attemptig to develop clear guidelies for idetifyig this type of iformatio ad what should be doe about it. Oe possibility is that we develop our ow versio of coceptual models that do iclude a more prescriptive otatio. c) Ambiguity i the Defiitio of Busiess Process Oe of the issues that we have cofroted is the lack of cosesus about precisely what ca be defied as a busiess process ad what caot. SSM has a umber of techiques for capturig multiple stakeholder perspectives o what the key busiess processes are ad how they should be moitored. We wat to preserve these multiple perspectives for as log as possible ito the developmet process. At preset we take the Cosesus Primary Task Model produced i SSM to be a objective descriptio of what is required but we have foud that it is ofte difficult to gai cosesus i developig this model ad the to preserve that cosesus as we move o. VIII. CONCLUSION AND FUTURE WORK The work doe i this paper reviewed ad highlighted the eed for a multimethodology framework that ca hadle both soft ad hard issues of domai busiess process modellig ad implemetatio as a software support system. The ew proposed framework is developed based o the idea of Domai-Drive Desig (DDD) ad Soft Systems Methodology (SSM). We have added a soft perspective o DDD to form Soft Domai-Drive Desig. The approach is described as a systemic framework for domai busiess process modellig ad implemetatio. The framework is proposed ad justified as a multimethodology framework, icorporatig guidig steps through various key stages i the developmet process. The framework is beig evaluated ad further developed i a actio research programme. We briefly provided the example of a Peer-Tutorig-System (PTS) case study to show how the proposed framework ca be applied to a real problem situatio. The evaluatio work is ogoig i other cases icludig a Combied Studies Programme Developmet (CSPD) ad the Placemet Uit Maagemet System withi our istitutios. More details will be the target of future publicatios. REFERENCES [1] Joseph Barjis, The importace of busiess process modellig i software systems desig, Sciece of Computer Programmig Joural, vol 71,pp 73 87, 2008. [2] Alter, S., The work system method: Coectig people, processes ad IT for busiess results, Work System Press, Larkspur, CA, 2007. [3] Sewchurra, K. & Petkov D, A systemic Framework for Busiess Process Modellig Combiig Soft Systems Methodology ad UML, Iformatio Resources Magemet Joural, 20, 3, IGI Publishig, PA,USA, P. 46-62., 2007. [4] Salahat, M., Wade, S., Lu, J., A systemic Framework for Busiess Process Modellig ad Implemetatio, I the proceedig of 5 th Iteratioal Coferece o Iovatios of Iformatio Techology (Iovatios 08), UAE Uiversity, Al Ai, UAE, i IEEE xplore 978-1- 4244-3397-1/08., 2008. [5] Mohammed Salahat, Steve Wade. A Systems Thikig Approach to Domai-Drive Desig. I the proceedig of UKAIS2009 coferece, Oxford Uiversity, Oxford, UK, 2009. [6] Eric Eva, Domai-Drive Desig Tacklig Complexity i the Heart of Software, Addisso Wesley, 2004. [7] Al Humaida, F., Evaluatio ad Developmet Models for Busiess Processes, PhD thesis, Uiversity of Newcastle, UK, 2006 [8] Checklad, P., ad Poulter J., Learig for Actio. A short Defiitive Accout of Soft Systems Methodology ad its use for Practitioers, Teachers ad Studets, Joh Wiley ad Sos Ltd, West Sussex, Eglad, 2006. [9] Checklad, P., Systems Thikig, Systems Practice, Joh Wiley ad Sos Ltd, West Sussex, Eglad, 1999. [10] Checklad, P. ad Holwell, S.E., Iformatio, Systems ad Iformatio Systems, Makig sese of the field, Joh Wiley ad Sos Ltd, West Sussex, Eglad, 1998. [11] Bustard, D. W., Dobbi, T. J., ad Carey, B. N., Itegratig Soft Systems ad Object-Orieted Aalysis, IEEE Iteratioal Coferece o Requiremets Egieerig, Colorado Sprigs, Colorado, pp. 52-59, 1996. [12] Wade, S. ad Hopkis, J., A Framework for Icorporatig Systems Thikig ito Object Orieted Desig Seveth CAiSE/IFIP8.1 Iteratioal Workshop o Evaluatio of Modelig Methods i Systems Aalysis ad Desig (EMMSAD 02), Toroto, Caada, May,27-28,2002. [13] Al-Humaida, F., & Rossiter, N., Busiess Process Modellig with OBPM combiig soft ad hard approaches, i Proceedig of 1st Workshop o Computer Supported Activity Coordiatio (CSAC), 6th Iteratioal Coferece o Eterprise Iformatio Systems, Porto,, pp 253-260, 13-14 April,, 2004. [14] Erikkso, H. E., & Peker, M., UML busiess process modellig at work, Joh Wiley ad Sos, New York, 2000. [15] Joh Migers, Combiig IS Research Methods: Towards a Pluralist Methodology, Iformatio Systems Research, 12, 3, Istitute for Operatios Research ad the Maagemet Scieces (INFORMS), pp. 240-259, 2001. [16] D. Platt, Process Modellig ad Process Support Eviromet to Desig Maagemet, Departmet of Civil Egieerig, Faculty of Egieerig, Uiversity of Bristol, UK, 1994. 2229

[17] Daveport, T. h. Process iovatio: Reegieerig work through iformatio techology, Harvard Busiess School Press, Bosto, Mass, 1993. [18] Warboys, Bria, Kawalek, Peter, Robertso, Ia, ad Greewood, Mark, Busiess Iformatio Systems-A process approach, McGraw-Hill, UK, 1999. [19] Svatopluk Štolfa, Ivo Vodrák, Mappig from Busiess Processes to Requiremets Specificatio, Retrieved o 7th Aug, 2008 from 85.255.195.219/cof/esm/esm2006/abstract.pdf [20] Pawso R. & Mathews R., Naked Objects, Joh Wiley ad Sos Ltd, West Sussex, Eglad, 2002. [21] Goodlad, S. ad Hirst, B. Peer Tutorig: A Guide to Learig by Teachig,Lodo: Koga Page; New York: Nickols Publishig, 1989. [22] Garder, H. (1993) Multiple itelligeces: the theory i practice. New York, NY:Basic Books. [23] Miliszewska Iwoa, Ta Grace. Befriedig Computer Programmig: A Proposed Approach to Teachig Itroductory Programmig. Issues i Iformatio Sciece ad Iformatio Techology, volume 4, 277-289., 2007. [24] Hu Xiaohui. Improvig teachig i Computer Programmig by adoptig studet-cetred learig strategies, Chia papers, issue 6. 46-51., 2006. 2230