Regardless of its origin in business reengineering as of [Da93] and [HC93] works, development of BPM (Business Process Management) as a management discipline has always been enriched by solid computer science field contributions, e.g. [Aa98], [Aa03], [AHW03] and [We10], which were decisive to develop required methods and technologies to enable modelling, analysing, simulating, implementing and monitoring operational business processes. In this regard, BPM plays an important role in bridging the gap between business and IT areas in organizations, providing support to strategic alignment of business processes, e.g. [Mo12] and [Ro15]. Indeed, BPM initiatives in organizations are not rarely associated to EIS (Enterprise Information Systems) adoption or adjustment, aiming to implement designed business processes, fully or partly automated, in order to enhance effectiveness of process improvement programs. This demands the deployment of process-oriented EIS that support business processes (e.g. [He14]) and are able to quickly adapt to changes in operations in a constantly changing business scenario. (e.g. [WSR09]). In this context, existing EIS can be considered in a broader view, e.g. [Xu11] and [AW04], and may or may not be originally designed from a process-driven perspective.
To design such systems, business processes must be taken into account during RE (Requirements Engineering) phase. This may involve deriving software requirements from business process models and proceeding with regular software development processes, but might also embrace adoption of commercial BPMS (Business Process Management Suite). In fact, according to [DAH05], any information system developed under process-driven orientation can be considered a PAIS (Process-Aware Information System), which suggests there may be different RE approaches to address processoriented EIS. In any way, RE and BPM fields must be interrelated to address the design of EIS aligned to business context, e.g. [HKW14].
Discovering varied RE approaches to derive EIS may contribute to strategic alignment of organizations, since they can guide different strategies of PAIS adoption that can be considered to orient EIS to business processes. The main motivation of this study is to provide the state-of-the-art academic overview on how RE can be applied to derive EIS from BPM. The study describes a systematic literature review on this research topic, presenting the review phases and discussion of the results found.
To perform the SLR (Systematic Literature Review), guidelines from [KC07] were followed. To start planning the review, an initial mapping study was performed in order to expand research keywords used to search for primary studies. "Business process management" and "requirements engineering" main keywords were used as a starting point to search for article title, abstract and keywords in Scopus database [Sc17]. On 12/09/2017, this search produced 55 document results, whose metadata were exported to VOSviewer software tool [EW10]. A co-occurrence network of keywords was then created, based on the number of publications in which two keywords occur together in the bibliographic data. The keywords co-occurrence map can be viewed in Fig. 1. This initial mapping study resulted 395 keywords which were later analysed to compose SLR search queries. SLR protocol items were defined as follows:
•
Objective: to provide a background in order to appropriately position research activities in the field of BPM and RE for EIS.
requirements model, requirements specification, software engineering, software requirement, strategic alignment, system engineering, system requirement.
• Sources selection criteria definition: abstract and citation databases of peerreviewed literature. • Source list: Scopus [Sc17] and Web of Science [We17].
• Primary studies' language: English. These predefined approach types were adapted from the common framework for the concept of BPM in an EIS perspective from [MMT08].
To conduct the review, a search strategy was defined based on the relevance of each keyword and synonym in the keywords co-occurrence map identified during the initial mapping study. Expanded search queries for each source were developed as follows:
• Scopus: TITLE-ABS-KEY ( ( "business process management" OR "BPMN" OR "BPMS" ) AND ( ( "requirement*" AND ( "business it alignment" OR "strategic alignment" OR "information system*" OR "system* engineering" OR "software engineering" ) ) OR ( "system* requirement*" OR "software requirement*" OR "requirement* engineering" OR "requirement* analysis" OR "requirement* elicitation" OR "requirement* gathering" OR "requirement* management" OR "requirement* model*" OR "requirement* specification*" ) ) ) AND ( LIMIT-TO ( DOCTYPE , "cp " ) OR LIMIT-TO ( DOCTYPE , " ar " ) ) AND ( LIMIT-TO ( PUBYEAR , 2018 ) OR LIMIT-TO ( PUBYEAR , 2017 ) OR LIMIT-TO ( PUBYEAR , 2016 ) OR LIMIT-TO ( PUBYEAR , 2015 ) OR LIMIT-TO ( PUBYEAR , 2014 ) ) AND ( LIMIT-TO ( LANGUAGE , "English " ) ).
• Web of Science: (TS=(( "business process management" OR "BPMN" OR "BPMS" ) AND ( ( "requirement*" AND ( "business it alignment" OR "strategic alignment" OR "information system*" OR "system* engineering" OR "software engineering" ) ) OR ( "system* requirement*" OR "software requirement*" OR "requirement* engineering" OR "requirement* analysis" OR "requirement* elicitation" OR "requirement* gathering" OR "requirement* management" OR "requirement* model*" OR "requirement* specification*" ) ))) AND Languages: (English) AND
Review results encompasses not only accepted papers, but rejected ones which could be grouped into subcategories for each exclusion criterium, based on similarity of subjects covered. Tab. 1 presents SLR results for rejected papers. analysis practices such as those from BABOK (a guide to the Business Analysis Body Of Knowledge) [II15], which consider multiple perspectives to look at organizations (business, systems, data, processes, technology, among other ones), usually also related to business and IT strategic alignment. Enterprise architecture is a large and important research field, although not necessarily addressing information systems development. Additionally, it cannot be considered a process-oriented approach, even though some BPM methods and techniques may be eventually applied.
• EC1.2 -management initiative: application of BPM as a management discipline only, focused on business process improvement not related to information systems development. Some BPM tools such as business process simulation might be used, but not related to derivation of software requirements to develop enterprise information systems that support business process operations. Management initiatives may also apply software engineering methodologies such as elicitation, quality assurance or agile methods on business process discovery.
• EC1.3 -process mining: application of techniques to extract business process models from event logs of existing information systems. Process mining initiatives demands availability of enterprise information systems and proper data, which allows business process models to be derived from them. However, process mining might be used for system maintenance or continuous requirements engineering [AW04].
• EC2.1 -BPMN for software design: application of process modelling techniques or notations during software design or deployment, without focusing on requirements specification from business process models.
• EC2.2 -modelling languages: application of ontologies to create, evaluate or compare alterative languages or notations for process modelling, without focusing on requirements specification.
Even though EC6 had been included in the search strings, during full-text analysis 2 papers had invalid document types and were then rejected. No similar condition occurred with EC3 and EC5 which resulted in no rejections at this phase. EC4 was also applied during study quality assessment based on full-text availability, and 9 papers were rejected due to this criterium.
Analysis of accepted papers could also group similar papers into subcategories of each approach type. Tab. 2 presents SLR results for accepted papers. Specific approach types are described as follows: • AT1.1 -BPM4EIS: this approach type distinguishes itself by application of BPM as a management initiative related to the development of enterprise information system able to support business processes and improve business and IT alignment [VSP08].
Papers using this approach usually present business process models in different levels of abstractions (notably business and IT levels) and describe how process models are considered during information system requirements specification.
• AT1.2 -BPMN4EIS: application of business process modelling to support requirements engineering and information systems modelling, as described in BABOK [II15], SWEBOK (guide to the Software Engineering Body Of Knowledge) [BF14] or REBoK (Requirements Engineering Body of Knowledge ) [Pe13] guides. Papers using this approach also cover ontologies for business process application to requirements engineering, notably the development of methodologies, languages and heuristics to derive software requirements from business process models (such as UML artefacts from BPMN models, for example).
• AT1.3 -goal2BPMN: emergent research field with applies goal frameworks such as KAOS (Knowledge Acquisition in autOmated Specification) [Da97] and i* [Fr16] for GORE (Goal-Oriented Requirements Engineering) [La01], which develops software requirements engineering oriented to organizations strategic goals, in order to improve business and IT strategic alignment. Papers using this approach use business process modelling notations (such as BPMN) to formalize software requirements.
• AT2.1 -extended BPMN: this approach type benefits from native extensibility of BPMN to improve requirements specification. It acknowledges notation power to capture software requirements and extends its elements to model specific requirements, notably non-functional ones [BD11], such as those related to security, risk, ubiquity, and others. Usually, extension proposal includes methods to convert models using extended notations to software design/deployment artefacts.
• AT2.2 -BPMN2services: semiautomatic generation of information systems from process-based requirements using SOA. This approach type is built upon MDA (Model-Driven Architecture) paradigm [OM14], in which software design is independent on the technology or programming language and involves CIM
Requirements Engineering approaches to derive Enterprise Information Systems from Business Process Management: a systematic literature review 267
(Computation Independent Model) to PIM (Platform Independent Model) transformations.
• AT3.1 -PAIS: academic research field for business process automation [WRR08], this approach type covers the development of technologies such as ACM (Adaptive Case Management) [HK11] and others that allow automatic generation of processoriented applications.
• AT3.2 -BPMS: commercial suites for application generation from automation of business process modelled in the software platform [Re06].
Analyzing
It is worth mentioning that findings on this study address an interdisciplinary research gap between operations management and computer science fields, by bringing together strategic alignment approaches that can be considered from BPM or RE perspectives.
Referring to the research question, this SLR provided a comprehensive academic overview on how RE can be applied to derive EIS from BPM, exposing the various existing approaches in the current academic literature. Content analysis of selected papers allowed identification of six subcategories of approach types, extending initial data extraction criteria and unveiling emergent research fields that had not been considered previously. For that matter, it can be claimed that the objective defined in the SLR protocol has been successfully fulfilled.
Also, discovering different approach types that vary depending on software development automation degree they demand, suggests that PAIS adoption for EIS goes beyond a build-or-buy decision, but may embrace different strategies for different organizations (e.g. varying on size, BPM maturity, business process' characteristics, IT budget or other aspects). These findings can guide further research activities, providing a base framework to support surveying RE approaches to derive EIS from BPM in the practice of organizations.
It is important to remark the contribution of initial mapping study based on keywords cooccurrence map to the research methodology, which expanded the research search and enabled such broad overview as a result. However, the SLR results offer only a first approach to the research topic and therefore a wide but limited knowledge about the subject. Further research including a detailed analysis of the literature must be carried out in order to reference findings and enable proposal of a research framework.