Exploring Requirements Engineering for Machine Learning via a Product Case Study

Requirements problems are widely acknowledged to have impact on the effectiveness of the software development process. In order to improve the requirements engineering (RE) process and to reduce requirements problems, Sommerville et al. [1] have developed a requirements maturity model. Literature shows that the measurement process, designed in this model, is very confused and can lead organizations to incorrect results. This is because the measurement process is ambiguous and no strategic and systematic approach is used to decide different scores for various RE practices.The objective of this paper is to propose a measurement instrument for Sommerville et al.’s model to effectively measure the maturity of the RE process. The main purpose of proposing this measurement instrument is to develop better ways to assist practitioners in effectively measuring the maturity of the RE process. This instrument provides a very practical structure with which to measure the maturity of the RE process. I have tested this instrument in one case study where only one category of RE process, i.e. ‘requirements elicitation’ was used as an exemplar. The case study results show that the measurement instrument has potential to assist practitioners in effectively measuring the maturity of ‘requirements elicitation’ category of the RE process. Thus, I recommend organizations trial this instrument for other categories of RE process in order to further evaluate its effectiveness in the domain of RE process.

Adoption of the most suitable requirements engineering (RE) process and selection of the most appropriate RE techniques for a given project is a common challenge faced by industry. This paper presents a methodology for RE process development for a given project. The novelty of this work is four-fold. First, a RE process knowledge base (REPKB) is established and it helps during RE process development. Second, a decision support mechanism is provided during RE process development. Third, this methodology uses three components: process building blocks, standard templates of the RE process and development guidelines, to help process development. Fourth, it explicitly links project characteristics with RE process development so that the most suitable RE process can be developed. The theory behind this methodology is described in the paper. Preliminary results suggest that the methodology is of valuable help to requirements engineers during RE process development and process customization.

Adoption of the most suitable requirements engineering (RE) process and selection of the most appropriate RE techniques for a given project are challenging issues for industry. In order to alleviate the problems, a framework for RE process development (FRERE) was developed based on our research in the last six years. The overall architecture of the framework and a brief case study are presented in the paper. The framework addresses several core issues of the RE process: engineering issues, domain issues, characteristics of the software project, and issues of conformity to software process standards. FRERE consists of three parts: (1) The RE process knowledge base (REPKB) which contains RE process knowledge, (2) methodologies which provide guidance to develop a customized software process model in an iterative way; and (3) assessment models which provide information about the suitability of the newly developed RE process model for the software project under development.

Context: Requirements Engineering (RE) is one of the key processes in software development. With the advent of agile software development methods, new challenges have emerged for traditional, prescriptive maturity models aiming to support the improvement of RE process. One of the main problems is that frequently the guidelines prescribed by agile approaches have to be adapted to a project’s context to provide benefits. Therefore, it might be naive to believe that it is possible to propose a prescriptive method of RE process improvement that will suit all agile projects without any alteration. Objective: The aim of the paper is to evaluate a hybrid approach to assessing the maturity of agile RE (REMMA), which combines elements of prescriptive and problem-oriented improvement methods. Method: The usefulness, ease of use, and cost-effectiveness of REMMA were investigated through a case study performed in one of the biggest software houses in Central Europe. Results: The results of the case study suggest that the method seems easy to use, affordable, and is perceived as a useful tool to support the process of improving RE practices in agile projects. Its feature of taking into account the dependencies between practices and the necessity to adapt them to a certain project context was regarded as well suited for the agile context. Conclusions: REMMA, which includes two main components: a maturity model for agile RE (a set of state-of-the-art agile RE practices) and an assessment method that makes it possible to evaluate how well the agile RE practices are implemented, seems to be a useful tool supporting improvement of RE in agile projects.

The volume of software development outsourcing is growing enormously owing to the associated benefits of outsourcing and limitations of organizations. However, a large number of the projects outsourced for software development are failed to achieve anticipated results. In most of such cases, the reasons for failure are traced back to the Requirements Engineering (RE) process. In spite of this state of affairs, adequate efforts have not been made to avoid the failure of outsourced software development projects due to RE process issues. This research has been conducted with the aim of cultivating the RE process for software development outsourcing by identifying the significant RE practices for this process. The study is based upon Sommerville and Sawyer's RE practices for six key areas of the RE process. A survey research method has been used to collect data from 108 practitioners belonging to 18 national and multinational organizations associated with software development outsourcing. After analysis of the obtained data, we have identified the RE practices which are significant in order to improve the RE process for software development outsourcing and hence to achieve desired results.

Selecting a suitable Requirements Engineering (RE) process is usually based on personal preferences or existing company practices rather than on the characteristics of the project at hand (project situation). Feature-oriented software development is the overall process of developing software systems in terms of their features. The Software Product Line (SPL) approach is a paradigm for systematic reuse of software products, and a Software Factory is a SPL aimed at the industrialization of software development. Based on the notion that a software/RE process can be developed via an engineering process, this research aims to provide a feature-based RE process factory to develop RE processes based on project situations. In our approach, the project situation is modeled as the problem domain (resulting in a situation model). A feature model can encapsulate all the features in an SPL; therefore, the variations and commonalities among RE processes can be represented in the form of a feature model, considered as a model of the solution domain. A mapping for translating the situation model to the RE process feature model is proposed with the specific aim of promoting traceability and rationality in the selection of RE process features.

Requirement engineering (RE) process clear description can be an important factor for guiding the team members involved in the RE process, which may help organizations not exceeding the estimated schedule and budget for the software project. There can be many reasons of not having efficient RE process such as changing situations among the organizations involved in the RE process. It is certainly one of the various other reasons of having non-efficient RE process. Due to these changing situations the RE process customization accordingly should be performed. As part of our current research project, where we are investigating situational RE in Global Software Development (GSD), first we need to identify the factors which may results in changing situations. This paper we have critically reviewed the domain of “standards and models”, and identified the initial list of situational factors from them. These standards and models are related to software engineering, which directly or indirectly discusses RE process. We have adopted the constant comparison and memoing techniques of the Grounded theory for identification of unique initial list of the situational factors. This initial list is a significant consideration for a comprehensive list of situational factors affecting RE process in GSD.

Software industry is adopting global software development (GSD) due to its potential to produce quality products at a lower cost. However, the GSD firms face many challenges that make development activities more complicated, especially related to the requirements engineering (RE) process. The objectives of this article are to investigate and prioritize the barriers faced by the GSD organizations during the RE process. First, we identified 17 barriers related to the RE process in the GSD projects. Next, the identified barriers were further validated with real‐world GSD practitioners using a questionnaire survey. Finally, we applied the analytical hierarchy process to prioritize the investigated barriers with respect to their significance for the RE process in the GSD domain. The results show that coordination is the most significant barrier category for the RE process in GSD projects. Lack of standard and procedure for RE in GSD, lack of synchronized communication infrastructure, and lack of mutual understanding between the overseas RE teams are also high‐ranked barriers for the RE process in GSD. The authors believe that the findings of this study will assist practitioners and researchers in developing effective strategies and plans for the successful implementation of the RE process in the GSD context.

Requirement Engineering (RE) process during product development produces core artifacts to be analyzed during domain analysis phase of Software Product Line Engineering (SPLE). The product RE process can be a Software Requirement Specification (SRS) document, use case models or other artifacts containing both functional and nonfunctional requirements of the product. The analysis of these artifacts is a time consuming and error-prone process when performed manually. There is also a need for creating consistent and complete set of non-functional requirements from user-specific individual projects in SPL. Therefore, we propose a methodology to create Domain Non-Functional Requirements (DNFRs) from Product Non-Functional Requirements (PNFRs) using model driven approach. Two model transformations are performed to automate the process of DNFR creation. First transformation converts PNFR into Product Line Non-Functional Requirements (PLNFR) using a PNFR meta model. The second transformation converts the PLNFR into DNFR using DNFR meta model. This transformation exploits commonality and variability among the products in terms of their non-functional requirements. The resulting DNFR can serve as a useful baseline for domain analysis in SPLE.

Software requirement engineering (RE) process is one of the most important phases of the software development life cycle (SDLC) that affect its overall success, since it has a significant role in determining the software quality and software development process effectiveness, due to the increased consideration for Software requirement engineering process and process improvements at the SDLC life cycle. Several standards and RE improvements model were constructed on the aim at helping organizations in improving and assessing their RE process, however they suffer from several problems that limits their acceptance by the organization that are interest in the RE process improvement.The paper proposed a new requirement engineering process capability maturity model based on the capability maturity model integration for development (CMMI-DEV). The intention is to provide a generic maturity model that based on international standards and literature on software RE to help developing the area of RE process improvement.

Requirement Engineering (RE) plays an integral role throughout the process of software development. Requirement identification and prioritisation are the foremost phases of the RE process. Latest RE research work uses Machine Learning (ML) algorithms to tackle RE problems such as identifying requirements and assigning priorities to requirements, which have given better results than that of traditional natural language processing methods. An adequate understanding of these ML methods, however, is still lacking. The aim of this study is to understand which of the ML algorithms is likely to classify and prioritise the requirements efficiently and how they can be evaluated. It is observed that the current approaches are having constraints of scalability and complexity. Different methods used for the text preprocessing of requirements from SRS and user reviews are also proposed. 6 different ML algorithms and 6 different prioritisation algorithms, which are most common methods, are found. The most popular performance parameters used are accuracy, precision and recall. The limitations of these ML approaches are irrespective of dependency of requirements, priorities are assigned to requirements, the results with respect to scalability and speed is inferior.

Purpose This paper presents a recommendation on how one requirement elicitation technique supports the other techniques in defining system requirement for a case-based system. A case-based lesson planning system aims to assist teachers in constructing quality lesson plans through its cycle which begins with case retrieval. To retrieve relevant lesson plans, appropriate inputs should be used and the intended output needs to be identified via suitable requirement elicitation techniques. The use of a single technique might result in inadequate requirement specification, thus affecting the quality of the output requirements as well as quality of the final information system. Design/methodology/approach Requirement elicitation was carried out in three phases: phase I involved document review, phase II was an interview and phase III used a survey. Respondents of the study comprised experienced teachers as well as new teachers. This research used both qualitative and quantitative approaches to answer the research questions, which involved semi-structured interviews, document review and survey to collect the relevant data. Documents were reviewed by analysing lesson plans from three different countries. In addition, a review of lesson plans prepared by teachers and the standard syllabus were carried out. Findings from the document review were used in structured interviews using a teach-back technique, sorting and matrix of attribute-values. A questionnaire was then constructed based on the interviews and document review. Findings The findings of this initial study, as part of a larger research investigation, would help in knowledge modelling and representation. This will contribute to effective case retrieval via good design of the system input and output. The study identifies important elements of a lesson plan according to their ranking. Keywords that were used by teachers as input for retrieval were identified together with the expected output. Research limitations/implications The main goal of requirement elicitation is to specify complete and detailed requirements of the proposed system. There are two main types of requirement: functional and non-functional requirements. This paper only focuses on functional requirements – specifically case retrieval with appropriate input and output. Practical implications Various requirement engineering (RE) techniques can be applied in different phases of requirement elicitation. Suitable technique should be chosen at different phases of RE, as it is important for triangulation purposes. Incomplete RE will affect the modelling part of system development, and, thus, affect the design and implementation of an information system. Social implications Software engineer or anybody involved in system development should plan accordingly for the RE process. They should be creative and reasonable in selecting suitable RE techniques to be applied. Originality/value This study aims to gain understanding of the various aspects of lesson planning. Crucial knowledge in lesson planning that was gathered from the elicitation phase is modelled to have a good understanding of the problems and constraints among teachers. The findings of this initial study, as part of a larger research investigation, would help in knowledge modelling and representation. This will contribute to effective case retrieval via a good design of the system input and output.

The Requirements Engineering (RE) process is crucial for the success of software projects, as it establishes the foundation for developing efficient systems aligned with the needs of end users and stakeholders. In this context, soft skills play a fundamental role, being as valued as technical skills due to their impact on the success of software projects. This article aims to map and categorize the soft skills commonly cited in academia and industry concerning the RE process in software development in light of the perspectives of the CHAOS Report: Beyond Infinity. For this, a pilot study was carried out to validate the interview protocol that will be conducted. The main contributions of this study were identifying essential soft skills for RE practices, analyzing soft skills concerning CHAOS Report: Beyond Infinity, and categorizing the mapped skills.

[Context and motivation] The increasingly complex business and development environment brings challenges to IT system development and requirements engineering (RE) activities. [Question/problem] The goal of the case study was to investigate what the challenges of a customer organization’s RE process are in the outsourced development environment and what demands these challenges bring to RE process development. The case study was conducted in a Finnish insurance company. [Principal ideas/results] The results are based on 17 interviews and the analysis of 15 large projects. The case study indicates that one of the biggest challenges is to develop business and IT as a whole. When combining business process and IT system development, requirements are an important tool. Another critical challenge in the outsourced environment is that the RE process is distributed between the customer organization and the supplier. Furthermore, highly integrated IT systems and enterprise architecture bring demands to RE process. [Contribution] The paper describes a complex environment in which the customer organization develops IT systems, and systematically defines challenges related to the RE process.

The anticipated benefits of Software Development Outsourcing (SDO) are not achieved in case of several projects because of the issues that emanate from Requirements Engineering (RE) process. This research work presents a Requirements Engineering Practices (REP) model to cope with the customarily occurring issues of the RE process for SDO. To formulate the model, five workshops have been conducted and Root Cause Analysis has been performed by considering 43 commonly occurring SDO RE process issues, and 147 RE practices to tackle the issues. To discover the root causes for commonly transpiring issues, 5- Whys technique has been employed. The relevant RE practices that can be used to deal with the root causes, have been endorsed by applying Brainstorming technique. For the 43 frequently occurring issues, 89 root causes have been discovered. Afterwards, 124 relevant RE practices have been recommended to eradicate the root causes and hence to address the corresponding issues. Thus, REP model postulates the root causes for commonly occurring issues of the SDO RE process, maps the root causes to the relevant best RE practices to address the corresponding issues. The model has been evaluated by an expert panel and evaluation results have been analysed through Inter-Rater Reliability analysis and Analysis of Means. The REP model supports the RE process for SDO by i). evading the adoption of random and inappropriate RE practices for dealing with the common issues of the process, ii) helping to attain the expected benefits of SDO.