[pt] A engenharia de sistemas habilitados em Machine Learning (ML) bem-sucedidos apresenta vários desafios, tanto do lado teórico quanto prático. Entre esses desafios estão como abordar eficazmente às expectativas irrealistas das capacidades de ML por parte de clientes, gestores e até mesmo outros membros da equipe de desenvolvimento, e como ligar o valor do negócio às atividades de engenharia e ciência de dados compostas por equipes interdisciplinares. Nesta tese, estudamos o estado da prática e da literatura da engenharia de requisitos para ML para propor PerSpecML, uma abordagem baseada em perspectiva para especificar sistemas habilitados para ML que ajuda os profissionais a identificar quais atributos, incluindo componentes de ML e não-ML, são importantes para contribuir para a qualidade geral do sistema. A abordagem envolve a análise de 60 preocupações relacionadas a 28 tarefas que os profissionais normalmente enfrentam em projetos de ML, agrupando-as em cinco perspectivas: objetivos do sistema, experiência do usuário, infraestrutura, modelo e dados. Juntas, essas perspectivas servem para mediar a comunicação entre gestores de projeto, especialistas de domínio, designers, engenheiros de software/ML e cientistas de dados. A criação da PerSpecML envolveu uma série de validações realizadas em diferentes contextos: (i) na academia, (ii) com representantes da indústria e (iii) em dois estudos de casos industriais reais. Como resultado das diversas validações e melhorias contínuas, PerSpecML se destaca como uma abordagem promissora, preparada para impactar positivamente a especificação de sistemas habilitados para ML, ajudando particularmente a revelar componentes-chave que, de outra forma, teriam sido perdidos sem o uso da PerSpecML. / [en] Engineering successful machine learning (ML)-enabled systems poses various challenges from both a theoretical and a practical side. Among those challenges are how to effectively address unrealistic expectations of ML capabilities from customers, managers and even other team members, and how to connect business value to engineering and data science activities composed by interdisciplinary teams. In this thesis, we studied the state of the practice and literature of requirements engineering (RE) for ML to propose PerSpecML, a perspective-based approach for specifying ML-enabled systems that helps practitioners identify which attributes, including ML and non-ML components, are important to contribute to the overall system s quality. The approach involves analyzing 60 concerns related to 28 tasks that practitioners typically face in ML projects, grouping them into five perspectives: system objectives, user experience, infrastructure, model, and data. Together, these perspectives serve to mediate the communication between business owners, domain experts, designers, software and ML engineers, and data scientists. The conception of PerSpecML involved a series of validations conducted in different contexts: (i) in academia, (ii) with industry representatives, and (iii) in two real industrial case studies. As a result of the diverse validations and continuous improvements, PerSpecML stands as a promising approach, poised to positively impact the specification of ML-enabled systems, particularly helping to reveal key components that would have been otherwise missed without using PerSpecML.

Understanding and communicating user requirements in a software requirement analysis effort is very important. Misunderstandings of user requirements between stakeholders will cause problems in terms of satisfying their needs, reduction of defects, cost and schedule during the software development process. This dissertation presents a new technique that has the ability to represent the mental models of the user, developers, project managers and sponsors (collectively referred to as ?stakeholders?) as network representations. The requirements are modeled as nodes and the perception of stakeholders is modeled as the interrelationships (links) among the requirements. The requirements are first extracted from a requirements document. The requirements are then categorized into related groups as perceived by each stakeholder. The relatedness (proximity) data collected from the categories is then fed into the Pathfinder generation program that results in the generation of pathfinder network(PFNETs). The PFNETs of stakeholders are then compared for similarities/dissimilarities using a graph similarity metric referred to as a correlation coefficient. During preliminary research work, this technique was applied to multiple student projects with real customers at Mississippi State University (MSU), and to a project at NORTEL, Dallas, Texas with encouraging results. This research was successful in identifying duplicate, ambiguous and misunderstood requirements. The next step was to validate this technique on small-scale and medium-scale projects in an industrial setting. During the summer of 2003, the National Science Foundation (NSF) and AmerInd Inc. jointly sponsored a collaborative industry-university research effort to validate the proposed technique. It was found that this technique is easy to apply and useful to gauge an overall understanding of requirements and identify potentially misunderstood requirements for small and medium scale projects. This technique scaled well from a small-scale project with two stakeholders to a medium-scale project with a little over one hundred requirements and six stakeholders. The correlations helped focus discussions on the requirements that were potentially misunderstood among stakeholders. Duplicate, misunderstood and ambiguous requirements were identified during the facilitation sessions. We also present a new technique that applies information theory-based software metrics to measure consensus about requirements among stakeholders.

A lot of research is done in requirements engineering and testing but often the extensive literature is missing on defining good methods for linking functional requirements with test cases. Most of the delays occurring in the software development projects are because of incomplete or inaccurate functional requirements. The two main goals of our project are to achieve a successful software project by First, to design a template, which will merge functional requirements with test cases and second is to find the benefits of the aligning requirements to test cases. Changing, updating and tracing the requirements during the development of the project is not an easy task. The main reason for project failure is due to possibility of not fulfilling specified project requirements, so one way to solve this problem is to merge functional requirements with test cases. Thus removes the need of creating a separate requirements document, which will improve the traceability process between requirements and testing, thus leads to high quality and efficient development. The template helps us to drive a successful project by identifying the issues at an earlier stage of the development cycle.

Building a shared understanding of non-functional requirements (NFRs) is a known but understudied challenge in requirements engineering, primarily in organizations that adopt continuous software engineering (CSE) practices. During the peak of the COVID-19 pandemic, many CSE organizations complied with working remotely due to the imposed health restrictions; some continued with remote work while implementing business processes to facilitate team communication and productivity. In remote CSE organizations, managing NFRs becomes more challenging due to the limitations to team communication coupled with the incentive to deliver products quickly. While previous research has identified the factors that lead to a lack of shared understanding of NFRs in CSE, we still have a significant gap in understanding how CSE organizations, particularly in remote work, build a shared understanding of NFRs in their software development. This thesis presents a study that explores how a remote CSE organization builds a shared understanding of NFRs. We conducted a six-month case study of a remote CSE organization using ethnography-informed methods and methods from grounded theory. Through thematic analysis of our qualitative data from interviews and observations, we identify some practices in building a shared understanding of NFRs, such as validating NFRs through feedback. In addition, we identified some of the impediments to building a shared understanding of NFRs in the organization, such as gaps in communication and the limited understanding of customer context. Furthermore, we conducted member-checking interviews to validate our findings for relevance and to gain additional insights on the shared understanding of NFRs within the organization. The collaborative workspace the organization uses for remote interaction is Gather, which simulates physical workspaces, and which our findings suggest allows for informal communications instrumental for building shared understanding. As actionable insights, we discuss our findings in light of proactive practices that represent opportunities for software organizations to invest in building a shared understanding of NFRs in their development. / Graduate

We propose a Capabilities-based approach for constructing complex emergent systems such that they are change-tolerant, and the development effort promotes change-reduction. The inherent complexity of software systems increases their susceptibility to change when subjected to the vagaries of user needs, technology advances, market demands and other change inducing factors. Despite the inevitability of change, traditional Requirements Engineering strives to develop systems based on a fixed solution. This is a mostly unsuccessful approach as evidenced by the history of system failures. In contrast, we utilize Capabilities — functional abstractions that are neither as amorphous as user needs nor as rigid as system requirements — to architect systems to accommodate change with minimum impact. These entities are designed to exhibit desirable characteristics of high cohesion, low coupling and balanced abstraction levels. Capabilities are generated by a two-phased process called Capabilities Engineering. Phase I mathematically exploits the structural semantics of the Function Decomposition graph — a representation of user needs — to formulate change-tolerant Capabilities. Phase II optimizes these Capabilities to conform to schedule and technology constraints. Results from an empirical evaluation of a real-world Course Evaluation System indicate, with statistical significance, that a Capabilities-based design is more change-tolerant than a requirements-based design. In addition, we observe that the use of the CE process inherently reduces change, otherwise generated, during the regular development effort. Empirical analysis on the change-requests of Sakai, a complex emergent system, supports this claim. Finally, we observe that the process of Capabilities Engineering assists in pre-requirement specification traceability by bridging the complexity gap between the problem and solution spaces. / Ph. D.

Many software systems fail as a direct consequence of errors in requirements analysis. Establishing formal metrics early in the design process, using attributes like critical parameters, enables designers to properly assess software success. While critical parameters alone do not have the potential to drive design, establishing requirements tied to critical parameters helps designers capture design objectives. For the design of interactive systems, the use of scenario-based approaches offers natural user centricity and facilitates knowledge reuse through the generation of claims. Unfortunately, the requirements-analysis phase of scenario-based design does not offer sufficient built-in and explicit techniques needed for capturing the critical-parameter requirements of a system. Because success depends heavily on user involvement and proper requirements, there is a crucial need for a requirements-analysis technique that bridges the gap between scenarios and critical parameters. Better establishing requirements will benefit design. By adapting task-modeling techniques to support critical parameters within the requirements-analysis phase of scenario-based design, we are able to provide designers with a systematic technique for capturing requirements in a reusable form that enables and encourages knowledge transfer early in the development process. The research work presented concentrates on the domain of notification systems, as previous research efforts led to the identification of three critical parameters. Contributions of this work include establishment of a structured process for capturing critical-parameter requirements within a user-centric design framework and introduction of knowledge reuse at the requirements phase. On one hand, adapting task models to capture requirements bridges the gap between scenarios and critical parameters, which benefits design from user involvement and accurate requirements. On the other hand, using task models as a reusable component leverages requirements reuse which benefits design by increasing quality while reducing development costs and time-to-market. / Master of Science

In recent years, the requirements engineering community has proposed a number of models for the generation of a well-formulated, complete set of requirements. However, these models are often highly abstract or narrowly focused, providing only pieces of structure and parts of guidance to the requirements generation process. Furthermore, many of the models fail to identify methods that can be employed to achieve the activity objectives. As a consequence of these problems, the requirements engineer lacks the necessary guidance to effectively apply the requirements generation process, and thus, resulting in the production of an inadequate set of requirements. To address these concerns, we propose the expanded Requirements Generation Model (x-RGM), which consists of activities at a more appropriate level of abstraction. This decomposition of the model ensures that the requirements engineer has a clear understanding of the activities involved in the requirements generation process. In addition, the objectives of all the activities defined by the x-RGM are identified and explicitly stated so that no assumptions are made about the goals of the activities involved in the generation of requirements. We also identify sets of methods that can be used during each activity to effectively achieve its objectives. The mapping of methods to activities guides the requirements engineer in selecting the appropriate techniques for a particular activity in the requirements engineering process. Furthermore, we prescribe small subsets of methods for each activity based on commonly used selection criteria such that the chosen criterion is optimized. This list of methods is created with the intention of simplifying the task of choosing methods for the activities defined by the x-RGM that best meet the selection criterion goal / Master of Science

In den 80er und 90er Jahren hatten große Anwendungssysteme in Unternehmen einige hundert bis tausend Anwender. Der Software-Entwicklungsprozess für diese Anwendungen war innerhalb der Unternehmen relativ klar geregelt. Fachinformatiker und Fachabteilungen standen einander dabei gegenüber. Oft wurden auch externe Fachleute und Komponentenlieferanten integriert. Entwicklungsmethoden und Werkzeuge waren auf diese Situation ausgelegt. Seit dieser Zeit haben wesentliche Veränderungen stattgefunden. Internettechnologien haben neue Klassen von Applikationen ermöglicht, die wie folgt charakterisiert werden können: - Die Applikationen sind kooperativ (unternehmensübergreifend). Nicht selten sind 20-50 oder mehr Unternehmen z. B. bei Zulieferketten beteiligt. - Eine eigene Klasse bilden mandantenfähige Systeme sowie Business-to-Consumer Systeme (B2C) bei denen sehr große Nutzerzahlen konnektiert werden. - Die Entwicklungszeiten liegen im Bereich von Monaten statt Jahren für eine erste Bereitstellung einer Basislösung. - Die Systeme werden inkrementell unter starker Anwenderbeteiligung bis hin zur Endbenutzerbeteiligung weiterentwickelt. (...)

Syfte – Idag saknas det ett standardiserat arbetssätt för att hantera kundkrav samtidigt som det finns stora utmaningar i hanteringen av dem. Därmed är syftet med denna studie:  Att definiera ett ramverk för hantering av kundkrav inom Grafisk kommunikation. För att kunna besvara syftet i studien har en frågeställning tagits fram som utgångspunkt för studien. Hur kan Requirements Engineering, sociala relationer och kommunikation tillsammans bidra till att förbättra hantering av kundkrav inom arbetet med Grafisk kommunikation? Metod – För att besvara frågeställningen har studien arbetat med datainsamlingsmetoden, semistrukturerade djupintervjuer. Intervjuerna har genomförts  för att få fram primärdata. Metoden har valts ut för att säkerställa en kvalitativ ansats i studien. I studien har även litteraturstudier genomförts som en viktig del med syfte att ge en bakgrund kring befintliga teorier, metoder och verktyg. Resultat – Studien är av tvärvetenskaplig art och dess resultat behandlar områden så som Requirements Engineering, Grafisk kommunikation, sociala relationer och kommunikation. Områdena har studerats för att se hur de kan samverka för att säkerställa en högre och jämnare nivå av de projektresultat som reklambyråer och konsulter inom reklam- och kommunikationsbranschen levererar. Studien resulterade i ett ramverk för hantering av kundkrav inom Grafisk kommunikation. Ett ramverk beståendes av två olika kategorier som berör både den praktiska processen samt den social kommunikativ dimensionen kopplat till hantering av krav. Studiens resultat beskriver också vilka element som bör finnas i en iterativ designprocess samt hur ramverket kan användas tillsammans med en sådan typ av utvecklingsprocess. Implikationer – Studien anses kunna bidra till forskningsvärlden då dess resultat har öppnat upp nya dörrar för ytterligare områden att applicera Requirements Engineering på. För näringslivet och det vardagliga arbetet inom reklam- och kommunikationsbranschen har resultatet visat att ramverk för hantering av kundkrav inom Grafisk kommunikation bör kunna medverka till att stödja verksamheters arbetsprocesser och säkerställande av en mer jämn kvalitetsnivå. Utvecklingsprocesser och dess slutresultat skulle ha större möjlighet att bli framgångsrika, samtidigt som användningen av projektets resurser med stor sannolikhet skulle kunna effektiviseras genom den struktur som ramverket för med sig. / Purpose – There is no standardized approach for how to managing Customer Requirements, while there are significant challenges in managing them. Thus, the aim of this study is: To define a framework for the managing of Customer Requirements in Graphic Communication. In order to find the purpose of the study, a research question was developed and used as basis for the study. How can Requirements Engineering, Social Relationships and Communication improve the managing of Customer Requirements in Graphic Communication? Method – The study worked is based on one data collection method, semi-structured interviews, in order to answer the research question. The interviews were conducted to obtain the primary data. The method were selected to ensure a qualitative approach in the study. The study has also literature studies conducted as an important element in order to provide a background on existing theories, methods and tools. Findings – The study is of an interdisciplinary nature and its results concern areas such as: Requirements Engineering, Graphic Communication, Social Relationships and Communication. These areas have been studied to see how they can interact and ensure a higher and uniformed level of project results, as advertising agencies, consultants in communication and advertising industry delivers. The study resulted in a framework for managing Customer Requirements in Graphic Communications. A framework consisting of two different categories that can affect both practical processes and a social communicative dimension linked to the handling of requirements. The study’s result describes the elements that should be in an iterative design process and how the framework can be used in conjunction with such development. Implications – The study is considered to contribute to the research world, since its results have opened up new doors for additional areas for applying the Requirements Engineering. For business and everyday-work in the advertising and communication industry, the result demonstrated that the framework for managing Customer Requirements in Graphic Communications should help to support operations work process, and ensuring a more consistent level of quality. Development processes and its final result would have greater opportunity to be successful, while the use of project resources with high probability could be improved and more efficient through the structure framework brings.

Abstract Requirements engineering (RE) is an important process in systems development. This research was carried out in the context of Very Large-Scale Requirements Engineering (VLSRE) within the scope of a requirement screening (RS) process. The RS process is defined as a front-end process for screening incoming requests, which are received in a constant flow. The goal of the RS process is to efficiently identify the most promising requests for further analysis, development and implementation while filtering out non-valuable ones as early as possible. The objective of this study was to understand the challenges related to the RS process and develop solutions to address those challenges. A qualitative research approach was utilised to achieve the research goals. The overall research process follows an action research method, in which each action research cycle includes at least one individually defined and executed case study. Action research and case studies are research methods that are well suited to studying real-life phenomena in their natural settings. This research was carried out in two case companies in the information and communication technology domain. Data from 45 interviews were analysed for preparing publications I–V, which are included in this thesis. In addition, during the longitudinal action research study described in this thesis, data from 26 interviews and 132 workshops were utilised to develop solutions for the RS process, which is an industrial implementation of the VLSRE process. The conducted action research contributes to the field of software engineering, in which such research efforts are currently lacking. This research has identified a number of significant challenges that different stakeholders face related to requirements processing and decision making in the VLSRE context. Examples of these challenges are the great number of incoming requirements, the lack of information for decision making and the feasibility of utilised tools. To address the identified challenges, a requirements architecting method was developed. The method includes a dynamic requirement template, which gathers structured information content for eliciting requests, documenting and communicating requirements and forming features while considering the needs of different stakeholders. The method was piloted, validated and deployed in industry. / Tiivistelmä Tutkimus toteutettiin laajamittaisen vaatimusmäärittelyprosessin kontekstissa keskittyen vaatimusten seulontaprosessiin. Vaatimusten seulontaprosessi määritellään tuotekehityksen alkuvaiheen prosessiksi, jossa käsitellään jatkuvana vuona tulevia kehityspyyntöjä. Vaatimusten seulontaprosessissa pyritään tunnistamaan tehokkaasti lupaavimmat pyynnöt jatkoanalyysiä, tuotekehitystä ja toteutusta ajatellen sekä suodattamaan pois niin aikaisessa vaiheessa, kun mahdollista ne pyynnöt, joilla ei ole arvontuotto-odotuksia. Tutkimuksen tavoite oli ymmärtää haasteita, jotka liittyvät vaatimusten seulontaprosessiin sekä kehittää ratkaisuja näihin haasteisiin. Tutkimuksessa käytettiin laadullisen tutkimuksen menetelmiä. Kokonaisuutena tutkimusprosessi noudattaa toimintatutkimuksen periaatteita siten, että jokainen sykli tai sen vaihe sisältää yhden tai useamman itsenäisesti määritellyn tapaustutkimuksen suunnittelun ja läpiviennin. Valitut tutkimusmenetelmät soveltuvat hyvin tilanteisiin, joissa tutkimuskohteina ovat reaalimaailman ilmiöt niiden luonnollisissa ympäristöissä havainnoituina. Tutkimusaineisto kerättiin kahdesta informaatio- ja kommunikaatioteknologia-alan kohdeorganisaatiosta. Väitöskirjaan sisällytettyihin julkaisuihin I-V on analysoitu 45 haastattelun aineisto. Näiden lisäksi väitöskirjassa kuvatun pitkäkestoisen toimintatutkimuksen aikana hyödynnettiin 26 haastattelun ja 132 työpajan aineistoa kehitettäessä ratkaisuja vaatimusten seulontaprosessin haasteisiin. Vaatimusten seulontaprosessi on laajamittaisen vaatimusmäärittelyprosessin teollinen toteutus. Tutkimuksessa tunnistettiin useita merkittäviä haasteta, joita eri sidosryhmillä on liittyen vaatimusten seulontaprosessiin ja päätöksentekoon laajamittaisessa vaatimusmäärittelyprosessissa. Vaatimusten suuri määrä, päätöksentekoon tarvittavan tiedon puute ja käytössä olevien työkalujen soveltumattomuus ovat esimerkkejä tunnistetuista haasteista. Ratkaisuna haasteisiin kehitettiin vaatimusten seulonta- ja analyysimenetelmä. Kehitetty menetelmä sisältää dynaamisen vaatimusdokumentin, jonka avulla voidaan kerätä kehityspyyntöjen tietosisältö jäsennellysti, dokumentoida ja kommunikoida vaatimukset sekä muodostaa niistä tuotteisiin toteutettavia ominaisuuksia ottaen huomioon eri sidosryhmien tarpeet. Kehitetty menetelmä on koestettu, validoitu ja soveltuvin osin otettu käyttöön teollisuudessa.