Softwareudvikling i et foranderligt landskab - Et tværfagligt softwareudviklingsprojekt til understøttelse af organisationen Ledige Med Drive

摘要

This thesis is about exploring the challenges associated with software developmentwithin organizations. To discover and investigate these challenges, a practicaldevelopment project was conducted within the organization Ledige Med Drive inRoskilde Denmark.Through utilization of the four principles of the Participatory Design Method, MUST,and Robert K. Wysocki’s unified theory about Software Development and ProjectManagement, the project were initialized by scoping and visioning activities. Theseactivities lead to a rationalized choice of software development strategy in accordanceto Wysocki’s software development landscape model.Utilizing the agile development method Scrum, a seven-week development processwere undertaken within the organization. This process was focused around developinga webpage solution to support the activities within the organization while fulfilling theneeds and wishes of its users. Throughout the development process, the Scrumactivities resulted in a high degree of re-evaluation and adjustment of the processesevaluated. The utilization of Scrum principles and continuously examination andadaption of our development process lead to a better understanding of the challengesassociated with the software development. By assessing different possible webtechnologies that could be used for implementing the identified system requirement, anumber of applied technologies were decided upon. Through a stepwise implementationof the software solution, we identified a rising need for a clearer separation betweenthe webpages frontend and backend functionality as the complexity of the developedsolution increases.As an exploration of potentials for improving the matching between unemployed andemployers in relation to job application and recruitment, there was conducted anexperiment using a selection of Machine Learning techniques. The assessed techniquesinvolves classification and clustering algorithms applied on a bag-of-wordrepresentation of text-documents. Throughout this experiment, we discover greatpotential for utilizing the assessed techniques for recommendation purposes in relationto job application and recruitment.