Searching for Value : Using Apache Solr to Find Relationships Between Work Items in a Product Backlog
Kiiski, Teemu (2023)
2023
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2023121537833
https://urn.fi/URN:NBN:fi:amk-2023121537833
Tiivistelmä
Software bugs are an inevitable part of software development. The question often arises as to who should pay for the bugfix. In a dual business model where a software product is developed under internal product development and customer projects, assigning a bugfix to the correct category is crucial. If a bug cannot be assigned to the customer project that caused it, it will become technical debt that competes for the capacity reserved for internal product development.
This Master’s thesis project was commissioned by Innofactor, a Finnish software company that operates in the Nordic countries and provides Microsoft and in-house solutions for promoting a digital organization. The project tackled the problem of identifying bugs that have been caused by customer-requested code changes. The proposed solution was a helper tool that looks for similarities between a new bug report and other work items in a product backlog.
The problem of looking for similarities between work items based on text data is an information retrieval problem. In order to quantify the problem, quantitative data collection and analysis was carried out. Likewise, the processing of software bugs is related to software quality assurance. The domains of information retrieval and software quality assurance formed the theoretical framework for the project. Action research was chosen as the research method.
The goal of the project was to create a proof of concept for a helper tool, which could be used to find the work item related to the code change that caused the bug. A proof of concept was built on the Apache Solr search platform. The backlog data was exported from a task management tool and indexed in Solr. The MoreLikeThis feature was used to test the solution, which produced promising results but also revealed further development needs.
This Master’s thesis project was commissioned by Innofactor, a Finnish software company that operates in the Nordic countries and provides Microsoft and in-house solutions for promoting a digital organization. The project tackled the problem of identifying bugs that have been caused by customer-requested code changes. The proposed solution was a helper tool that looks for similarities between a new bug report and other work items in a product backlog.
The problem of looking for similarities between work items based on text data is an information retrieval problem. In order to quantify the problem, quantitative data collection and analysis was carried out. Likewise, the processing of software bugs is related to software quality assurance. The domains of information retrieval and software quality assurance formed the theoretical framework for the project. Action research was chosen as the research method.
The goal of the project was to create a proof of concept for a helper tool, which could be used to find the work item related to the code change that caused the bug. A proof of concept was built on the Apache Solr search platform. The backlog data was exported from a task management tool and indexed in Solr. The MoreLikeThis feature was used to test the solution, which produced promising results but also revealed further development needs.