Finlands Akademi
Sökande / Kontakt person Mäntylä, Mika
Organisation Uleåborgs universitet
Projektets titel Detecting Technical Debt with Natural Language Processing
Beslutnr 328058
Beslutsdatum 04.09.2019
Finansierings period 01.01.2020 - 31.12.2022
Finansiering 415 356
WebFOCUS Report
Beskrivning av projektet
We think that by analyzing millions of source code files, we can capture the collective mind of software developers. Software developers’ thoughts can be analyzed from the natural language comments that provide opinions about source code. In this research proposal, we focus on thoughts about technical debt in the source code. Technical debt utilizes the concept of financial debt to illustrate the technical problems that result in unnecessary increases in software development costs. Gartner has evaluated that the cost of technical debt can be several hundred millions of dollars to the software industry. Currently, thousands of open source projects enable large-scale probabilistic reasoning on human thoughts about source code. We build a natural language processing (NLP) based technical debt detector. Our work is a part of a greater path that is on making computers understand source code as humans understand it.