Abstract
Abstract. A climate model is an executable theory of the climate; the model encapsulates climatological theories in software so that they can be simulated and their implications investigated. Thus, in order to trust a climate model, one must trust that the software it is built from is built correctly. Our study explores the nature of software quality in the context of climate modelling. We performed an analysis of defect reports and defect fixes in several versions of leading global climate models by collecting defect data from bug tracking systems and version control repository comments. We found that the climate models all have very low defect densities compared to well-known, similarly sized open-source projects. We discuss the implications of our findings for the assessment of climate model software trustworthiness.
Highlights
In this paper we report on our investigation into the software quality of climate models
There is a long history of software research focused on industrial and commercial software development but it is only recently that scientific software development has been seen as an important area of research (Kelly, 2007)
"taken with other evidence, the T experiments suggest that the results of scientific calculations carried out by many software packages should be treated with the same measure of disbelief researchers have traditionally attached to the results of unconfirmed physical experiments." if Hatton’s findings are any indication of quality of scientific software in general, improvements in software quality assessment of scientific software is dearly needed
Summary
In this paper we report on our investigation into the software quality of climate models. A study by Easterbrook and Johns (2009) of the software development practices at the UK Met Office Hadley Centre estimates an extremely low defect density for the climate model produced there, which suggests an extraordinary level of software quality. Our purpose in this study is to conduct a rigorous defect density analysis across several climate models to confirm whether this high level of quality holds, and whether it is true of other models. Defect density measures the number of problems fixed by the developers of the software, normalised by the size of the body of code. We chose defect density as our indicator of quality because it is well-known and widely used across the software industry as a rough measure of quality, and because of its ease of comparison with published statistics. The measure is general and does not rely on many assumptions about how software quality should be measured, other than the notion that fewer defects indicate greater software quality
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
