| Scientific and Engineering Software (SES) is different from commercial software, it often targets computational analysis of a problem without a prior solution. This work is aimed at systematically reviewing the literature for extracting particular challenges and solutions of SES development. We also conducted a case study, by developing decision support and optimization software for the oil industry to bring our findings of the review into practice and provide evidence of challenges/solutions of developing SES.;Following software engineering practices, such as adopting object-oriented technology, iterative development approach, MVC architectural pattern, unit and GUI testing, we could successfully develop and commercialize the optimization software system.;Our development experience confirmed observations of the literature on these challenges, especially those of requirement elicitation and testing. The requirements in SES are often unknown upfront, developers are often domain experts and software validation is complex; for the scientific/engineering core for which no certain test oracle exists. |
