Re-architecting a Legacy Codebase with FOSS Tools

A useful climate data record is inherently persistent. The longer the record is, the more useful it is for analyzing trends in the subtle shifts in the Earth's radiation budget. In kind, ensuring the technical relevance of data formats and coding standards will contribute to the longevity of the dataset. Whether the code or datasets are always useful to scientists and the general public is dependent on their modern-day, technical relevance.

Clouds and the Earth's Radiant Energy System (CERES) maintains a long-term climate data record. The hope is that data collection, of some form, will continue for many years to come. Approximately twenty years into raw data collection, the CERES science team is reassessing the usability of its code and data products. Free and Open Source (FOSS) tools have greatly contributed to the re-architecting of the codebase and to the analyzing the usability of data products.

Some tools have aided primarily in code re-architecture. For example, Copy/Paste Detector (CPD) software has provided a simple way to highlight regions of copied code or logic in CERES' enormous codebase, to collapse the lines-of-code count. Other tools have contributed to the improvement of modularity, readability, and efficiency of the codebase. Employing Git for version control has provided developers with a new way to collaborate remotely and mitigate tricky code merges. To reinforce proper programming practices, the science team has hosted Software Carpentry courses, which are two-day software development courses for scientists, taught using a Creative Commons curriculum. As long as CERES datasets are generated, the underlying code must persist in a sustainable fashion.

A user-centric philosophy would have these datasets consistently viewable, without exception, within the most popular software tools. These tools include QGIS. Going forward, the CERES data management team is working in conjunction with the Atmospheric Science Data Center (ASDC) to ensure usability of the data, including within popular open-source options.

From data production to distribution, CERES is modernizing. Certain FOSS tools are greatly facilitating this process. They will be highlighted in this presentation focusing on sustainable practices of programming and data production.