Context. Maintaining good software documentation quality is an important aspect of software maintenance. To decrease associated costs, interest in automating documentation tasks has recently increased. However, there is a lack of guidance regarding how practitioners may apply related research findings to pre-existing documentation.

Objectives. In this case study we investigate what the most important documentation properties for manually produced component documentation quality are and how these can be linked to issues in real-world documentation. We also examine to what extent they can be tackled based on current tools and approaches.

Methods. A literature study is employed to identify previously reported documentation properties and related automation strategies. These are then mapped against issues identified in the documentation of a telecommunications company department. Finally, strategies for maintaining documentation quality are proposed and the applicability of these is retrospectively evaluated.

Results. Current deeply automated strategies are found to be largely ineffective for the analyzed documentation issues specifically and for improving documentation content not tightly linked to source code in general. Manually produced automated tests and version control as well as raw documentation formatting enabling these are identified as more promising approaches for improving documentation maintenance.

Propositions for diminishing department documentation issues. Four suggestions are made on how documentation issues could be tackled. Particular attention is given to making code samples testable.

Conclusions and Future Work. Manual documentation is still often needed, and writing it in a way that enables automation where possible will become increasingly important. Containerization technology is widely used but there is a dearth of corresponding testing knowledge and tooling. Problems with code samples in documentation commonly cause critical issues. Improving methods for maintaining code-related sections of documentation, especially containerization-related code samples, is an urgent challenge.