THIS REPORT HAS BEEN WITHDRAWN

Lisa Rennels and Sarah Chasins

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2022-82
May 12, 2022

Software and data analytics are increasingly central to the work of domain experts across myriad fields, many of which lie entirely outside of traditional computer science. This trend makes domain specific languages (DSL) a valuable and popular tool for software engineers to support to research and work in particular domains, including the earth sciences. In particular, embedded DSLs are an approach to efficiently developing powerful and rich languages and tools for niche domains. Improving understanding of how embedded DSL programmers use the language including usage patterns, pain points, and the pros and cons of various intrinsic and extrinsic language features, is key to informing future design that prioritizes usability. We conducted a contextual inquiry including observation and short semi-structured interviews with the users of an embedded DSL, Mimi, written as a platform for climate economics domain experts in the integrated assessment modeling field. We then used thematic analysis to build up a hierarchy of primary themes from the audiovisual data. These including five primary findings, each with a specific implication for future embedded DSL design. These findings include that the host language of an embedded DSL is very consequential for the user, not just the designer, and that personal engagement with community is crucial to attracting and retaining users.

Advisor: Koushik Sen and Sarah Chasins

Author Comments: withdrawn to upload clean formatted copy next