Metraipolis › archive › dc_970b2ede44b8
Beyond the Prompt: Linking What Developers Ask, Do, and Understand with Coding Agents — arXiv:2609.33712
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.33712
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T23:02:12.369Z
- SHA-256 of the archived text
ea0c3305f2b4b80c694c29a6887907e3b8fc31213d36a3ca1fba03bcdb6715a5- Stable URL of this page
https://metraipolis.com/documento/dc_970b2ede44b8
How to cite
Beyond the Prompt: Linking What Developers Ask, Do, and Understand with Coding Agents — arXiv:2609.33712 — arXiv (en), https://arxiv.org/abs/2609.33712 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_970b2ede44b8 (Metraipolis archive dc_970b2ede44b8, sha256 ea0c3305f2b4b80c694c29a6887907e3b8fc31213d36a3ca1fba03bcdb6715a5)
BibTeX
@misc{dc970b2ede44b8,
title = {Beyond the Prompt: Linking What Developers Ask, Do, and Understand with Coding Agents — arXiv:2609.33712},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.33712},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_970b2ede44b8; sha256 ea0c3305f2b4b80c694c29a6887907e3b8fc31213d36a3ca1fba03bcdb6715a5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33712. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Coding agents can now change code for developers, who describe goals, supply context, and respond to the agent's work. Yet prompts, screen activity, and task success each tell only part of this story. We present Say, Do, Understand, an end-to-end workflow for analyzing what developers write to an agent, what they do while it works, and what they can explain afterwards. The workflow has five stages (Capture, Prepare, Analyze, Integrate, and Interpret) and three instruments: a prompt codebook, a scheme for coding screen-recorded activities and events, and separate rubrics for explaining the process and the solution. We applied it in an observational study of ten experienced developers who used GitHub Copilot on an unfamiliar codebase. Crossing task performance with understanding produced four personas. The two measures agreed for eight developers but split for two: one passed most tests but could not explain the solution, and another passed few tests but explained it well. In this sample, the personas that most often asked the agent to check its work spent the least time testing on their own. These patterns are descriptive and do not generalize beyond the sample. We recommend the workflow to computing educators, industry practitioners, and researchers to adapt and evaluate human--AI communication in software engineering. Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-29 in cs.SE). Categorie arXiv: Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2609.33712 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T22:39:35.202Z dalla pagina https://arxiv.org/catchup/cs.SE/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).