Call for Papers - EMSE Special Issue on “Empirical Studies on Prompts and other Instruction Artifacts in Software Engineering”

A special issue of the Empirical Software Engineering Journal.

Editors of the Special Issue

• Lead Guest Editor:
  • Vincenzo De Martino (Universitat Politècnica de Catalunya, Spain), vincenzo.de.martino@upc.edu
• Guest Editors:
  • Giovanna Broccia (ISTI-CNR, Italy), giovanna.broccia@isti.cnr.it
  • Fabiano Pecorelli (Pegaso University, Italy), fabiano.pecorelli@unipegaso.it
  • Jennifer Horkoff (Chalmers University of Technology and the University of Gothenburg, Sweden), jenho@chalmers.se

Description of the Special Issue

Language Models (LMs), both large (LLMs) and small (SLMs), have significantly advanced the state of the art in Software Engineering (SE), both in enhancing standard SE activities and in driving SE research for LM-enabled systems. In the LM for SE direction, LMs have enabled new forms of automation and assistance across a wide range of tasks, such as code generation, bug fixing, and documentation generation. Their ability to understand and produce natural and programming languages has transformed how developers interact with software artifacts and development tools. In the SE for LM direction, software systems now often contain one or more LM agents, assigned specific tasks and coordinating with other agents, tools, or standard software through APIs. However, the effectiveness of LMs in SE largely depends on how they are instructed to perform a given task. Instructions to LMs can include prompts, system prompts, message templates, agent communication schemas, and behavior/tool guardrail policies. These objects now become first-class objects of study. This includes the practice of prompt engineering, the practice of designing, structuring, and refining prompts to guide model behavior. Well-crafted prompts can significantly improve output quality, consistency, and reliability, while poor prompting can lead to irrelevant, incorrect, or inefficient responses. Empirical studies should move beyond prompts to systematically investigate the structure and nature of any instruction artifact associated with LMs for software, or with LMs in software. In this context, emerging perspectives such as Instruction Artifacts for LM-based SE and Task Specification for Language-Model Systems highlight the need to study not only natural-language instructions but also their interaction with structured schemas, orchestration logic, and agent-based communication patterns. With the rise of agentic systems powered by LMs, understanding how instruction artifacts are designed, combined, and evolved - both for and between agents - becomes a critical research challenge. This special issue aims to provide a dedicated venue for empirical research where instruction artifacts are treated as first-class objects of study. In particular, we target studies that empirically evaluate how changes to these artifacts affect software engineering outcomes (e.g., correctness, reliability, robustness, cost, sustainability, and developer experience), and that report instruction artifacts and execution settings in sufficient detail to support replication.

Topics of Interest

The special issue welcomes contributions on (but not limited to) the following topics:

• Empirical evaluations of prompting strategies and patterns for Software Engineering tasks.
• Development and validation of metrics for prompt quality, effectiveness, cost, energy consumption, and reproducibility.
• Studies on prompt testing, validation, and quality assurance.
• Integration of prompt engineering practices into traditional SE workflows and pipelines.
• Assessment of tools and environments supporting prompt creation, optimization, and management.
• User studies on the experience of software engineers designing and using prompts in LM-based workflows.
• Industrial case studies of prompt engineering applications in Software Engineering.
• Investigations into the cognitive and human factors influencing prompt design and understanding.
• Challenges and lessons learned from applying prompt engineering in real-world SE settings.
• Replication studies and empirical findings, including negative or non-significant results.
• Green and Sustainable prompt engineering practices

Schedule

• Submission deadline: March 1, 2027;
• Date first review round completed: April 30, 2027;
• Date revised manuscripts due: July 31, 2027;
• Date completion of the review and revision process (final notification): November 30, 2027.

Submission Instructions

Papers should be submitted through the Empirical Software Engineering editorial manager website (http://www.editorialmanager.com/emse/) as follows (1) select “Research Papers” and (2) later on the Additional Information page:

• Answer “Yes” to “Does this paper belong to a special issue?”
• And select “PROMPT-SE 2026” for “Please select the issue your manuscript belongs to”.
• In the cover letter, please explicitly explain why the submission fits with the special issue and, in case of extensions of conference papers, what additional research is included with respect to the conference version.

For formatting guidelines as well as submission instructions, visit http://www.springer.com/computer/swe/journal/10664?detailsPage=pltci_2530593