Large Language Models (LLMs) are extensively used nowadays for various software engineering tasks, primarily code generation. Previous research has shown how suitable prompt engineering can help developers improve their code-generation prompts. However, so far, there are no specific guidelines guiding developers in writing suitable prompts for code generation. In this work, we derive and evaluate development-specific prompt optimization guidelines. We use an iterative, test-driven approach to automatically refine code-generation prompts, and we analyze the results of this process to identify prompt improvement items that lead to test passes. We use such elements to elicit 10 guidelines for prompt improvement, related to better specifying I/O, pre/post conditions, providing examples, various types of details, or clarifying ambiguities. We conducted an assessment with 50 practitioners, who reported their use of the elicited prompt improvement patterns and their perceived usefulness. Our results have implications not only for practitioners and educators, but also for those creating LLM-aided software development tools.

In the Python ecosystem, the adoption of idiomatic constructs has been fostered because of their expressiveness, increasing productivity and even efficiency, despite controversial arguments concerning familiarity or understandability issues. Recent research contributions have proposed approaches-based on static code analysis and transformation-to automatically identify and enact refactoring opportunities of non-idiomatic code into idiomatic ones. Given the potential recently offered by Large Language Models (LLMs) for code-related tasks, in this paper, we present the results of a replication study in which we investigate GPT-4 effectiveness in recommending and suggesting idiomatic refactoring actions. Our results reveal that GPT-4 not only identifies idiomatic constructs effectively but frequently exceeds the benchmark in proposing refactoring actions where the existing baseline failed. A manual analysis of a random sample shows the correctness of the obtained recommendations. Our findings underscore the potential of LLMs to achieve tasks where, in the past, implementing recommenders based on complex code analyses was required.

Land Use Land Cover (LULC) classification plays a crucial role in the optimisation of agricultural lands to cope with global warming and increase crops. Many existing methodologies for LULC classification rely on machine learning algorithms and/or hyperspectral imaging, both of which necessitate extensive annotated datasets and substantial computational resources. Moreover, these methods could yield results that are in some cases inaccurate or lack interpretability. This paper introduces an innovative and effective approach to extract characterising colours from different land types. The proposed solution manages to automatically generate the sets of colours corresponding to categories of land, from a small set of annotated images, in just one iteration. Such colours were then used to classify lands in other images by analysing colour distributions and assessing pixel densities. The results demonstrated the precision and accuracy of our approach, which successfully classifies up to three categories of land across multiple images.

A large amount of test cases is very useful to check the correctness of a software system while it is developed. Often a considerable time is dedicated by human programmers to designing effective test cases. This paper proposes an approach for automatically generating test cases tailored to the characteristics of the code under test. For this, the classes of a software system to be tested are characterised by a static code analysis aiming at summarising and representing their behaviour. As test cases check the behaviour of code, classes that exhibit a close behaviour may be checked using similar test cases. Therefore, in the approach proposed, for classes having a comparable behaviour, test cases are generated by taking as a template the test cases available for one of the classes among the similar ones. The approach has been assessed on a few open source projects and has proved to be viable for generating applicable and effective test cases for the classes.