What is it about?
When programmers write code, their tools often try to guess what they want to type next and suggest possible options — just like how your phone suggests words while you’re texting. However, as software projects get larger and more complex, these suggestions can become overwhelming and less helpful, sometimes making it harder for developers to find the right option. Our work introduces a new approach called CoCoCoLa, which gives programmers more control over these suggestions. Instead of only relying on the names of code elements, CoCoCoLa lets users filter suggestions based on specific characteristics, like the type of result they want or the number of inputs a function takes. This makes coding faster, more accurate, and less frustrating, especially in big projects.
Featured Image
Why is it important?
Unlike traditional code completion tools that mostly rely on the names of code elements and try to guess what the developer wants, CoCoCoLa allows users to directly specify what kind of code suggestions they are interested in, using a simple and flexible mini-language. This is particularly timely as codebases and developer tools have become much more complex, often leading to a “paradox of choice” where too many options slow down productivity. Our method stands out by providing granular control without disrupting the existing workflow, and it’s designed to be compatible with more than 50 popular programming languages. By returning control to the developer, CoCoCoLa has the potential to make code completion smarter, more relevant, and less overwhelming, which could significantly enhance productivity for software teams worldwide.
Perspectives
When writing code, I personally get bored and annoyed quickly with long lists of irrelevant code suggestions in modern IDEs and tend to ignore them entirely to focus on the code I want to write. This works well if I know the language and its libraries well, but otherwise that's really a missed opportunity where the IDE could have helped me but doesn't. With CoCoCoLa, I wanted to build a bridge between intelligent code suggestions and developer intuition, making coding not only faster but also more enjoyable. I believe this approach will open new directions for making development tools more user-centric in the future.
Vadim Zaytsev
Universiteit Twente
Read the Original
This page is a summary of: CoCoCoLa: Code Completion Control Language, June 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3742876.3742883.
You can read the full text:
Contributors
The following have contributed to this page
