A new tool connects novice programmers with projects that match their experience and interests.
What’s new: Github’s Good First Issues tool uses deep learning to find easy-to-fix issues among the collaborative software development platform’s multitude of open source projects.
How it works: Github, which is owned by Microsoft, enables developers to collaborate freely worldwide. Participants often flag bugs to be fixed or features to be implemented, but beginners may have trouble figuring out which are appropriate to their skill level.
- Github trained a deep learning model on a dataset of issues labeled with designations like “beginner friendly,” “low-hanging fruit,” and “easy bug fix.” The metadata also noted whether issues were closed by someone who had not previously contributed to the repository and how many lines of code were involved.
- The model assigns a probability score to new issues based on how likely they are to be easily fixed. Users can browse by problem type or project. If they’ve been sufficiently active on Github, they can receive a list of open issues suited to their previous contributions.
Behind the news: An earlier version used traditional computing to query a list of 300 beginner-friendly labels. However, it surfaced only about 40 percent of relevant issues, the company said.
Why it matters: Github is a focal point of software development and the heart of the open source movement. Helping people figure out where they can have the most impact can only make it more productive.
We’re thinking: What a great tool for aspiring developers! When jumping into AI (or, indeed, most disciplines), it’s better to start small and succeed than to start too big and fail.