GSoC’20 RoboComp project: Syntax-error highlighter for LearnBlock
Hello! I’m José Miguel, a Computer Engineering student at the University of Extremadura in Cáceres, Spain. I’ve grown up with a passion for technology. Over time I’ve played around with computer graphics, programming language theory, operating system design, game modding, web design and programming and a bit of computer security… I love learning something from every field I discover!
About the Project
LearnBlock is a visual robot programming tool that eases the path towards learning full-fledged robot programming (in Python). Putting some more straightforward layers in between makes it easier for newcomers to implement their ideas and get results quickly.
I was amazed when I discovered LearnBlock. not only you have a productive visual block environment to code (and a simplified text language in the middle), but the resulting code is portable among a variety of robots!
My goal with this project is to fix one critical flaw it has right now: lack of proper error reporting. In learning tools, it is crucial to have useful guidance through error resolution, since the target user usually lacks the intuition or experience that an expert developer has. The environment should not only point to the place where it found a problem, but also provide some tips which make the solution clear and explain why it is required. Then, the user has not only solved their problem but has also learned why the previous solution was incorrect and how to fix it in the future.
I plan to implement a core error handling mechanism, not tied to any representation but to an internal tree-like mirror of the code, which allows for centralized error checking, handling and reporting. A typechecker which annotates its nodes and finds type mismatches to help further the user locate errors, as well as, an appropiate frontend signaling of these errors for each of the available representations (currently, the visual block-based one and the simplified text one).
José Miguel Sánchez García