I'm new to compiler development and want to create a language-checking tool for my college portfolio. The idea is to give it source code—possibly a small subset of Rust—and have it report errors. I'm not sure where to begin or how the project should be structured. Should I think of this as a compiler, a linter, or something else? I'd appreciate a learning path and recommendations for beginner-friendly resources.
4 Answers
Learn how grammars are written, especially context-free grammars and notation such as BNF, along with parsing and abstract syntax trees. A focused project could accept a small fragment of an existing language and determine whether it is a valid loop or expression, then explain where invalid input occurs. That gives you a manageable scope while teaching the core ideas.
“Error checking” can mean several different things. Validating formatting and syntax is one level; checking types and whether every function returns correctly is considerably harder; determining whether the program does what the author intended is a much more advanced problem. A practical starting point is a tiny language with variables, arithmetic, conditionals, and loops. Write a recursive-descent parser or interpreter for that language, then add useful diagnostics gradually.
Start much smaller than a full Rust compiler. A good first version can process a tiny language through the usual front end: tokenize the text, parse the tokens, build an abstract syntax tree, and then perform semantic checks. You could report syntax errors, undeclared variables, or simple type mismatches. That would be a substantial and impressive project even without generating executable code. Crafting Interpreters is an excellent beginner resource and is available to read online for free.
This would most accurately be called a compiler front end, or possibly a linter depending on what it checks. A compiler usually includes stages that translate source code into another representation or executable output. A checker that only analyzes syntax and meaning is not a complete compiler, but it still covers important compiler concepts and makes a strong portfolio project.
It’s also worth working through the examples instead of only reading the chapters. Implementing a small interpreter alongside the book makes the lexer, parser, AST, and semantic-analysis stages much easier to understand.