APL Interpreter – An implementation of APL, written in Haskell (2024)

a year ago

APL is an array programming language with a compact and expressive syntax using single Unicode symbols.
The project involved creating an APL interpreter in Haskell, which was more challenging than anticipated due to Haskell's learning curve.
The interpreter follows standard interpreter steps: lexing, parsing, evaluating, and printing results.
Parsing evolved through versions, starting from context-free to incorporating context and monads for better handling of APL's grammar.
Evaluation in APL involves handling functions as data, with higher-order functions forming trees naturally.
The interpreter uses monads for state handling, with a focus on practical typeclasses to manage different evaluation contexts.
Selective assignment in APL allows expressions on the left-hand side of assignments, provided they use specific functions.
Working with higher-dimensional arrays in APL is primarily about index arithmetic and shape manipulation.
The interpreter mimics Dyalog APL but has discrepancies in behavior, printing formats, and supported features.
Haskell's strengths include compiler guarantees, useful libraries, and functional programming features, but it has a steep learning curve and challenges with debugging and efficiency.

Hasty Briefsbeta