Lush: My favorite small programming language

Scott Locklin reflects on Lush, a small programming language he once favored, which has historical significance in the development of machine learning algorithms. Initially popular among its creators, including Yann LeCun and Leon Bottou, Lush has faded from mainstream use as other languages like R and Python gained traction. Locklin recently revisited Lush, discovering updates and fixes made by Bottou on its older version. He appreciates Lush for its ergonomic design, self-documenting features, and ease of integrating C libraries, which contrasts sharply with the complexities found in other languages like R. Lush's array handling and object system are highlighted as particularly useful for machine learning tasks. Despite its age, dating back to 1987, Locklin argues that Lush remains a cozy and effective tool for developing algorithms, emphasizing its programmer-centric design. He notes some limitations, such as outdated documentation and minor usability issues, but overall, he advocates for a revival of Lush, suggesting that it could have significantly influenced the machine learning landscape had it not been overshadowed by more popular languages.

- Lush is a small programming language with historical importance in machine learning.

- It features self-documenting capabilities and easy integration with C libraries.

- The language is noted for its ergonomic design and user-friendly object system.

- Locklin suggests that Lush could be revived to better serve modern programming needs.

- The language's decline is attributed to the rise of more popular languages like R and Python.