|
|
|
|
|
|
by xelxebar
2433 days ago
|
|
|
Just starting to learn J, myself, and this brings to mind Aaron Hsu's "Does APL Need A Type System?" talk: https://www.youtube.com/watch?v=z8MVKianh54 which argues that APL already provides the development ergonomics that are typically acribed to type systems. I.e. why type check your tacit function when it's implementation is about the same length as its type annotation? Anyway, as someone trained in Category Theory and a huge fan of type systems, I found Hsu's argument surprisingly pursuasive. I would love to hear any counterpoint arguments and the motivations behind attempts at giving J a type system. |
|
|