You seem to have been frustrated with prescriptive programming advice and language evangelism, which is understandable. Yet it looks like you're conflating Hickey's design philosophy with language zealotry. The talk's core message about reducing accidental complexity remains valuable, even if the presentation style and some specifics, sure, can be characterized as debatable.
I can buy your vocabulary criticism - fair point, okay. Although vocabulary often can help crystallize concepts. Clojure's internal mutation - alright, accurate observation. Clojure uses mutation internally for performance while presenting an immutable interface. This does show pragmatism over purity. "Tooling doesn't fix bad coders" - true. No language or paradigm prevents bad code if developers don't understand the principles. Immutability isn't always best - correct. Even functional languages acknowledge this (IO monads, ST monad, etc.), but Clojure doesn't force immutability - it provides default immutability with escape hatches.
Now things that I can't agree with:
"Academic blabla" is outright dismissive. The talk addresses real architectural problems many developers face. The criticism of ORMs isn't just about "complexity" - it's about impedance mismatch, hidden performance costs, and leaky abstractions. These are well-documented issues. "Must be simple" is a misrepresentation of Hickey's point. He advocates for simplicity as a goal, not an absolute rule. Builder pattern equivalence is oversimplification. Persistent data structures offer different guarantees than builders (structural sharing, thread safety, etc.).
I can buy your vocabulary criticism - fair point, okay. Although vocabulary often can help crystallize concepts. Clojure's internal mutation - alright, accurate observation. Clojure uses mutation internally for performance while presenting an immutable interface. This does show pragmatism over purity. "Tooling doesn't fix bad coders" - true. No language or paradigm prevents bad code if developers don't understand the principles. Immutability isn't always best - correct. Even functional languages acknowledge this (IO monads, ST monad, etc.), but Clojure doesn't force immutability - it provides default immutability with escape hatches.
Now things that I can't agree with:
"Academic blabla" is outright dismissive. The talk addresses real architectural problems many developers face. The criticism of ORMs isn't just about "complexity" - it's about impedance mismatch, hidden performance costs, and leaky abstractions. These are well-documented issues. "Must be simple" is a misrepresentation of Hickey's point. He advocates for simplicity as a goal, not an absolute rule. Builder pattern equivalence is oversimplification. Persistent data structures offer different guarantees than builders (structural sharing, thread safety, etc.).