[odipar]

Thanks for the reply!

Unfortunately, persistent data structures tend to produce (short-lived) garbage which the JVM is very good at collecting!

So yes, Clojure benefits immensely from the JVM.

It is also an interesting research topic whether (optimised) reference counting would be a better approach.

Regarding objects, there is also a "middle ground" to consider:

Split big (immutable) arrays in smaller ones, connect them with some pointers in between, and you are still cache friendly.

Also, you can do a lot on the application level to reduce garbage, and most Java programmers don't care for that exactly because of JVM.

[c-cube]

About the refcounting approach, you may want to look at the Perceus paper. It's refcounting with dynamic reuse of memory that isn't shared (like a sort of runtime linear typing), and it's used in Koka for functional programming.