Rich Hickey:

>No, I don't. The post is for Clojure users and shows them a relative 
>comparison they care about, persistent vs transient.<
>Unless that code is returning a persistent vector and provides thread 
>isolation, it is not doing the same job.<

Sorry, I didn't mean to offend you in any way. Sometimes I am not gentle 
enough. I am sorry.

You are right that my comparison with D was wrong because the semantics of that 
Clojure code and data structures is richer.

On the other hand a comparison with a simpler/basic implementation (like just a 
C array) can be quite useful anyway, to know how much you have to pay for such 
useful extra semantics.

I have seen several online benchmarks of many languages, Clojure too. Often 
they show only numbers for a single language. In such situations a comparison 
with other languages is useful to put numbers in perspective. So a comparison 
with (for example) a baseline C program is useful (hopefully with the same 
semantics, where possible). For Clojure a comparison with Java can be equally 
useful.

Bye,
bearophile

Reply via email to