[Openmcl-devel] Speed, compilers and multi-core processors

mikel evins mevins at mac.com
Tue May 19 16:39:07 PDT 2009

On May 19, 2009, at 1:39 PM, Jeremy Jones wrote:

> On Tue, May 19, 2009 at 2:12 PM, Dan Weinreb <dlw at itasoftware.com>  
> wrote:
>> ...
>> And you know what other language works this way?
>> Clojure, a very promising new dialect of Lisp.
>> -- Dan
> I'd like to endorse the approach taken by Clojure (the other Closure)
> too.  It is a dialect of Lisp that where a central design decision was
> to make almost all data structures immutable.  This obviates most of
> the need for locking, a common performance issue for multi-threaded
> programs.  It is easy to write programs that rarely use shared,
> mutable data structures.  And when you do need shared, mutable data
> structures, the locking is explicit.  Take a look at a
> http://clojure.blip.tv/file/1313398/ for an introduction to Clojure
> for Lisp programmers.  I wasn't expecting much, but it's really good.
> Someone could define a version of Common Lisp with these  
> characteristics...

It wouldn't be Common Lisp, or, if it was still Common Lisp, the  
Clojure features wouldn't buy you much. See my other message on the  
subject. In order to get Clojure's benefits for concurrency, you have  
to say goodbye to SETQ.


More information about the Openmcl-devel mailing list