On Aug 11, 2009, at 5:14 PM, fft1976 wrote:

> On Aug 11, 1:50 pm, Stuart Sierra <the.stuart.sie...@gmail.com> wrote:
>
>> I agree wholeheartedly.  Let's optimize wherever possible, but drop  
>> on
>> the side-by-side comparisons.
>
> I think it's very immature to dictate to others what should be
> important to them. Speed may be of minor concern to some Ruby-on-Rails
> types, but it is important to me, and if the documentation of Clojure,
> its author himself and many of the fans say that it should be exactly
> as fast as Java, when doing the same thing, I think it's valid to
> wonder why it doesn't happen in reality.

Let's try to take a little of the hyperbole out of things.

No one has tried to dictate anything to anyone.  Performance is  
important to you?  It's important to me, too, and I'm hungry to get  
more speed out of clojure than I do already (being locked in to v1.0  
through the end of the week).  But I know that performance is highly  
application-dependent; for one other data point, our partial  
reimplementation of a very well-tested and highly-optimized 5-year-old  
Java codebase that is completely CPU-bound has resulted in roughly  
equivalent performance (quite literally, less than 10% on either side  
depending on input data).  I can't wait to take advantage of newnew,  
transients, etc.  Your application domain may expose some rough edges  
that could be improved, in which case whatever you can do to help will  
surely be welcomed by *everyone*.

That said, I don't think I've ever heard Rich Hickey, or anyone else  
for that matter, say that clojure would ever be "exactly as fast as  
Java" in every domain.  Such a claim would be somewhat absurd to make,  
especially when other jvm languages that have far more in common with  
Java and that have been around for far longer (here I'm thinking of  
Scala) haven't yet risen to that level across the board.

I know I've heard Rich say things like "clojure can be as fast as  
Java", or in the case of certain numeric benchmarks (which is always  
an exercise in cherry-picking), it can be exactly as fast.  But in the  
end, performance is a land full of shades of gray, and benchmarks  
generally don't do that landscape's complexity the justice it deserves.

If your domain is one where clojure doesn't meet your standards, I  
hope you'll recognize the language's potential, and do whatever you  
can to help -- but please don't assume that everyone else in the  
community shares your specific priorities.  Those who know the deep  
innards of the clojure compiler and data structures are hardly  
standing still, but it's a big sandbox these days, and there's a lot  
of different concerns.

Cheers,

- Chas

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your 
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to