I was wondering the same thing, Jörg. This thread 
<https://groups.google.com/forum/#!topic/clojure/VQj0E9TJWYY> talks about 
it as well. I posted a note there which I will reproduce here for your 
convenience:

I think it's safe to assume that since `ArrayList` uses unsynchronized 
mutability internally (a quick review of the GrepCode entry for `ArrayList` 
confirms this 
<http://grepcode.com/file/repository.grepcode.com/java/root/jdk/openjdk/8u40-b25/java/util/ArrayList.java>),
 
then we can rest assured that a `volatile` box as opposed to a totally 
unsynchronized mutable variable is unnecessary, even in the context of 
`fold`. After all, `reduce` (and by extension, `transduce`) is only ever 
going to be single-threaded unless the data structure in question 
unexpectedly implements a multithreaded reduce, which should never happen 
(and if it does, you likely have bigger problems). To be honest, I'm not 
sure why `volatile` is used in transducers instead of e.g. an 
`unsynchronized-mutable` box. There may be a good reason, but I'm not 
seeing it immediately. I'd love to learn.

On Friday, January 2, 2015 at 10:06:25 AM UTC-5, Jörg Winter wrote:
>
> So if I'd need that extra performance, say in a private library of 
> transducers not intended to be shared with other Clojure developers, it is 
> perfectly Ok to use a java.lang.Object instead of a volatile, right ?
>
> J 
>
> 2015-01-02 15:59 GMT+01:00 Timothy Baldridge <tbald...@gmail.com 
> <javascript:>>:
>
>> "As far as I understand, the step-function of a transducer is never(?) 
>> accessed concurrently by more than 1 thread." 
>>
>> It's actually "one thread at a time". And you're right, stuff like 
>> Core.async may bounce a transducer between several different threads, but 
>> only 1 thread "owns" it at a given time. 
>>
>> Timothy
>>
>> On Fri, Jan 2, 2015 at 7:11 AM, Jörg Winter <jwin...@gmail.com 
>> <javascript:>> wrote:
>>
>>> Hi,
>>>
>>> As seen in this example of a stateful transducer...
>>>
>>> http://crossclj.info/ns/org.clojure/clojure/latest/clojure.core.html#_partition-by
>>>
>>> ... I am wondering what is the concrete motivation behind using 
>>> 'volatile!' instead of say a simple (mutable) Java-Object wrapper ?
>>> In the partition-all example, an ArrayList is used for aggregating the 
>>> 'temporary' results for the step-function, so this mutable state is not 
>>> concerned with threading at all.
>>> Why then is there a threading-concern with pv (the volatile!) ?
>>>
>>> As far as I understand, the step-function of a transducer is never(?) 
>>> accessed concurrently by more than 1 thread.
>>>
>>> Is volatile! necessary because transducers should be usable with 
>>> core.async ?
>>> Or is it just an easy way to get a mutable object in Clojure ?
>>>
>>>
>>> Best,
>>> Joerg
>>>
>>> -- 
>>> You received this message because you are subscribed to the Google
>>> Groups "Clojure" group.
>>> To post to this group, send email to clo...@googlegroups.com 
>>> <javascript:>
>>> Note that posts from new members are moderated - please be patient with 
>>> your first post.
>>> To unsubscribe from this group, send email to
>>> clojure+u...@googlegroups.com <javascript:>
>>> For more options, visit this group at
>>> http://groups.google.com/group/clojure?hl=en
>>> --- 
>>> You received this message because you are subscribed to the Google 
>>> Groups "Clojure" group.
>>> To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to clojure+u...@googlegroups.com <javascript:>.
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>
>>
>>
>> -- 
>> “One of the main causes of the fall of the Roman Empire was that–lacking 
>> zero–they had no way to indicate successful termination of their C 
>> programs.”
>> (Robert Firth) 
>>
>> -- 
>> You received this message because you are subscribed to the Google
>> Groups "Clojure" group.
>> To post to this group, send email to clo...@googlegroups.com 
>> <javascript:>
>> Note that posts from new members are moderated - please be patient with 
>> your first post.
>> To unsubscribe from this group, send email to
>> clojure+u...@googlegroups.com <javascript:>
>> For more options, visit this group at
>> http://groups.google.com/group/clojure?hl=en
>> --- 
>> You received this message because you are subscribed to a topic in the 
>> Google Groups "Clojure" group.
>> To unsubscribe from this topic, visit 
>> https://groups.google.com/d/topic/clojure/CjxK7xEsOKQ/unsubscribe.
>> To unsubscribe from this group and all its topics, send an email to 
>> clojure+u...@googlegroups.com <javascript:>.
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 
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
--- 
You received this message because you are subscribed to the Google Groups 
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to clojure+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to