[concurrency-interest] on happens-before formalism

Hanson Char hanson.char at gmail.com
Sun Feb 19 12:19:56 EST 2012


Alternatively, it seems "happens-before" really means
"happens-on-or-before", which probably is easier to say or write, and
fits more closely with it's actual definition as a partial order (and
not a strict partial order).

Hanson

On Fri, Feb 3, 2012 at 8:04 AM, Joe Bowbeer <joe.bowbeer at gmail.com> wrote:
> When this question arose on The Art of Multiprocessor Programming discussion
> list a couple years ago, the resolution was that happens-before (<=) is a
> strict partial order, therefore reflexive.
>
> In other words, happens-before can be interpreted as
> "happens-before-or-is-equal" -- but no one wants to say or write that.
>
> --Joe
>
>
> On Thu, Feb 2, 2012 at 11:28 PM, David Holmes <davidcholmes at aapt.net.au>
> wrote:
>>
>> It seems to me to be somewhat arbitrary to define this as either reflexive
>> or irreflexive as it makes no practical difference to the semantics. I
>> think
>> irreflexive would be more appropriate in this case as intuitively it
>> doesn't
>> make sense to say that "A happens-before A". I suspect that by selecting
>> reflexive and using a normal/simple notion of poset that the overall
>> formalism is simplified.
>>
>> If you really want to know ask on the Java Memory Model list cc'ed.
>>
>> Cheers,
>> David Holmes
>>
>> > -----Original Message-----
>> > From: concurrency-interest-bounces at cs.oswego.edu
>> > [mailto:concurrency-interest-bounces at cs.oswego.edu]On Behalf Of Pavel
>> > Rappo
>> > Sent: Friday, 3 February 2012 5:18 PM
>> > To: Concurrency-interest at cs.oswego.edu
>> > Subject: Re: [concurrency-interest] on happens-before formalism
>> >
>> >
>> > Then I think it's very different from Lamport's definition of
>> > happened-before which they reference to.
>> > Maybe that's why Lamport calls it "happenED-before" (not
>> > "happenS-before").
>> >
>> > Btw, there's such thing as strict partial order, which I guess in
>> > this case is more suitable. It's irreflexive, asymmetric and transitive.
>> >
>> > On 3 Feb 2012, at 09:16, Qingzhou Luo wrote:
>> >
>> > > I think they did mean reflexive.
>> > >
>> > > See partial order definition:
>> > > http://en.wikipedia.org/wiki/Partially_ordered_set
>> > >
>> > > On Thu, Feb 2, 2012 at 5:51 PM, Pavel Rappo
>> > <pavel.rappo at gmail.com> wrote:
>> > > I've been reading "SPECIAL POPL ISSUE The Java Memory Model" by Jeremy
>> > > Manson, William Pugh and Sarita Adve.
>> > > There's one (yet) thing seems strange to me. Though I think
>> > it's actually a
>> > > typo it's still worth mentioning.
>> > >
>> > > page. 8, 2.1: "...Note that all of this means that happens-before is a
>> > > partial order: it is reflexive, transitive and anti-symmetric..."
>> > >
>> > > Am I right saying authors actually meant "irreflexive"?
>> > >
>> > > --
>> > > Sincerely yours, Pavel Rappo.
>> > > _______________________________________________
>> > > Concurrency-interest mailing list
>> > > Concurrency-interest at cs.oswego.edu
>> > > http://cs.oswego.edu/mailman/listinfo/concurrency-interest
>> > >
>> > >
>> > >
>> > > --
>> > > Qingzhou Luo
>> > > http://mir.cs.illinois.edu/~qluo2/
>> > >
>> > > Department of Computer Science
>> > > University of Illinois, Urbana Champaign
>> > >
>> >
>> >
>> > _______________________________________________
>> > Concurrency-interest mailing list
>> > Concurrency-interest at cs.oswego.edu
>> > http://cs.oswego.edu/mailman/listinfo/concurrency-interest
>> >
>>
>> _______________________________________________
>> Concurrency-interest mailing list
>> Concurrency-interest at cs.oswego.edu
>> http://cs.oswego.edu/mailman/listinfo/concurrency-interest
>
>
>
> _______________________________________________
> Concurrency-interest mailing list
> Concurrency-interest at cs.oswego.edu
> http://cs.oswego.edu/mailman/listinfo/concurrency-interest
>


More information about the Concurrency-interest mailing list