[concurrency-interest] Questions on ConcurrentHashMap

David Holmes dcholmes at optusnet.com.au
Thu Nov 29 22:14:54 EST 2007


Hi Joe,

> I found the actual documentation later.  It's up in the into section,
> not down in the method documentation which is apparently an artifact of
how javadoc works.

Well we could have redefined the method Javadoc to include additional text,
but instead we just put a couple of general sentences in the class javadoc.

> Sidestepping the issue of almost nobody knowing how to specify semantics
> for concurrency, one of the few things that is understood is atomicity. So
> hypothetically you could say something like all mutating operations are
atomic
> w.r.t. other mutating operations, even the aggregate ones.

You could. But that's not the way the implementation works. All I was saying
was what could be said for the existing implementation.

Cheers,
David



More information about the Concurrency-interest mailing list