Lock-Free Actor-Based Flow Programming in D2 for GSOC2011?

[dsimcha]

== Quote from eris (jvburnes at gmail.com)'s article
> Walter et al...
> I'm interested in finishing my proof-of-concept multi-core actor/flow-based
> programming system and I'd like to do it for D2.  I've already implemented a
> .9 version in D1 and would like to leverage the additional features and
> concurrency system in D2.
> I'm currently tweaking it for lock-free, multi-core, reactor-based
> environment. The lock-free intercore message FIFOs are currently running
> around 50 MT (mega transfers) a second.
> Is it too late?
> eris

Can you please explain what this brings to the table beyond what's currently
available in std.concurrency and std.parallelism?  I'm very curious.