[DPP-70] Should we remove Async.AsyncChan

[qnikst]

[Imported from JIRA. Reported by Tim Watson [Administrator] @hyperthunk) as DPP-70 on 2013-02-01 08:23:20]
AsyncChan is a bit of an aberration. You can use it just like AsyncSTM, but once you've obtained the result, it's not longer useful, because subsequent reads will block.

I think we should either think of a useful API for this - e.g,. a handle that can be used to refer to the results of multiple asynchronous tasks via a compositional API - or get rid of it altogether.

I'm open to suggestions here. If we do remove it, we can get rid of the abstraction layer in Async and pull AsyncSTM up.