[OKS] Async IO

[Bill Davidsen]

Sounds like the right way to go. I always thought blocking i/o was for the
benefit of the o/s, but with read-ahead and buffered writes the gain will
be essentially invisible for many things. I admit that the first
development team I ever joined used async io for everything, writing a new
o/s for GE, then in the mainframe business, so I may not be neutral;-)

This has been less of an issue with low cost process creation, several
processes using IPC and blocking io are probably no harder to get right
than a bunch of async io, but this is a good time to add the capability.

-- 
bill davidsen <[e-mail]>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.