Meyer, OO and concurrency

[Thomas Lindgren]

--- todd <todd@REDACTED> wrote:

> When we have people litteraly searching out
> nanoseconds in their 
> algorithms and doing everything possible to
> guarantee task switch 
> latency, then I do remain unconvinced.

The sensible approach to this argument is probably to
concede that sort of problem to the low-level
programmers (for now :-) and use Erlang for more
suitable work.

My impression is that Erlang even so produces quite
respectable performance in many real applications. For
instance, my most recent effort, a mobile network
node, exceeded the performance requirements by 2x (at
least -- that's when the testers stopped). For this
sort of application, there's really no need to wring
out nanoseconds or fret about task switching times. In
fact, doing so is just wasting time and money.

Best,
Thomas


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com