In response to Tony Davies question: Whether continuation & stream I/O
are equally expressive I'm not actually sure, I have been told they're
not & that continuation is more so - but I don't know they argument,
however certain things are definitely a LOT easier using continuations
so they certainly "feel" more expressive! But I'm not sure
expressiveness is the right test (isn't assembler very expressive :-))
- for instance synchronisation between reqests & responses and between
transactions is "guaranteed" by software for streams and by the I/O
semantics for "result" continuations. (This is one reason why the
"stream" module supplied in Hope+C isn't real streams - if you get
your request/responses out of synch it sends you a rude message - a
REAL stream system deadlocks/screws up in some other way :-) :-))

Joes question: No difference that I know of between continuation &
stream models for parallelism. But concurrency (user specified &
needed parallelism) is easier to specify/control in the continuation
style - but then you could argue that is merely a subjective view.

Nigel
---
Dr Nigel Perry                    Email: [EMAIL PROTECTED]
Department of Computer Science    Tel: +64 6 356 9099 ext 8900
Massey University                 Fax: +64 6 350 5611
Palmerston North
New Zealand

Reply via email to