RE: Friendly amendment #2c [Re: Straw poll on "synchronous" definitions]

If you want to define synchronous first, then those examples should go down in the asynchronous part. For the synchronous part we could say: "A typical example is receiving the response on the same open socket"

Ugo

> -----Original Message-----
> From: Walden Mathews [mailto:waldenm@optonline.net]
> Sent: Friday, March 14, 2003 5:28 PM
> To: David Booth
> Cc: www-ws-arch@w3.org
> Subject: Re: Friendly amendment #2c [Re: Straw poll on "synchronous"
> definitions]
> 
> 
> I have two recommendations:
> 
> (1) reverse the nature of the definitions to the positive mode, so
> that asychronous is defined in terms of synchronous instead of vice
> versa.  This is mainly about refactoring out a logical double
> negative.  But something interesting happens with the examples.
> Please comment (see below).
> 
> (2) be careful about stating the nature of the relation of asynch
> to synch.  "Opposite" is ambiguous.  It's really set difference.
> If you subtract the synchronous cases from all r/r cases, you are
> left with the asynchronous cases.  Isn't that a clearer test?
> 
> Thusly:
> 
> Synchronous:
> A request/response interaction is said to be synchronous when 
> the request
> and response are chronologically coupled.  In other words, 
> the client agent
> has to "wait" for the response once it issues the initial 
> request.  The
> exact
> meaning of "wait" depends on the characteristics of the client agent
> (including
> the transfer protocol it uses).  Examples include waiting for 
> the response
> in a different thread, on a different socket or end-point, or 
> by polling the
> server.
> 
> Asyncronous:
> A request/response interaction that does not meet the constraints of a
> synchronous interaction (above) is said to be asynchronous.
> 
> 
> FWIW,
> 
> Walden
> 
>

Received on Friday, 14 March 2003 21:11:12 UTC