Re: [RTF] AC019 proposal to WSA WG

On Wed, Jul 10, 2002 at 07:35:56AM -0700, Cutler, Roger (RogerCutler) wrote:
> Or perhaps satisfying it has some inevitable negative
> consequences, which is another interesting result.
> 
> I may be mistaken, but I am gathering that you are claiming that the last
> alternative above is the case.

That's correct.

>  Wouldn't it be more productive either to
> suggest alternative solutions to the requirement itself, or document the
> limitations imposed by satisfying it, than to resist the requirement itself?

I've touched on what a REST based reliability solution would look like,
by talking about the value of idempotent methods and state transfer.
But for me to propose this would be asking the WG to accept a REST based
solution before we agree to accept REST itself.  While I'd like this to
happen, I would expect it would meet with more pushback than what I've
been suggesting.

> I see a strong need for reliable messaging, but I am also extremely
> interested in understanding the circumstances it is appropriate to use it in
> and the consequences of using it.

I'm glad to hear it.  FWIW, I believe that the larger the network, the
less appropriate "reliable messaging" is as a reliability solution.
Have you read the Waldo paper?  While it doesn't touch on reliable
messaging explicitly, the general issue of the problems with network
transparency are discussed in some depth.  I highly recommend it.

MB
-- 
Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
Ottawa, Ontario, CANADA.               distobj@acm.org
http://www.markbaker.ca        http://www.idokorro.com

Received on Wednesday, 10 July 2002 10:57:55 UTC