RE: Proposed text on reliability in the web services architecture

> -----Original Message-----
> From: Walden Mathews [mailto:waldenm@optonline.net]
> Sent: Thursday, January 09, 2003 10:04 PM
> To: Champion, Mike; www-ws-arch@w3.org
> Subject: Re: Proposed text on reliability in the web services
> architecture
> 

> Not sure about his example, but I believe we've established
> today that wrapping a legacy application is an easy and effective
> way to avoid a "ground up" redesign.  If not, please advise.

Well, it's not me you have to convince that it's easy.  It's all those
enterprise developers out there.  As far as I can tell they're clamoring for
something like WS-Reliability, ebXML Messaging, or something similar that
everyone supports to provide a reliable messaging substrate so they don't
HAVE to think too hard about this stuff.  Maybe they're misguided.  Time
will tell.  So far, I'd say that those who are making the argument that an
RM layer serves no purpose haven't made their case.  

C'mon folks, this is after all a consensus based industry consortium. The
industry consensus seems to be that RM is worth standardizing.  You've got
an ENORMOUS job ahead of you if you want to convince us to ignore this or
disparage it in the WSA document.

If instead you wish us to put in some sort of "best practice guideline" that
its dangerous to rely too much on an RM layer, and it's also desirable to
consider safe retrievals, idempotent updates, 2-phase commit protocols, and
generally put failure awareness and recovery logic into applications rather
than infrastructure, that has a reasonable amount of support.  Draft some
compelling text to this effect and let's discuss it.

Received on Thursday, 9 January 2003 23:15:13 UTC