- From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
- Date: Tue, 17 Dec 2002 16:36:00 -0500
- To: www-ws-arch@w3.org
> -----Original Message----- > From: Mark Baker [mailto:distobj@acm.org] > Sent: Tuesday, December 17, 2002 4:20 PM > To: www-ws-arch@w3.org > Subject: Does RM make a qualitative difference? > > > > > If that's correct, is it enough to actually make a qualitative > difference to an application developer? Or are they still > going to have > to deal with lost messages? I believe it's the latter, which is why I > suggest that our time would be best spent focusing on how to help > application developers deal with reliable coordination. 0.5% of the traffic on the Internet is an enormous number. I wouldn't scorn it, even if your guesstimates are correct. But anyway, this discussion is being driven by two things: First, the fact is that people (BEA mentioned themselves here) are rolling their own SOAP headers to provide acknowledgement services. We're exploring the idea of finding someone in the W3C (perhaps XMLP, perhaps a new WG) who might want to produce a Recommendation on how to do this. Whatever the costs and benefits of this at a theoretical/architectural level, the benefits can only be magnified by having a "standard". Second, this doesn't obviate the need for a better architectural understanding of "reliable coordination" at the application level, or even of a W3C Recommendation on how to do that in a standardized way. It does appear that leveraging the SOAP header/extensibility model to do RM at the message level is "low hanging fruit" (especially considering ebXML's prior work in this area) and some want to harvest it. Considering the wealth of experience in this area, and the quality of the discussion so far, I see no reason to rein them in. I'd suggest that you kick off a discussion of how to deal with this issue effectively at the application level rather than trying to discourage those who hope to deal with it at the infrastructure level. There is room in the WSA document for both. As usual, I'll argue that the best way forward is to discuss when one or the other is more appropriate, or if they can complement each other, and not place them in theoretical opposition to each other.
Received on Tuesday, 17 December 2002 16:36:03 UTC