- From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
- Date: Thu, 12 Dec 2002 20:21:13 -0500
- To: Mark Baker <distobj@acm.org>
- Cc: www-ws-arch@w3.org
> -----Original Message----- > From: Mark Baker [mailto:distobj@acm.org] > Sent: Thursday, December 12, 2002 1:44 PM > To: www-ws-arch@w3.org > Subject: Closing issue X > > > > In the interest of closing this issue so we can all get on with our > lives, I'm going to take a stab at providing some closing text based > on my best understanding of why Web services proponents believe that > REST's uniform interface constraint isn't important to the problems > they're trying to solve. Mark, The WSA WG considered your suggestion at some length today. We decided that it would NOT be a responsible act on our part to close this issue in a perfunctory manner. We will deliberate on how to address your original suggestion as the WS architecture document evolves and as we consider each of the issues that have been raised against it. So, we are NOT taking the procedural cop-out that I suggested earlier, we are saying that this isssue is too important to close without carefully crafted reasoning that references the WSA document. We believe that our time is now best spent on refining that document so that we have a strong foundation on which to argue our case if this issue ever gets to the TAG. Speaking only for myself, I suspect that the TAG would have plenty of fodder to chew on considering only the human-readable web as it actually exists, assuming they choose to consider the role that the uniform interface constraint might play in the Web architecture document. If, hypothetically, they choose to believe that this is an important principle of the Web architecture, we can at some future point consider how it might apply to the Web services architecture. Also, recall that SOAP's Web method feature, once supported in WSDL and widely deployed in actual tools and applications, may take us considerably closer to the goal that many of us share to more closely align Web services with the features of the Web as we know it. That will take time. Finally, to flog the Horse Tartare once more, I suspect it will be much easier to get Web services developers to agree on the pragmatics of making idempotent services GETable, and on the "best practice" of using the URI to identify a specific Web service resource in great detail, than on the theoretical principle of a "uniform resource constraint." If you have them by the wallet, their hearts and minds will follow :-) regards, Mike Champion
Received on Thursday, 12 December 2002 20:21:20 UTC