- From: Henrik Frystyk Nielsen <henrikn@microsoft.com>
- Date: Wed, 17 Jul 2002 17:28:32 -0700
- To: "Paul Prescod" <paul@prescod.net>
- Cc: "Paul Denning" <pauld@mitre.org>, <www-archive@w3.org>, <moreau@crf.canon.fr>, "Martin Gudgin" <mgudgin@microsoft.com>, "Marc Hadley" <marc.hadley@sun.com>
>There is one important part I do not understand. Why is this "best >practice" under the title "RPC" in both the SOAP primer and the SOAP >Part 2? If you were using a pure document-style of interaction >would not the same best practice apply? Because a major difference between RPC and the Web is that RPC doesn't provide the concept of identifying resources with URIs which is why the section starts: "The World Wide Web identifies resources with URIs, but common programming conventions convey identification information in the arguments to procedures, or in the names of those procedures." That is, in plain old HTTP, it is the common mode of operation to identify resources using URIs. >I believe that SOAP's definition of RPC is: "the exchange of messages >that map conveniently to definitions and invocations of method and >procedure calls in commonly used programming languages". > >If so, why would proper use of URIs on the Web and in HTTP be tied to >RPC? As you know, many of us see disciplined HTTP usage as an >*alternative* to RPC! Yeah, I have heard about the goals of HTTP ;) >Until I hear back, I consider this a bug, but my level of stridency on >the issue will depend upon the cost of fixing it. From a spec point of view, I actually think this does address the issue. >The primer, especially, seems quite easy to fix (at least >technically, I >don't know about procedurally). Merely replace the word "RPC" with >"message" or "message exchanges" (or some more appropriate word of your >choice). "Conveying web-friendly message exchanges" and "There are many >instances where message exchanges are designed for uses which..." > >If something like that were done I would not push for any >changes to the >normative specification. > >Also, the primer section is a little bit confusingly written in that it >seems as if 3.1.3 is entirely about "pure information retrieval" >applications until you get close to the end. If it is still possible to >do some re-arrangement that does not change the meaning then I would be >glad to propose alternate text. If you believe there are editorial improvements to the primer then I suggest you generate a new issue as this is somewhat different from this issue. Thanks, Henrik
Received on Wednesday, 17 July 2002 20:29:04 UTC