Re: Proposed resolution: issues 78, 16

 Paul,
 I disagree that moving "root" into core SOAP is necessarily a
good idea. Serialization root is a data-serialization term and in
my opinion the only part of SOAP that deals with data
serialization is part 5 - encoding. Once we start moving stuff
from encoding to the core, it could get messy. And I understand
multirefs as a data-serialization thing, too.

 We certainly could only try to specify multirefs (and roots) in
the core and leave the rest to the encoding, but I feel this
couldn't be done simply and that it wouldn't restrain other
encodings too much.

 On the other hand moving "start" attribute to the core could
help, for example with ordering of header processing. Then
section 2, for instance, could say:

 1) check mU
 2) process the headers, start with the one referenced by "start"
or with the first one.

 But this would extend to "when you processed the first one, it's
removed and then the next first one to start with is the former
second one, unless the processed header specified other
ordering."

 You might naturally want to extend the "start" attribute so that
any part (header) could have a "follow" attribute that would
specify which part (header) should follow. I don't think we want
to go that way.

                            Jacek Kopecky

                            Idoox
                            http://www.idoox.com/




On Tue, 31 Jul 2001, Paul Denning wrote:

 > I propose that the "root" attribute be made a global attribute and the text
 > describing it me moved to a new subsection under 4.1 (e.g., 4.1.3).  (4.1.1
 > is encodingStyle, and 4.1.2 is versioning model).
 >
 > This would allow the "root" concept to be used even when section 5
 > encodingStyle is not used.  It could be used for serialization roots inside
 > the SOAP header and/or body.
 >
 > I also like the "start" attribute for either/both header and body
 > elements.  I don't think an rpc namespace is needed for this.
 >
 > I just noticed there is no section 4.1, so perhaps we need to add a section
 > 4.1 titled "Global Attributes".  Since "actor" (4.2.2) and "mustUnderstand"
 > (4.2.3) are also global attributes, perhaps they should be moved under the
 > proposed section 4.1.  That would create something like this:
 > 4.1 Global Attributes and Versioning
 > 4.1.1 encodingStyle
 > 4.1.2 Versioning Model
 > 4.1.3 actor
 > 4.1.4 mustUnderstand
 > 4.1.5 start
 > 4.1.6 root
 >
 > Paul
 >
 > At 06:04 PM 2001-07-30, David Fallside wrote:
 > >1.) It makes it possible to know which element in the Body is the RPC
 > >element without having to parse the entire Body first. [This was a
 > >disadvantage of using the "root" attribute from Section 5.6.]
 > >
 > >
 > >2.) It can be used with any encoding.
 > >
 > >
 > >3.) It does not interfere with other RPC conventions currently in use,
 > >since the "start" attribute would be defined only in the new "rpc"
 > >namespace.
 > >
 > >
 > >One problem with this solution is that it does not address the problem of
 > >determining "serialization roots" inside the SOAP Header.
 >
 >

Received on Wednesday, 1 August 2001 15:14:06 UTC