Re: Comments on SOAP 1.2 part 2

----- Original Message -----
From: "John Ibbotson" <john_ibbotson@uk.ibm.com>
To: <xml-dist-app@w3.org>
Sent: Monday, April 08, 2002 9:48 PM
Subject: Comments on SOAP 1.2 part 2


> Here are some comments related to part 2 of the specification. I have used
> the Editor's copy dated Mar 23 2002.
>
> Abstract
> There is no mention of features, transport binding framework or the HTTP
> binding so for simplicity remove the final sentence.

Or add text about features, transport binding framework and HTTP binding?
Not done yet.

>
> Introduction
> In bullet 4 there is no mention of sections 5 and 6 for the binding
> framework.
> Suggest new bullet 4:
> The SOAP Protocol Binding Framework (see 5 A Convention for Describing
> Features and Bindings and 6 Message Exchange Patterns) defines a way of
> describing SOAP bindings to underlying transport protocols.
> New bullet 5:
> The SOAP HTTP Binding defines a binding of SOAP to HTTP [2] following the
> rules of the SOAP Protocol Binding Framwork (see 7 SOAP HTTP Binding).

Done

>
> Section 2 The SOAP Data Model
> First sentence:
> "The SOAP Data Model represents information ....."
> No mention of what type of information the data model represents. Suggest:
> "The SOAP Data Model represents instance information ....."

Replaced 'information' with 'application-defined data' for consistency with
Section 1

>
> Section 2.1 Graph Edges
> Problem here in mixing the use of the word node. This section uses it in
> the graph theoretic sense, the last sentence of section 2 uses it as a
SOAP
> Node !! Is there another word the mathematicians use to identify a graph
> "node" ? Not sure there is.

Section modified to always use 'graph node' as opposed to 'node'

>
> Section 2.3 Values
> Bullet 1:
> Rewrite as "A non-terminal is known as "generic" if the labels of its
> outbound edges are not unique ......."

Not sure I agree. We currently say

    'A non-terminal is known as a "generic" if the labels of its outbound
    edges need not be unique (i.e. if duplication of edge labels is
   allowed).'

This allows people to treat things as generics ( and access by name or
position ) even if all edge labels are unique. Your amendment would rule
this out, I'm not sure we want to do that although I'm open to argument.

> Bullet 2:
> Rewrite as "A non-terminal whose outbound edges are distinguished solely
by
> their labels is known as a "struct"."

Done

>
> Section 3.1.3 Encoding compund values
> Typo space missing "..... an array node MAY have ....."

Done

>
> Section 3.1.4.3 Constraints on id and ref attribute information items
> We constrain the value of a ref attribute to be the value of exactly one
id
> attribute. Does the converse apply ? That we cannot have ids without a
> matching ref ?

Can't see any reason to enforce that.

>
> Apologies for missing the Friday deadline,

No worries, thanks for the feedback

Gudge

> John
>
> Emerging ebusiness Industry Architecture ,
> XML Technology and Messaging,
> IBM UK Ltd, Hursley Park,
> Winchester, SO21 2JN
>
> Tel: (work) +44 (0)1962 815188        (home) +44 (0)1722 781271
> Fax: +44 (0)1962 816898
> Notes Id: John Ibbotson/UK/IBM
> email: john_ibbotson@uk.ibm.com
>
>

Received on Monday, 8 April 2002 19:00:01 UTC