W3C home > Mailing lists > Public > xml-dist-app@w3.org > April 2004

Re: Binding framework & XML Version / Infoset

From: <noah_mendelsohn@us.ibm.com>
Date: Fri, 23 Apr 2004 08:22:44 -0400
To: Yves Lafon <ylafon@w3.org>
Cc: xml-dist-app@w3.org
Message-ID: <OF75003075.8E9E7CA4-ON85256E7E.0055444A@lotus.com>

Answering my own mail, the proposal I sent a couple of days ago failed to 
deal with this very question (link not available as I'm offline, but it 
was sent to distApp on 04/21/2004 12:06 AM EDT, titled "Proposed text for 
issues rec20 and rec22").  Specifically, if a sender prepares an outbound 
message infoset, or if an intermediary attempts to relay such an infoset, 
and if the usual HTTP binding is being used with support only for 
application/soap+xml, in what way should the HTTP binding fault?  I think 
we should document this.

In fact, I should own up to the fact that my proposed text went a little 
further than fixing the fundamental error in SOAP 1.2, and signalled our 
intention to allow XML 1.1 etc.  I think this is the right long term text, 
but whether to do it all in an erratum, or to split between an erratum and 
a SOAP 1.2.1 I see as a close call.  If you want to allow only XML 1.0 in 
SOAP 1.2 and reserve XML 1.1 etc. for a future SOAP 1.2.1, then I think 
the necessary adaptations to my proposed text are straightforward.   Was 
this choice discussed on the Wed. telcon?  Sorry I missed it.  Either 
direction (go straight to XML 1.1 & future versions in the erratum, as I 
proposed, or XML 1.0 only in the erratum) is OK with me as long as we are 
absolutely clear what the rules are.  Thank you, and sorry for any 
confusion.

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------






        Noah Mendelsohn
        04/20/2004 09:53 PM
 
                 To: Yves Lafon <ylafon@w3.org>
                 cc: xml-dist-app@w3.org
                 Subject: Re: Binding framework & XML Version / Infoset


I wonder whether we shouldn't say something about what the HTTP binding is 

to do if it supports only the mandatory application/soap+xml and is 
confronted with an infoset containing XML 1.1 or later content?  Does it 
fault?  How?  I think this needs to be synced up with the erratum text for 

SOAP 1.0.  I'm supposed to be preparing that, and will try to get to it 
tonight.  I am on vacation this week and there is some chance I will be a 
bit delayed in fulfilling that action, depending on how extensive the 
changes appear to be.

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








Yves Lafon <ylafon@w3.org>
Sent by: xml-dist-app-request@w3.org
04/20/2004 11:39 AM

 
        To:     xml-dist-app@w3.org
        cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Binding framework & XML Version / Infoset



In Part 1, section 4:
<<
A SOAP binding specification:
[..]
* Describes how the services of the underlying protocol are used to 
transmit
  SOAP message infosets.
>>>
And then in 4.2:
<<<
The binding framework does not require that every binding use the XML 1.0
[XML 1.0]  serialization as the "on the wire" representation of the XML
infoset; compressed, encrypted, fragmented representations and so on can
be used if appropriate. A binding, if using XML 1.0 serialization of the
XML infoset, MAY mandate that a particular character encoding or set of
encodings be used.
>>>

So it looks like the default is to provide XML 1.0 serialization and if
something else is used, it has to be defined explicitely in the binding
specification.

In Part 2, the HTTP bindings does not define one single serialization
scheme, but allow different ones that can be identified by the mime type
used, but a conforming implementation MUST support application/soap+xml,
which is now restricted to XML 1.0 serialization.
In this case, the serialization scheme is not _explicitely_ stated but is
implicitly part of the media type definition, we may perhaps addsome text
in Part2 - 7.1.4 to state that the media type definition may have an
impact on the XML version used in the serialization.

I also noted that in section 5, part1, the definition of white space
characters always refer to XML 1.0, should this be changed to only "XML"
(or "the XML in which the infoset will be serialized in"?).
Thanks,

-- 
Yves Lafon - W3C
"Baroula que barouleras, au tiéu toujou t'entourneras."
Received on Friday, 23 April 2004 08:24:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:16 GMT