Re: [uddi-spec] Proposed changes to UDDI Version 2 specificationsrelated to UTF-8 / UTF-16 issue

[Keisuke Kibakura <kibakura@jp.fujitsu.com>]

Good work, Claus, 

I have one issue. 
A requirement that a registry node shall use the same encoding 
as the initial request for response messages seems too strict. 

Adding support for UTF-16 in addtion to UTF-8 would not be a difficult
task since an middleware for soap and xml generally handles both encodings. 
But making a uddi registry be sensitive to a encoding of 
an incoming request needs some more work. 
If a new errata requires a UDDI registry change its encoding policy 
for a response message depending upon the initial incoming request,
it would have a certain influence on existing products/implementations. 

Thus, I recommend the last sentence in section 7.2, "Response messages 
that are received from an Operator Site shall use the same character 
encoding as the initial request message that was sent to the Operator Site.", 
and "Response messages must use the same character encoding as the 
initial request message." in section 4.4.2 in Operator's spec. 
be removed. 

Comments?

--
Keisuke Kibakura
FUJITSU LIMITED


>>>>> On Mon, 07 Oct 2002 19:59:22 +0200, 
>>>>> "Von Riegen, Claus" <claus.von.riegen@sap.com> wrote :

> All,
> 
> At our last conference call I took an action item to specify those changes for the UDDI Version 2 specifications that would be necessary to adopt the UTF-16 character encoding. Currently, UDDI allows UTF-8 only. I have not specified what would be necessary for UDDI Version 3. It can be done at a later stage accordingly to the outcome of this discussion.
> 
> First, the Programmers API Specification has to be changed as follows:
> 
> Section 3.1.7 "XML Encoding"
> Replace
> "Similarly, XML allows for a large number of character set encoding choices.  UDDI Operators are required to only support a single XML encoding - UTF-8, and will support all compatibility characters defined for UTF-8.  See appendix B for more information related to the use of byte order marks and UTF-8 and the way the UDDI SOAP implementations convert all requests to Unicode prior to processing."
> with
> "Similarly, XML allows for a large number of character set encoding choices. UDDI Operators are required to support both UTF-8 and UTF-16, and will support all compatibility characters defined for these character encodings. See appendix B for more information related to the use of byte order marks in Unicode and the way the UDDI SOAP implementations convert all requests to Unicode prior to processing."
> 
> Section 6.7 "Support for Unicode: SOAP listener behavior"
> Remove
> "The decision to use the UTF-8 encoding in all requests simplified the number of encoding variations that need to be handled within the XML interchanges used in this API specification."
> Remove
> "All such responses will be encoded in UTF-8."
> 
> Section 7.2 "XML Encoding requirements"
> Replace
> "All messages sent to and received from the Operator Site shall be encoded as UTF-8, and shall specify the HTTP Content-Type header with a charset parameter of "utf-8". All such messages shall also have the 'encoding="UTF-8"' markup in the XML-DECL that appears on the initial line.  Other encoding name variants, such as UTF8, UTF_8, etc. shall not be used.  Therefore, to be explicit, the initial line shall be:
> 		     <?xml version="1.0" encoding="UTF-8" ?>
> 		and the Content-Type header shall be:
> 				Content-type: text/xml; charset="utf-8"
> Operator sites MUST reject messages that do not conform to this requirement."
> with
> "All messages sent to the Operator Site shall be encoded either as UTF-8 or as UTF-16, and shall specify the HTTP Content-Type header with a corresponding charset parameter of "utf-8" or "utf-16". All such messages shall also have a corresponding 'encoding="UTF-8"' or 'encoding="UTF-16"' markup in the XML-DECL that appears on the initial line.  Other encoding name variants, such as UTF8, UTF_8, UTF16 etc. shall not be used.  Therefore, to give an explicit example for UTF-8, the initial line shall be:
> 		     <?xml version="1.0" encoding="UTF-8" ?>
> 		and the Content-Type header shall be:
> 				Content-type: text/xml; charset="utf-8"
> Operator sites MUST reject messages that do not conform to these requirements.
> Response messages that are received from an Operator Site shall use the same character encoding as the initial request message that was sent to the Operator Site."
> 
> Second, the Operator's Specification has to changed as follows:
> 
> Section 4.4.2 "XML Encoding"
> Replace
> "All XML messages must be encoded in UTF-8."
> with
> "All XML messages must be encoded either in UTF-8 or UTF-16. Response messages must use the same character encoding as the initial request message."
> 
> There is no need for a change of both the Data Structure Reference and the Replication Specification.
> 
> Best regards,
>  Claus
> 
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>