- From: Henrik Frystyk Nielsen <frystyk@microsoft.com>
- Date: Sun, 25 Mar 2001 10:52:14 -0800
- To: "'Michael Brennan'" <Michael_Brennan@Allegis.com>, <xml-dist-app@w3.org>
>I understand there has been some controversy over RFC 3023,
>but in my mind, if we are going to have protocols piggybacking
>on top of HTTP and MIME, they better be doing so in a
>manner conforming to those standards (unless the W3C elects
>to offer explicit guidance for handling this in another manner).
The controversy over RFC 3023 is not related to the use of the charset
parameter but the requirement to use a centrally registered short-name
for a URI. In fact the SOAP HTTP binding is fairly clear on the rules -
from [1]
HTTP applications MUST use the media type "text/xml"
according to RFC 2376 [3] when including SOAP entity
bodies in HTTP messages.
Both RFC 2376 and 3023 say that
Although listed as an optional parameter, the use of the charset
parameter is STRONGLY RECOMMENDED, since this information can be
used by XML processors to determine authoritatively the character
encoding of the XML entity.
>The ebXML spec is taking the appropriate approach to offering
>explicit guidance to developers in this regard, IMO:
>http://discuss.develop.com/archives/wa.exe?A2=ind0103&L=soap&D=0&P=4501
3
In fact the guidelines that you mention are in conflict with RFC
2376/3023:
The MIME Content-Type header for the Header Container
MUST have the value text/xml in accordance with the
[SOAP] specification. The Content-Type header MAY
contain a charset attribute.
According to [2], MAY is not the same as RECOMMENDED not to mention that
RECOMMENDED here is further qualified by "STRONGLY". IMO, it is better
to refer to the authoritative spec rather than getting into a situation
where such conflicts occur.
Henrik
[1] http://www.w3.org/TR/SOAP/#_Toc478383526
[2] http://www.normos.org/ietf/rfc/rfc2119.txt
Received on Sunday, 25 March 2001 13:52:52 UTC