- From: Jonathan Marsh <jmarsh@microsoft.com>
- Date: Wed, 20 Dec 2000 09:42:53 -0800
- To: "'Martin v. Loewis'" <martin@loewis.home.cs.tu-berlin.de>, www-xml-xinclude-comments@w3.org
Thanks for the references. I agree with your conclusion that issue 67 appears bogus - it should be easily closed. > -----Original Message----- > From: Martin v. Loewis [mailto:martin@loewis.home.cs.tu-berlin.de] > Sent: Tuesday, December 19, 2000 4:14 PM > To: www-xml-xinclude-comments@w3.org > Subject: XInclude-67-determining-encoding > > > I believe there is no need to support specification of an encoding on > the include element, at least not if parse="xml". Instead, the > included resource will have its own encoding according to section > 4.3.3 of the XML recommendation. > > That says that entities encoded in UTF-8 or UTF-8 need no further > declaration; processors MUST be able to determine these encodings from > looking at the first bytes. It also says that entities using any other > encoding MUST have a text declaration indicating the encoding. That > mechanism works in all scenarios given in > XInclude-67-determining-encoding, in particular for nfs:, file:, afs:, > ftp:. > > Furthermore, if there was an encoding attribute on the include > element, the spec would need to explain which one takes precedence if > multiple encodings are given (e.g. in the include element and the HTTP > header, or in the include element and an UTF-16 little-endian BOM). > > Of course, the issue XInclude-62-text-encoding is different; text > resources don't have a specified encoding. Making the behaviour > implementation-defined may be good enough, though. > > Regards, > Martin >
Received on Wednesday, 20 December 2000 20:00:08 UTC