- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Wed, 02 May 2007 10:33:29 +0200
- To: Anne van Kesteren <annevk@opera.com>
- CC: HTML WG <public-html@w3.org>
Anne van Kesteren wrote: > On Wed, 02 May 2007 10:05:24 +0200, Julian Reschke > <julian.reschke@gmx.de> wrote: >> I know that nobody wants to hear that over here, but XML's draconian >> error handling seems to work well in many places. No, I'm not >> suggesting to use it for HTML5, but keep in mind that it *can* work. > > Actually, XML has character encoding issues (despite it being defined > clearly in the specification) and RFC3023 is mostly ignored in practice. > There are quite some feed readers which have a "non-draconian" XML > parser as well. I think XML 1.0 is pretty clear about character encoding. The problems as caused by RFC3023 overruling the encoding information of the payload (even if not encoding was specified in the Content-Type) are well-understood; the revisions of both RFC3023 and RFC2616 are likely to contain changes getting this right. Atom uses an application/...+xml content type, so the problem doesn't even appear. Furthermore, there is no such thing as a "non-draconian" XML parser. What you probably referring to is a "non-draconian XML parser", which is not an XML parser at all. That being said, the feed reading subsystem in IE7/Vista IMHO only accepts well-formed XML and seems to get away with it (ignoring the RFC3023 issue where almost everybody seems to agree it was a bad idea in the first place). Best regards, Julian
Received on Wednesday, 2 May 2007 08:34:01 UTC