- From: David Dorward <david@dorward.me.uk>
- Date: Fri, 06 Apr 2007 00:09:47 +0100
- To: Shane McCarron <shane@aptest.com>, www-validator@w3.org
Shane McCarron wrote: >> huh? That would be content sniffing and against Web architecture. >> The mime type has priority and for good reasons. > A media type in a header does indeed take priority... I know this, of > course. My point was that, as we all know, lots of XHTML 1 documents > get served as text/html when in fact they are really well-formed, valid > XML. Shame that, it mostly stops them being run through XML parsers. The RFC for text/html only mentions XHTML 1.0 documents that follow the additional guidelines in Appendix C though[1]. >> Imagine you want to serve the source code of this document sending it >> then as text/plain. With the above reasoning, it would mean that we >> do not respect the intent of the author. >> >> If you send it as text/html, it is HTML >> If you send it as application/xhtml+xml it is xml > If you send it as text/html, it *might* be XHTML. It *might* be XHTML 1.0 that conforms to the additional guidelines in Appendix C (assuming that it isn't non-standard). > Since we tell people they can do that. So, since we tell them that, > we should at least > provide some way to help them validate things that they serve up as > that. But the document you are serving up isn't XHTML 1.0 with Appendix C. [1] Well, it says "a profile of", but the note I mentioned earlier in the thread clarifies. -- David Dorward <http://dorward.me.uk/>
Received on Thursday, 5 April 2007 23:10:08 UTC