W3C home > Mailing lists > Public > www-tag@w3.org > September 2008

Re: rel=CURIE in RDFa, but rel=URI in Link:

From: Steven Pemberton <steven.pemberton@cwi.nl>
Date: Tue, 30 Sep 2008 15:14:19 +0200
To: "Julian Reschke" <julian.reschke@gmx.de>
Cc: "Shane McCarron" <shane@aptest.com>, "Dan Connolly" <connolly@w3.org>, "Henry S. Thompson" <ht@inf.ed.ac.uk>, "Jonathan Rees" <jar@creativecommons.org>, "www-tag@w3.org WG" <www-tag@w3.org>, "XHTML WG" <public-xhtml2@w3.org>
Message-ID: <op.uiapidc8smjzpq@acer3010.lan>

On Tue, 30 Sep 2008 14:52:08 +0200, Julian Reschke <julian.reschke@gmx.de>  
wrote:

> Steven Pemberton wrote:
>> ...
>>> Digging deeper shows that you rely on IE's content-sniffing kicking  
>>> in. This seems to depend on the extension being "html" (it won't work  
>>> with "xhtml", for example). So IE is interpreting the page as HTML,  
>>> not XHTML.
>>  Not me. That's someone else's page. I, like most people, tend to  
>> deliver XHTML to IE as text/html in most cases. But the point is that  
>> the authoring is XHTML, and the user experience at the end is correct.  
>> That's all I am worried about.
>
> Well, whether the user experience is correct depends a lot on authoring  
> practices (empty tags, scripts, ...).

Sure, that's what I meant with "forethought" in an earlier message. You  
have to make sure the XHTML follows certain guidelines; but that's true of  
CSS and Javascript too.

> So this is interesting for testing, but it wouldn't want to rely on IE's  
> content-sniffing habits for production use.

Ha! I don't know how you can ignore it! I once wrote a tutorial on HTML,  
with links to example files, and links to the source of the examples,  
where the two things were just the same file being served as text/html in  
one case, and text/plain in the other (preventing them ever getting out of  
sync).

On the browser I was using it looked fine, but then I went to IE, and of  
course the source links didn't work anymore...

>>> BTW: the page claims:
>>>
>>>    <meta http-equiv="Content-Type" content="text/html; charset=utf-8"  
>>> />
>>  Yes, presumably a mistake. It's a real pain that content-encoding is  
>> enmeshed with mime type in HTML. Lucky that HTTP headers are  
>> authoritative!
>
> But then, the page is relying on IE not trusting the header.

The HTTP headers are authoritative.

> As a matter of fact, it also *works* with
>
>    <meta http-equiv="Content-Type" content="application/xhtml+xml;  
> charset=utf-8" />

Yes, that's what the author should have used.

Best wishes,

Steven
Received on Tuesday, 30 September 2008 13:15:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:06 GMT