Re: xmlns in HTML5

On Fri, 17 Jul 2009 15:04:05 +0200, Julian Reschke <julian.reschke@gmx.de>  
wrote:

> Steven Pemberton wrote:
>> ...
>>> Well, the author can say it is anything they want, but that doesn't  
>>> change what it actually is.
>>>
>>> It is literally not possible to send XHTML5 as text/html, because as  
>>> soon
>>> as you label it as text/html, you are stating "it is HTML".
>>  I used to think that too, but then I realised that in the real world  
>> it is different. Browsers sniff, and media types are hard-wired into  
>> software, rather than being an extension point. You have to row with the
>
> This is misleading.
>
> (1) Browsers do sniff, but not in this case. If you sent XHTML labeled  
> as text/html, the browser indeed processes it as HTML. That's why  
> there's so much broken "XHTML" content in the wild; if it indeed would  
> be processed as XML, it would fail.

I was pointing out the general case: in the market place media types have  
failed. You cannot use them reliably to indicate the type of a document.  
We might want it to be so, but it just ain't so.

I agree about the bit about broken XHTML. I hope you don't imagine that I  
think this is a good situation.

> (2) Media types are supported as extension points. I'm successfully  
> using that (not for something related *HTML, though).

Great.

>> oars you have got. As I said, I send documents with media type  
>> text/html, not because they are necessarily HTML, but because I want  
>> them in the browser.
>
> ...because you want them in that browser that doesn't support XHTML.

I want it in front of my user, the person it is all about. I don't care  
how I get it there (I have preferences of course, but my aim is to get it  
to the user).

>> I agree that the document gets *processed* as HTML, but the document  
>> doesn't magically change type just because it gets sent with a certain  
>> media type.
>
> The document itself doesn't have a type. Type information is meta  
> information. In HTTP, that is done using the Content-Type header. From  
> the filesystem, I assume browsers look at the file name to decide what  
> to do (but again, for *HTML, not at the content itself).

You see, I am talking from the point of view of people, not machines. If  
the author says it is XHTML, then I assume they know what they are talking  
about; sure that is meta information, but personally I *can* derive that  
piece of meta information from looking at the document. If in order to get  
it to the user the author has to play tricks with the browser's little  
mind, that is fine by me. The purpose of the excercise is to get  
information to the user, not please the browser.

>> I'm not arguing about processing, I'm arguing about the document. And  
>> as far as I am concerned, when it comes to saying what sort of document  
>> it is, the author is most certainly normative.
>
> How can the author be normative, when there's no place to send that  
> information (except, in HTTP, through the content type)?

Like I say, media types have failed in the market place. Until the day  
that that has been fixed, we have to survive in a very regrettable  
environment.

If all the world has to offer is cowpaths, then you have to accept getting  
dung on your shoes every now and then.

Best wishes,

Steven

Received on Friday, 17 July 2009 13:39:31 UTC