Re: Validating XHTML5 with XML entities

Hi Henri,

On Aug 28, 2008, at 9:41 AM, Henri Sivonen wrote:

>
> On Aug 27, 2008, at 23:59, Sam Ruby wrote:
>
>> It occurs to me that if somebody were to do the work of producing a  
>> full DTD for XHTML5, such work could be considered by this  
>> workgroup either to be included in the HTML 5 standard, or as a  
>> standalone document.
>
>
> I think that's not a good idea.
>
> Writing a DTD for XHTML5 for the purpose of getting entities doesn't  
> make sense unless browsers recognize it.

Browsers already do support character references in XHTML served as  
XHTML. We've been through this over and over.

> Making the browsers recognize it would gratuitously break  
> compatibility with existing browsers compared to just using one of  
> the legacy public identifiers that browsers do recognize.

The browsers recognize the HTML character references in XML  
serializations (i.e., other XHTMLs). It will only confuse authors if  
we don't confuse it in XHTML5.

> (If one is willing to break compatibility with existing browsers,  
> one might as well implement SVG output from WordPress using the  
> commented-out SVG-in-text/html scheme.

It is your desire to keep character references out of XHTML that is  
breaking compatibility with existing browsers.

> Writing a DTD for validation doesn't make sense when DTDs are dying  
> technology that is documented to be inadequate and is mostly  
> abandoned by the XML community and better schema technology exists  
> and an actual liberally-licensed schema using the newer and better  
> schema technology exists.

Leaving a DTD out of our spec is not the way to convey that  
information. Providing a DTD and discouraging reliance upon it would  
be a better approach. Providing alternate schema using other schema  
languages that highlighted their benefits would be an even better way  
to convey that message.

We've been over this again and again and these arguments do not hold  
water. By providing these schema we make the task of adopting HTML5  
(in whatever serialization or form) much easier. It is ridiculous to  
not make an official DTD (and other schema) to make a point about how  
we don't (or one of our WG members doesn't) like DTDs. We shouldn't be  
forcing authors to independently and meticulously enter internal DTDs  
into their documents or validation tools just so you can make a point  
about how you don't like DTDs (who does; I don't know anyone).

Take care,
Rob

Received on Thursday, 28 August 2008 12:36:35 UTC