Re: The only name for the xml serialisation of html5

On 2007-09-30 02:51:19 +0200 Dean Edridge <dean@55.co.nz> wrote:
> Leif Halvard Silli wrote:
>> 2007-09-29 03:59:42 +0200 Dean Edridge <dean@55.co.nz>:

[ ... name logics ...]
>>> [... ] the most obvious and logical naming conventions for our spec:
[...]
>>> HTML 5 (text/html)
>>> XHTML 5 (application/xhtml+xml) or (application/xml)
> 
>> _Why_ this is is logical, semantic and simple seems to boil down to 3 
>> reasonable ideas: 	

[In short: names should 1) refer to the MIME type, 2) nourish the HTML/XHTML dicotomy 3) both names should contain the string 'html'.]

Dean, I would like to hear if I interpreted your repeated claims about logical, semantic, simple etc correctly. Or could you, finally, explain what the logic-ness, semantic-ness and simple-ness about your proposal is?

[... 'HTML 5 text' and 'HTML 5 xml' ... ]

>> 	HTML 5 text        (text/html)
>> 	HTML 5 application (application/xhtml+xml)/(application/xml)
>> [... or ] for the XHTML variant [...]
>> 	HTML 5 xml [...]

> Surely you're not serious with these ideas.
> How would these ideas be easily implemented in the real world?
> Isn't it already a complicated mess as it is?

Again - you did not _explain_ why your proposal is locial/semantic/simple. The same way, you cannot just say the magic world 'real world' and expect the real world to agree with you. If there is more than rhetorics in it, please explain. I mentioned the only 'real world' I know, where this name currently matters - namely the HTML 5 draft, and in that draft, neither mine or your name proposal currently occur.

Otherwise the idea (in singular form) builds quite directly on Karl's idea from 31. of August [1]:

>> For now, when I talk about serializations of html5, I use
>>     html5/html
>>     html5/xml
>> and try to use html5 only for the abstract model, not the syntax.

Currently, the draft has no ultra clear way for discerning between the semantics/the abstract model and the serialisations. But a model where 'HTML 5' is used for the sematnics, and 'HTML 5 xml' and 'HTML 5 text' would be used for the serialsations - would solve that question. Using 'XHTML 5' would not make that disction any clearer, anyway.

On Friday, 28th of September you replied to Karl, asking: «Can we call the XML serialisation of the spec XHTML 5 or not?» Then, consider that the HTML 5 draft doesn't ask for such a permission. It only uses 'HTML 5' once (in the title) and  never use 'XHTML 5'. (If we really mean that are defining 'XHTML 5', then 'XHTML 5' should be in the title of the draft!) The draft is only «referring» to the two formats as  «XHTML5» and «HTML5». From section 1.3. Conformance requirements:

«[...] one based on XML (referred to as XHTML5), and one using a custom format inspired by SGML (referred to as HTML5)»

My proposal would change the wording in those paranthesises into: (referred to as HTML 5 xml) and (referred to as HTML 5 text)». Or, if there is a point in concatenating the terms into one word: 'HTML5xml' and 'HTML5text'.

I have no special hang-up on whether we say 'HTML 5 html' or 'HTML 5 text'. Or whether we say 'HTML 5 xml' or 'HTML 5 application'. The important thing is that we should say 'HTML 5' and then add a qualifyer after that to signify whether we refer to the format «based on XML» or the «custom format inspired by SGML». 


>>   3. 'application' would in itself work as a warning against
>>      using the XHTML serialisation on the public web.
>
> Again, please tell me that you are kidding me! What's the use of XHTML if we 
> can't use it on the public web? I use it everyday.

I meant with 'public web' the same as (I suppose) Charles meant with just 'Web', when he, in another thread, compared «comaptibility with XHTML2» with «compatibility with the Web» [3].

You are serving pages as application/xhtml+xml every day for all UAs? Surely not. For compatibility with the Web, we have to serve XHTML as text/html - at least to some UAs. HTML 5 is meant to put an end to that hack.

[1] <http://lists.w3.org/Archives/Public/public-html/2007Aug/1205.html>
[2] <http://lists.w3.org/Archives/Public/public-html/2007Sep/0546.html>
[3] <http://lists.w3.org/Archives/Public/public-html/2007Sep/0446.html>
-- 
leif halvard silli

Received on Sunday, 30 September 2007 02:56:12 UTC