Re: XHTML1: Suggested improvements to Appendix C (PR#6232)

Ian,

Thanks for your comments.  In the call today we discussed how to try to 
address (what we felt was) your core objection - that of normatively 
defining handling for text/html.  While the group does NOT agree that 
this draft document normatively requires anything with regard to 
text/html, we do agree that the text in section 5.1 was poor.  We 
suggest the following new wording, also reflected in [1]:


    5.1. Internet Media Type

XHTML 1.0 documents SHOULD be labeled with the Internet Media Type 
"application/xhtml+xml" as defined in [RFC3236 
<http://www.w3.org/MarkUp/2009/ED-xhtml1-20090602/#ref-rfc3236>]. For 
informaiton on delivering XHTML 1.0 Documents to user agents that do not 
natively handle this media type, see the informative note [XHTMLMIME 
<http://www.w3.org/MarkUp/2009/ED-xhtml1-20090602/#ref-xhtmlmime>].

[1] http://www.w3.org/MarkUp/2009/ED-xhtml1-20090602



Ian Hickson wrote:
> On Thu, 28 May 2009, Shane McCarron wrote:
>   
>>> Do you have a link to the most recent editor's draft of the XHTML 
>>> specification so that I can see if the new text addresses my comments?
>>>       
>> Sure.  W3C XHTML 2 Drafts are publicly available via our drafts page [1] 
>> - within that page you will find an updated XHTML 1 editors draft we are 
>> preparing for PER - the latest is [2].  You might also want to look at 
>> the Working Group Note that was published earlier this year [3].
>>
>> [1] http://www.w3.org/MarkUp/Drafts
>> [2] http://www.w3.org/MarkUp/2009/ED-xhtml1-20090528/
>> [3] http://www.w3.org/TR/2009/NOTE-xhtml-media-types-20090116/
>>     
>
> Thanks.
>
> With respect to the issue quoted below (PR#6232), I do not find the 
> working group's response to be satisfactory, as the underlying problem is 
> still present in the XHTML 1.0 3rd edition draft.
>
> Specifically, I object to the following comment in section 5.1:
>
> :  XHTML Documents which follow the guidelines set forth in [XHTMLMIME] 
> :  may be labeled with the Internet Media Type "text/html" [RFC2854]
>  -- http://www.w3.org/MarkUp/2009/ED-xhtml1-20090528/#media
>
> My original comments stand, namely that I request that either this prose 
> be removed, disallowing XHTML documens to be sent as text/html, or that 
> the guidelines set forth in [XHTMLMIME] have the changes listed below made 
> to them, including in particular making the guidelines normative, and that 
> the text following the above phrase, namely:
>
> :  [...] as they are compatible with most HTML browsers.
>  -- http://www.w3.org/MarkUp/2009/ED-xhtml1-20090528/#media
>
> ...be changed to refer to "legacy Tag Soup user agents" or some similar 
> wording that admits that XHTML cannot be made compatible with HTML, only 
> with the error handling code of existing user agents.
>
> Thank you for responding to my comments.
>
>
>   
>>>>> I believe the XHTML1 spec is wrong to allow XHTML to be sent as
>>>>> text/html. While in theory XHTML1 can be made compatible with Tag Soup
>>>>> UAs while still being valid and correct, the reality is that few
>>>>> authors are able to do so.
>>>>>
>>>>> I recommend that the working group consider releasing another edition
>>>>> of XHTML1, that removes the ability to send XHTML as text/html.
>>>>>
>>>>> However, if the working group does not wish to do this, I believe the
>>>>> following changes need to be made to appendix C:
>>>>>
>>>>>   1. Make the appendix normative.
>>>>>
>>>>>   2. Change "on existing HTML user agents" to "on legacy Tag Soup user
>>>>>      agents" or some similar wording that admits that XHTML cannot be
>>>>>      made compatible with HTML, only with the error handling code of
>>>>>      existing user agents.
>>>>>
>>>>>   3. Change the suggestion that XML declarations should be omitted to
>>>>>      a more strongly worded recommendation, as XML PIs trigger quirks
>>>>>      mode in WinIE6 and are displayed verbatim on PocketIE.
>>>>>
>>>>>   4. Remove one of the duplicated sentences in "C.4. Embedded Style
>>>>>      Sheets and Scripts", and require that script and style blocks be
>>>>>      neither "commented out" (with <!--/-->), nor enclosed in CDATA
>>>>>      blocks, nor include any entities.
>>>>>
>>>>>   5. Add a section requiring that <tbody> not be omitted.
>>>>>
>>>>>   6. Change the "C.11. Document Object Model and XHTML" section
>>>>>      slightly so that it requires that scripts be aware that when
>>>>>      treated as XML, they should use the namespace-aware Core APIs,
>>>>>      and when treated as HTML, it should use the DOM1 Core APIs;
>>>>>      similarly, that all script compare tagNames and attributes by
>>>>>      lowercasing them first.
>>>>>
>>>>>   7. Require that stylesheets style the HTML element rather than the
>>>>>      BODY element.
>>>>>
>>>>>   8. Documents should not use the <meta http-equiv="Content-Type">
>>>>>      element. (Actually this applies to all XHTML.)
>>>>>
>>>>>   9. There should be no use of namespaces other than the XHTML one.
>>>>>      (This is true of all valid XHTML elements anyway.)
>>>>>
>>>>>  10. There should be no XML Stylesheet PIs anywhere. (See 3)
>>>>>
>>>>> Overall, I think the language should be made more strict ("MUST"s
>>>>> rather than "SHOULD" or "MAY"). Stricter requirements are a great help
>>>>> when evangelising the use of correct markup.
>>>>>           
>
>   

-- 
Shane P. McCarron                          Phone: +1 763 786-8160 x120
Managing Director                            Fax: +1 763 786-8180
ApTest Minnesota                            Inet: shane@aptest.com

Received on Tuesday, 2 June 2009 16:02:51 UTC