W3C home > Mailing lists > Public > public-html@w3.org > October 2008

Re: UA style sheet for <q>-- why required?

From: Robert J Burns <rob@robburns.com>
Date: Thu, 30 Oct 2008 23:07:52 -0500
Cc: "HTML WG" <public-html@w3.org>
Message-Id: <994D1639-974C-42D7-B161-8796F598CB5D@robburns.com>
To: Jim Jewett <jimjjewett@gmail.com>

Hi Jim,

On Oct 30, 2008, at 7:27 PM, Jim Jewett wrote:

>
> On Thu, Oct 30, 2008 at 7:37 PM, Robert J Burns <rob@robburns.com>  
> wrote:
>> On Oct 30, 2008, at 6:13 PM, Jim Jewett wrote:
>
>>>> I've seen no one expound a compelling reason for not leaving  
>>>> quotation
>>>> styling to a styling mechanism instead of hard coding the styling  
>>>> in
>>>> the HTML document.
>
>>> Correctness.  The HTML is supposed to be usable even with  
>>> stylesheets
>>> off.
>
>>> If Bob says <q>I'll be a monkey's uncle!</q>, then the stylesheets  
>>> are
>>> effectively required. ...
>
>> This still is not a reason. When stylesheets are off that means  
>> that the UA
>> is using the UA stylesheet to present the quotations.
>
> The UA stylesheet is still a stylesheet.  I really did mean "without
> stylesheets"  -- the sort of display that you get for an unknown
> element, where the element name and attributes are effectively lost.
>
> Perhaps this is no longer a reasonable goal; it already doesn't
> work for object, audio, and video, and sometimes not for img.
>
> Still, it seems odd to break that tradition for a single inline
> text element.

I don't really think there is such a tradition.  Without a default UA  
stylesheet (or some equivalent styling mechanism) then the best a UA  
could do would be to present the DOM tree as simply a tree (not even  
able to discern block from inline frames, though perhaps mixed content  
in an element provides some hint). The problem of unknown elements is  
a separate issue (like video and audio but not object or img). For  
text/html parsing, the browsers all parse those unknown elements in  
diverse ways including IE which treats all such tags as separate void  
element including the </closing> tag parsed as a "/closing" void  
element.

>
>
>>> If Bob says <q>I'll be a monkey's uncle!</q>, then the stylesheets  
>>> are
>>> effectively required.
>
>> will be rendered properly using the UA stylesheet.
>
> It isn't today, so there is not a good transition plan for authors.
>
> You know that (as yoiu mentioned in [2]), but figure it is OK,
> because the spec said "should not".
>
> I think it was the spec's advice that was wrong, and authors were
> entirely justified in ignoring that particular "should not", as they
> did indeed have good reason to do otherwise.

I think the spec's advice makes sense, but I agree that authors have  
been largely justified in ignoring the should not because IE did not  
follow the MUST requirements of the spec. Even as we're now reaching  
the goal of all major browsers supporting quotes, two reasons remain  
for ignoring the SHOULD NOT recommendation of HTML 4.01:

1) content targeting legacy browsers (particularly IE<8) needs to  
include quotation marks
2) authors desiring to present quotations according to the many edge  
case presentations we've enumerated (French, Russian collapsed final  
quotations marks, French quotations with omitted interceding quotation  
marks), will need to include such styling in their HTML documents  
until CSS adds support for these styling conventions (and browsers add  
support for that CSS).

> That said, if the support for reduction-to-text is going to be  
> dropped,
> then HTML5 is the right time.
>
>> [removing the explicit quotes] has been discussed at length in
>> this thread[1][2].  It is not entirely clear this is the best way  
>> forward,
> ...
>> [1]: <http://lists.w3.org/Archives/Public/public-html/2008Oct/0241.html 
>> >
>> [2]: <http://lists.w3.org/Archives/Public/public-html/2008Oct/0260.html 
>> >
>
> [1]  Is really about getting the visual presentation right to the last
> detail.  That is certainly nice to have, but it is too much too ask of
> a non-stylesheet view.  It is probably too much to ask of even a
> default stylesheet.

I agree for default styling. I think that UA support for CSS2 and  
HTML4.01 is all that is needed for UA default stylesheet styling of  
quotations. However, a separate issue is to encourage CSS WG to  
provide sufficient properties to span all of the quotations styling  
mechanisms internationally even if those are not needed for default UA  
stylesheet styling.

> I'm thinking not just of the big browsers, but of smaller tools, like
> plucker.  (www.plkr.org)  Arcane rules in the error-correction section
> are bad, but unavoidable.  Odd exceptions within even the section for
> perfectly valid documents is worse, and raises the barrier to entry.

I'm not clear on what you're saying here. However, since support for  
CSS 2 and HTML 4.01 is all we need for sufficient default styling of  
quotations, then I don't think that should be a barrier to entry. The  
proposed duplicate quotation mark elimination algorithm might add some  
complexity, but i don't think it's a show stopper. Since the :before  
and :after CSS properties tell the UA precisely where to look for  
duplicate quotation marks, the UA simply needs to check for matching  
immediately adjacent quotation marks on the boundary of the element  
where the :before and :after pseudo elements are applied (ignoring  
whitespace) and if found omit those quotation marks (the glyphs) from  
the rendered view. Such an algorithm (as Chris Wilson suggested)  
provides authors with the styling flexibility originally envisioned by  
HTML and CSS2, even while supporting legacy UAs by including the  
presentational quotation marks in the HTML source.

Take care,
Rob
Received on Friday, 31 October 2008 04:08:37 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:58 UTC