W3C home > Mailing lists > Public > www-html@w3.org > August 2002

Re: Section Separator (was In support of the "line space" (nee <hr>))

From: Alexander Savenkov <w3@hotbox.ru>
Date: Sun, 25 Aug 2002 17:05:26 +0400
Message-ID: <13615007060.20020825170526@hotbox.ru>
To: www-html@w3.org, Bill Daly <billdalynj@yahoo.com>, Lorenzo De Tomasi <lorenzo.detomasi@libero.it>, fantasai <fantasai@escape.com>
CC: Micah Dubinko <MDubinko@cardiff.com>, Tim Bagot <tsb-w3-html-0005@earth.li>

Hello everyone and sorry for the delay.

Trying to cover all the hr-related mails.

Bill Daly:
> Which leads me to wonder, if hr is going to be
> removed, then would there be a need for change in the
> way borders are implemented?  I'm sure there are cases
> where we wouldn't want the border having the full
> width of the section.
Me too. However is there a real need for this? The 'hr' shading cannot
be achieved via CSS, is this bad? I don't think so. Anyway, if
somebody wants to customize the border look it's time to refer to CSS.
I believe XHTML2 should not be burdened with historical elements only
because CSS cannot offer (for now) some mechanisms to replace them.

What about me, if there's a strong wish to have a half-width or so
border I guess one may be able to come across some weird combination
of padding, margin, float, or position properties of CSS2.
Furthermore, CSS3 is not out yet perhaps it's time to let the folks
know about the issue. No need to place empty div's into the markup
(as Toby suggested).

Lorenzo De Tomasi:
> I think that <hr> is immediate, everyone can use it easily.
> To use css people must learn css and sometimes it's not simple (like in this
> case).
Usability and simplicity are for sure one of the most important
principles of the W3C. However, do you really want to put "CSS is not
simple" as a crushing argument for leaving 'hr' in?

> Following this way isn't xhtml becoming too much complex for common users?
> especially in this trasitional period?
If you find it too complex (and I guess a good piece of code is not
just "yay, lemme put all these into the body, where're my font tags,
here they are, it's easier to put'em here than to go read CSS, well
done, how do I make paragraphs, oh yeap, the br element's the right
one" etc.), it's time for you to switch to some standard-compliant
WYSIWYG stuff instead of hand coding and requesting for presentational
and similar crap which is fighted against for years.

> I have great dubts. This e-mail is only a list of them without much logic,
> maybe difficult to understand.
...

Micah Dubinko:
> For <hr>, I propose the following decision tree:
> 
> 1. Does <hr> have any legitimate non-presentational uses?
>    If no, get rid of it entirely.
>    If yes, rename it and preserve the correct aspects of it.
Entirely agree.

> 1. English-language-specifics
> 
> Is the "line space" technique specific to English-language texts? Can anyone
> provide examples for/against?
The variety of ways to separate blocks of information is great. Giving
the palm to the horizontal rules seems to be one of many English or
rather country (raise your hands) specific rudiments pretending to be
world-wide. A page break, an image, an empty space, a horizontal
ellipsis, almost anything can serve as a separator.

Instead, separator should depend on the content (read: 'section').

> 2. Unnamed <section>s
> 
> Does an unnamed <section> capture the same meaning?
> 
> There will always be many answers to the question of how to map
> chapters/verses, etc. to <section>s.
> 
>> I. e. the author considered the second subsection too small
>> to give it a heading.
> The part I quoted was just the very end of one block of text and the very
> beginning of the next. And this particular story doesn't have any named
> sections at all (besides chapters),
Sure it does not. You can't see the underlying markup in the book. In
fact traditional maker-ups have no structure in mind. It's their task
to position things on the only specific page. This approach satisfties
them as a paper book can't be presented on other media.

*On the other hand*, authors of the web documents should *markup* the
text at first and apply the styling later. If we could see the markup
of the book you cited I'm sure we would see the section
margins/borders where you think there's just an empty space (asterisks
or whatever).

> which is pretty common (at least in my
> library). So it wasn't that the author thought the subsection was "too
> small", but rather that a 'pause' was necessary before continuing. An
> unnamed <section> doesn't indicate any kind of a pause.
It does. Do you mean if a section has no 'h' there should be no
margins, aural cues, and such?

Tim Bagot:
> I'd consider
> it pretty odd to go straight from one section to the next without any sort
> of break, in the same way that I'd consider it pretty odd not to indicate
> paragraph breaks in any way (and naming paragraphs is certainly unusual
> for most types of document). The exact way in which such a pause is
> presented to the reader (or listener, etc.) is a matter for style sheets,
> but in any sane one it would definitely be there.
I agree with Tim. It doesn't matter whether the 'section' has a 'h'
element or not. The default stylesheet will provide styling for both
'section's and 'h's.

>>The example [of 3 asterisks] you give dates from the old days.
> 
> Well, 1999 in any case. :-) It's written from the viewpoint of a writer,
> where there is still a strong expectation to deliver dead-tree manuscripts.
> I wanted to show one possible presentation of this element.
Hmm, what I wanted to say was "There are now other ways to separate
blocks of information on the Web, e. g. borders, margins, parantheses
etc."

>>Adding a new empty element with no information inside is
>>not a good thing for the future Web
> 
> Just curious, why do you say this? Sometimes the lack of information carries
> meaning.
Examples are welcome. An empty table cell? Even this one, strictly,
speaking has to contain some information, a - or a x (which are
different and represent the lack of information, or inapplicability).

> What I really don't want to see is:
> <p class="center">*&nbsp;&nbsp;&nbsp;*&nbsp;&nbsp;&nbsp;*</p>
No one is proposing that. (Hello, Lorenzo)

Fantasai:
>> I rather doubt it. They are not part of the story while the paragraphs
>> they separate are. I would say you're confused by their presentation.
>> In fact they're plain subsections, but unlike the first one ("Graff
>> seemed...") the second ("He was...") has no heading. I. e. the author
>> considered the second subsection too small to give it a heading.
> 
> This is not always the case. In less formal writing, such as novels
> and newspaper articles, section breaks exist where the two sections
> are not always clearly defined.
They will exist exactly where they (somehow) managed to define breaks.
Enclose the text before and after the break into 'section's and here
we go.

> Sometimes the text immediately below
> a header refers to the header but gradually flows away from the
> topic so that it no longer refers to the header--but there's no
> exact point where you can say *here is where this section ends*.
> 
>> <section>
>>  <h>Title of the book</h>
>>  <section>
>>   <h>Title of this subsection</h>
>>   <p>...Some text here</p>
>>   <p>Graff seemed unconcerned. But then, Graff always seemed
>>   unconcerned. The next day everything changed. Graff went away, and
>>   in his place they gave Ender a companion.</p>
>>  </section>
>>  <section>
>>  <p>He was in the room when Ender awoke in the morning. He was an old man.</p>
>> </section>
> 
> Most novels don't have subsections. If you meant "chapter",
I just meant the 'section' element.

> then
> your structure is wrong because, IIRC, the last paragraph is part
> of the same chapter as the other paragraphs.
I see you weren't able to get the point as I mistyped the example.
I meant the following:
...
<section>
 <h>Title of the book</h>
 <section>
  <h>Title of this subsection</h>
   <section><!-- Can't see me on paper -->
    <p>...Some text here</p>
    <p>Graff seemed unconcerned. But then, Graff always seemed
    unconcerned. The next day everything changed. Graff went away, and
    in his place they gave Ender a companion.</p>
   </section>
   <section><!-- Can see a white space or asterisks before me, or an aural cue -->
    <p>He was in the room when Ender awoke in the morning. He was an old man.</p>
   </section>
 </section>
</section>
...
This one looks better. Objections? Btw, I suspect it's the time to
reject comments like "hey, I'm not gonna mark all these". For those
people I would recommend marking everything up with 'pre' and line
breaks and minuses (for 'hr').

>> Using CSS2/3 selectors one can insert an image, a phrase, or anything
>> else instead of just 'hr'. Once again, neither the line itself, nor
>> line break can cary semantics. They tell nothing about the relationship
>> between paragraphs/sections.
> 
> A line break, however, can be the default presentation for a semantic
> element, and CSS can style empty elements as well as non-empty ones.
> <break> would perhaps be a better name. A speech browser would render
> it as a pause, a visual one as a horizontal rule or three asterisks
> or an extra line break.
An empty space after section can be the default presenation for
'section' margins. A speech browser would render it as a pause, etc.

In any case, what does a 'section' imply? I presume it's a block of
information. The 'section' depth level tells how big the block is. The
presence or absence of the 'h' element explains if the author wanted
to name this particular small subsub(sub...)section or not.

> Typically this is a presentational, rather than a structural device.
I disagree. It's always presentational.

> The headings aren't really there to convey information, but rather
> to break up a long column visually.
Let's welcome 'section's or 'p'aragraphs, depending on your neeeds.

Fantasai:
> I've seen them on short columns, too. They do convey structure:
"They" are not breaks, but 'section's.

> they show
> where there's a break in the flow of the text.
You missed my point.

> Have you ever seen a section break between a spoken question and its
> answer?
What's wrong with paragraphs now?

> I doubt it, because there's no natural break between a question and its
> answer in the flow of dialog text.
Paragraphs should be used for them. A change of the speaker does not
imply a 'section' break. Use paragraphs.

> If a section break was presentational,
> it would have no relation to what's happening in the text--and the layout
> person, not the author, would put in the breaks.
I'm tired of repeating a line break means nothing. Look, if I take a
'hr' out of the context how do you tell me its meaning? How are you
going to pronounce asterisks? It's completely meaningless as *it*
carries no information ('section's do).

Perhaps the XHTML 2.0 has to explicitly state that the 'section'
element is not necessarily a section. It divides the document into
blocks of structured information and that's it.

>> I think newspapers almost always use them arbitrarily, although novels
>> may use them in a half presentational, hafl structural, manner
> 
> You mean section breaks make the page look pretty? I really doubt that.
> The author doesn't even know where section breaks fall when s/he sends
> in the manuscript--and often not even whether they'll be rendered as
> extra space between paragraphs, asterisks, or horizontal rules. The
> author only knows what the text is, yet still can point out where
> there is a break.
AWGTHTGTTA? I always thought the author is the only person knowing
where particular changes take place. If a publisher/editor inserts a
line break in the text without notifying the author I would say the
publishing house is wicked.

> Think about this: A voice is not a visual broadcaster. If the break was
> purely presentational, why would it be appropriate for a speaker to pause
> at the break--and even awkward in many cases if s/he did not?
Because the pieces of text (or anything) that surround this break do
carry structure.

Again, empty elements without attributes *of any kind* cannot cary
meaning because:
a. Meaning is expressed via information.
b. Information is expressed via data.
c. If there's no data there's no meaning.
Tell me now, where's the data in 'hr'? Or where's the data in 'br'?

The letter turned out to be a pretty long one, sorry for that.

Regards,
---
  Alexander "Croll" Savenkov                  http://www.thecroll.com/
  w3@hotbox.ru                                     http://croll.da.ru/
Received on Sunday, 25 August 2002 09:10:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:52 GMT