[whatwg] Interpretation issue: can <section> be used for "extended paragraphs"?

[Bjartur Thorlacius]

On 3/11/11, Jukka K. Korpela <jkorpela at cs.tut.fi> wrote:
> James Graham wrote:
>
>> On 03/10/2011 09:20 AM, Jukka K. Korpela wrote:
>>
>>> My question is: Is this acceptable use of the SECTION element, even
>>> in a flow that mostly consists of P elements, not wrapped inside
>>> SECTION elements of their own?
>>
>> If I understand you correctly, it is not the intended use of
>> <section> — i.e. <section> conveys a different semantic to the one
>> that you want — and could have a number of undesirable consequences.
>> In particular it would insert a (presumably untitled) entry into the
>> document outline.
>
> That's a good point that had not occurred to me. On the other hand, this
> "outline" thing seems to be somewhat theoretical for the time being...
> Browsers haven't implemented it, and the few online "outliners" I've seen
> mentioned seem to respond by cryptic error messages. This is astonishing,
> because the idea seems to be fairly simple.
>
> Then again, the "outline" concept virtually exists in HTML 4, too. You can
> construct a section nesting tree on the basis of the implicit sectioning
> defined by heading elements. But that's not something that browsers do, or
> authors care about. Maybe HTML5 changes this, somehow.
>
Table of contents may be quite useful, from time to time, and there
are at least
a few implementations generating them
[http://www.niquelao.net/headingsmap-my-firefox-addon/].

> But, admittedly, it would go against the intuitive idea of "section" to
> divide, say, a section so that some components are sections and some are
> just paragraphs, lists, or something.
>
> What if I used section markup for _all_ paragraphs in some context? Suppose
> I have section consisting of p elements, some of which are coupled with ol
> or ul elements (or something else). Then could divide the entire contents
> into inner sections, each containing either a single paragraph or a
> paragraph and something else. This sounds logically solid, though clumsy.
> The inner sections would effectively be "extended paragraphs", just with a
> simple p element as the sole content in trivial cases.
>
Sounds harmful to me. This brakes assumptions the outline algorithm, probably
among other things, rely on.