Re: revisiting heading advice

2014-01-05 22:02, Leif Halvard Silli wrote:
>
> What about h1-h6 inside blockquote elements?

The blockquote element has a long history. From the beginning, it has 
been defined as indicating quotation from an external source and used as 
a way of indenting a block (which might casually be a quotation, or at 
least explainable as a quotation with some contrived reasoning). I don’t 
intend to raise again this issue in general; I am just pointing at the 
history of implementations, which have shown no serious attempt at 
applying the idea of quotation – blockquote is just an indented block to 
them, unless we count some experiments with implementing the cite attribute.

It is rather difficult to, even at the level of ideas, how a quotation 
should appear in HTML. Should it use the formatting in the original 
(there are no handy tools for this), or the formatting principles of the 
host document, or some mix? Presumably, authors are supposed to fix 
relative URLs if the quoted original is in HTML format. What about 
JavaScript? Should it be simply dropped? What if the quoted HTML content 
has id attributes clashing with those in the host document? My point is 
that when inserting HTML content from another source into an HTML 
document there are several tricky issues

The issue of headings is, at the practical level, mainly part of the 
general problem of formatting quotations. It seems natural to adjust the 
ranks of headings in quoted material so that the highest-ranking heading 
is one larger than the rank of the heading under which the quotation 
will sit in the host document. It is not unreasonable to require that 
authors do this by hand or by using special tools, instead of expecting 
browsers to do that. Authors need to consider much more difficult issues 
anyway, when quoting HTML content in HTML documents.

> Take this snippet from the
> example of that section:
>
> ]]<blockquote>
>    <h3>Bla</h3>
>    </blockquote>
>    <p>Baz</p>
>    <h2>Quux</h2>
>    <section>
>    <h3>Thud</h3> [[
>
> 1: Why did the editor place a *h3* element inside the blockquote?

I can’t tell, but it makes perfect sense according to my reasoning 
above. In the quoted document, “Bla” could be a heading of any rank, but 
in the context, it needs to be adjusted to rank (and level) 3. It would 
minimally help authors if they could retain the rank as in the original, 
but the cost is that old browsers (and AT) will not treat it that way


> 2: Are headings inside blockquote confusing to users?

They could, but it may be necessary, for the purposes of quotation, to 
quote headings, too (sometimes only a heading, namely when commenting 
specifically a heading and nothing else)

> The answer
> probably depends on whether the AT/UA properly makes the blockquote
> context clear to the users. OTOH, if that h3 *was* ”reset” to level 1,
> then that level jump could probably have the effect on the user that
> he/she gets notified that something happened. In other words: Switching
> to level 1 would probably help making the user aware that this heading
> does not belong to the outline of *this* page.

I think this would be a completely wrong approach to a potential 
problem. If AT does not render quotations properly, then this is a 
problem that should be addressed as a general problem, rather than 
focusing on a special case where a heading is included. Moreover, by the 
current rules of HTML5, too, the h3 heading should be handled as level 3 
heading – it is level 3 in the document, even though it is topmost-level 
within its own section.

Well, I might be wrong, but I think you got it wrong. This not 
surprising, given the complexity of the definitions.

> But, I would also encourage you to think about whether one should give 
> advice about which heading (level) to use inside blockquote and other 
> sectioning roots (such as td), in view of how things are *actually* 
> implemented. 

If we think how things are actually implemented, we should tell authors 
to use headings so that their rank corresponds to their level.

-- 
Yucca, http://www.cs.tut.fi/~jkorpela/

Received on Thursday, 9 January 2014 21:55:04 UTC