W3C home > Mailing lists > Public > public-html@w3.org > June 2009

Re: Footnotes and sectioning roots. <figure> and <table>

From: Ian Hickson <ian@hixie.ch>
Date: Fri, 5 Jun 2009 23:04:18 +0000 (UTC)
To: Leif Halvard Silli <lhs@malform.no>
Cc: HTMLWG <public-html@w3.org>
Message-ID: <Pine.LNX.4.62.0906052248540.1648@hixie.dreamhostps.com>
On Sun, 3 May 2009, Leif Halvard Silli wrote:
>
> The draft text has a section with recommendation about footnotes. Those 
> recommendations do not give any advice about footnotes inside elements 
> that belongs to the "Sectioning roots" category. Just as headers of a 
> sectioning roots element do not take part in the general outline of the 
> document, it also seems logical that links inside such elements should 
> be kept inside the element itself if the links are of "footnote nature".

I think it depends on the source of the notes. For example, one could 
imagine a <blockquote> element containing some sections and having 
endnotes that are included within the block quote. One could also imagine 
having some annotations in the quote that are actually asides from the 
quoter, rather than the quotee; the targets of those links would 
presumably be outside the element, as they are not part of the quote.

This seems very specific; do we really want to include text about this 
kind of thing? It seems that we're really reaching into the fringe of what 
is going to be useful here.


> In particular the figure element, which represents a sectioning roots 
> element that "can be moved away from the main flow of the document 
> without affecting the document's meaning", needs footnotes advice.  
> Because, if the footnote of a <figure> is placed outside the figure 
> element itself, then it isn't possible to move the <figure> out the page 
> without affecting the document's meaning. If the figure was removed, and 
> the footnote natured notes remained in the page, those notes would be 
> entirely meaningless. The draft should say that footnote natured links 
> inside a figure element should point to footnotes inside the figure 
> element itself. (This does not prevent that the figure could contain 
> links that points to other, independent texts outside the figure 
> element.)

You're misinterpreting "can be moved away from the main flow of the 
document". It doesn't mean "removed", it literally means "move away", e.g. 
to lower on the page.

If you removed the figure altogether, then anyone referring to the figure 
would have its meaning changed too. It's not limited to footnotes.


> Figure elements used as table containers should be given special 
> attention. Firstly, table footnotes is a feature that is sought for - 
> see for instance Ferg [1].  Text in table cells often needs to be short. 
> Thus a link to a footnote might be required to explain what the short 
> text means. Such footnotes needs to be close to the table.

I don't see why they need to be close to the table. Why can't they be at 
the bottom of the document, e.g. as in Wikipedia?

> [1] http://www.ferg.org/section508/htp04_proposal.html


> The best advice is probably that such tables are placed inside a 
> <figure> element, and then that the footnotes are placed immediately 
> after the table. It does not seem right to use <tfoot> for this, and the 
> draft should speak against use of <tfoot> for that.

I think you may be overthinking this.


> But such table footnotes would also affect the use of the <caption> 
> element, however, as the the draft currently says that "When a table 
> element is in a figure element alone but for the figure's legend, the 
> caption element should be omitted in favour of the legend." Thus, if 
> table footnotes are placed inside the <figure> element - as they should, 
> it is suddenly no longer conforming to use the <legend> as caption of 
> the <table>. Thus the draft should, in addition to <legend>, also allow 
> table footnotes inside such figure elements.

I don't see why the footnotes can't just be placed at the bottom of the 
page or some other place like that.


> Using <figure> as a table container begs another question however: 
> display. The CSS display of <figure> and <legend> is block. But whenever 
> the <figure> is a <table> container, then a figure{display:table} seems 
> more appropriate, to ensure that the legend looks like a caption to the 
> table.

In general we want to keep the default styles as simple as possible. 
However, I agree that authors may wish to consider doing what you 
describe.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Friday, 5 June 2009 23:04:50 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:38 GMT