W3C home > Mailing lists > Public > public-html@w3.org > September 2007

Re: [html] Semantics of "aside", "header", and "footer"

From: Robert Burns <rob@robburns.com>
Date: Mon, 10 Sep 2007 10:50:27 -0500
Message-Id: <CAC9A2E1-B53F-452B-856D-8960A83508BF@robburns.com>
Cc: public-html@w3.org
To: Jens Meiert <jens.meiert@erde3.com>

Hi jens and Karl,

On Sep 10, 2007, at 6:16 AM, Jens Meiert wrote:

>
>> "header", "footer", "aside" seem to be bad names for "banner",
>> "masthead" and "note". But maybe people understand them better as
>> well.
>
> Right, but I guess there are quite a few elements then that could  
> use new names. It also doesn't solve the problem that these names  
> imply a certain kind of presentation, and beside the general  
> problem with that, does it really work everywhere?
>
>> I guess the proposed names come from the class names found on the
>> Web.
>
> I don't really consider this to be a good argument. Even if you  
> don't think about other things people do that are not necessarily  
> “beneficial”, Ian's stats also show popular class names like  
> “small”, “text”, or even “white” [1]. Like nobody seriously  
> considered to introduce a “white” element, there probably shouldn't  
> be something like “header” or “footer”.
>
> As far as I also understand recent feedback on that matter, we  
> certainly don't make a mistake by reconsidering the names and  
> probably even modify the semantics of “header”, “footer”, and  
> “aside”. (Probably flanked by a quick survey?)
>
>
> [1] http://code.google.com/webstats/2005-12/classes.html

To me the issue is not merely that these elements may be  
presentational. Certainly the use of a class name of footer and  
header that I've seen are meant to imply the element is at the bottom  
or top of the page respectively. I like Karl's suggestion of  
'masthead' since I think makes the association even clearer.

The more severe problem is then to try to take these presentational  
concepts and turn them into semantic uses. If we're going to use  
statistics to guide our choice of new element tag names, it shouldn't  
be in name only. Those class names are not just class names that  
sound presentational, but are used for non-presentational purposes.  
They are presentational. So we shouldn't just adapt the  
presentational class names to be non-presentational tag names.

This to me is just as misguided as taking something like the 'small'  
element (or B or I ) and trying to turn it into a non-presentational  
element. For example, the current draft cautions authors that the  
styling of B, I and SMALL could be changed at any time, rather than  
cautioning authors that these should not be restyled. This, to me, is  
completely backwards. That would be like cautioning authors that  
though a P element should be used to markup a paragraph, authors and  
users should be aware that it could instead be used as a quotation at  
any time or even a line-break.  Granted it could be used as a  
quotation as a misuse, but saying so just encourages that misuse.  
Similarly, a SMALL element might be mist-styled as larger than the  
surrounding text, but that would be a misuse of the purpose of the  
SMALL element.

Similarly, we have a proposed 'header' element that is drawn from  
statistical results showing many authors use <div class='header''>  
(and probably <div id='header'> too). In my experience this is used  
to markup a section of a document for the top of a document. It is in  
that sense presentational —used as a masthead just like the name Karl  
suggests. I raise little objection to defining such semantics with a  
'masthead' or 'header' tag name. The problem to me arises when we try  
to repurpose that tag name to be non-presentational: to treat it as a  
heading element (though still called 'header'). Why not just use 'h1'  
or 'h' or 'heading' for headings and allow them to be nested for  
subheadings? Why deliberately introduce confusion for authors by  
calling the heading element a 'header'? Similar issues arise for  
'footer' Especially since the attempt is also made to make use of  
'footer' as a scoped element: scoped to sections. This further  
undermines its use as a presentational element for content at the  
bottom of the page. Sure it could be for material at the end of a  
section, but I don't think that really needs markup. The order of  
content in a section already indicates what portion comes at the end  
of the section. If there are particular metadata semantics we need to  
markup that traditionally get displayed in footers, then lets add  
those metadata semantics (like a 'copyright' element for instance).

The issues of footnotes and endnotes, raised by Karl, do not really  
fit with aside. I'm still working on my review and proposal for this,  
but I have started a wiki page to deal with subordinate text[1]. This  
subordinate text might be presented as a parenthetical or as endnotes  
or footnotes (or scrolling-frame notes, etc). This is a very  
different semantics from 'aside' such as the right or left sidebar of  
a web page.

Take care,
Rob

[1]: <http://esw.w3.org/topic/HTML/Subtext>
Received on Monday, 10 September 2007 15:50:50 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:07 GMT