W3C home > Mailing lists > Public > public-html-a11y@w3.org > August 2011

Re: HTML5 Marking up content that repeats on related pages

From: Greg Lowney <gcl-0039@access-research.org>
Date: Mon, 01 Aug 2011 23:44:11 -0800
Message-ID: <4E37AACB.3000406@access-research.org>
To: public-html-a11y@w3.org
Hi! According to my notes from today's conference call I was instructed to add use cases and submit this as a bug marked a11y but not, for now, a11ytf. However, the Wiki merely says "Discuss", and I can't find any mention of it at all in the IRC transcript (where according to my notes it came between discussion of tag-cloud <http://www.w3.org/TR/2011/WD-html5-20110525/links.html#tag-cloud> and Interactive_Elements <http://www.w3.org/WAI/PF/HTML/wiki/Spec_Review/Interactive_Elements>). Does anyone else have notes or memories that would help clear it up?

In this draft I've four possible technical approaches to this, the suggestion that the aside element should take the for attribute, and two use cases:

Title: Identifying repeated and non-repeated content

Recommendation: HTML5 should provide a mechanism by which site authors can identify blocks of content that repeat across pages, and portions within repeated blocks that are customized for a particular page.

One method would be to allow elements inside nav, header, or footer blocks to be marked with an attribute that identified them as page-specific, while another method would be to enclose those same elements in an element (like aside) that identifies them as being related to other content (like the main article). A third would be that rather than assume nav, header and footer are repeating across pages, to allow any element to be marked with an attribute that identified it as repeating, not repeating (page-specific), or unspecified (e.g. repeated=yes, =no, or blank meaning unspecified). A fourth would be to allow marking up an element with an arbitrary name identifying a group of pages on which the element repeated (e.g. pages at sample.com/software have <header group="software">, while pages at sample.com/hardware have <header group="hardware">).

Issue: Shouldn't the aside element accept the for attribute (as does the label element) so that it can identify itself as related to something on the page (e.g. the main article) even if it is located elsewhere (e.g. in a side panel)? The spec currently says that aside is always related to the content that contains the aside element, and it's not clear whether or how the author could then position it in a side panel or navigation area.

Use case: Ellen is using a screen reader and as she visits successive pages on a site, she does not want to listen to repeated blocks of content such as the header, footer, and bar. The simple solution is to ask her screen reader to skip over or her web browser to hide elements with the header, footer, and nav roles. However, she frequently visits a site on which these blocks, while generally repeating, contain information that is customized for each page. For example, each page also contain a side bar with several sections that repeat on each page (Calendar, Top Stories, Users Online) but also some sections that are specific to the page (a tag cloud and set of "Related" links). Similarly,  the  the navigation bar always shows links to the high-level segments of the site, but it also shows links to the sub-segments under the current segment; thus, any of the Samples pages would have links for Home, Biography, Bibliography, and Samples with subsidiary links for Sample 1, 
Sample 2, and Sample 3, while any of the Bibliography pages would have links for Home, Biography, Bibliography with subsidiary links for Non-Fiction, Fiction, and Short Works, and finally Samples. Ideally these non-repeated sections of the navigation and side bars would be marked as non-repeating, so that the user could have only those segments read to her.

Use case: Ronald uses a mobile device and visits the same site. Because space is precious, and he doesn't want to have to scroll through many, many lines of repeated content, an add-in for his browser provides a mode wherein all repeated content is hidden, including everything in the navigation bar and side bar except those portions marked as page-specific. This greatly reduces the number of inputs he has to use to navigate the page, as well as reducing bandwidth. When he decides he wants to use the repeated navigation links, he can chooses one of the mostly hidden regions and have it shown by itself, again keeping the amount of screen space and navigation commands to a minimum.


-------- Original Message  --------
Subject: HTML5 Marking up content that repeats on related pages
From: Greg Lowney <gcl-0039@access-research.org>
To: WAI-UA list <w3c-wai-ua@w3.org>, Janina Sajka <janina@rednote.net>
Date: 7/28/2011 2:15 AM
> Hi! The following was added to the UAWG wiki at http://www.w3.org/WAI/UA/work/wiki/HTML5_review_by_UAWG_notes#Marking_up_content_that_repeats_on_related_pages:
>     Marking up content that repeats on related pages
> Added by Greg Lowney <http://www.w3.org/WAI/UA/work/wiki/index.php?title=User:Glowney&action=edit&redlink=1> 09:11, 28 July 2011 (UTC)
> Can you mark up content that is repeated across pages on the site or subsite? For example, an example given in the HTML5 spec for the <footer> element describes it as showing a "site wide footer", but there is no easy way to distinguish that from a page-specific footer. This is particularly, but not exclusively, an issue for <header>, <footer>, and <nav> elements. This is an issue because assistive technology may often want to hide or skip over things that are repeated on multiple pages, but not skip over equivalents that are unique to the current page. Note that repeated content may still vary somewhat from page to page, as in when items in <nav> are all links except the one representing the current page, or the fact that different pages may have different copyright dates. One mechanism might be to allow an string attribute on these elements that could be compared to elements on other pages of the same site, and if the strings match the user agent can assume that they are 
> equivalent (e.g. <nav sitewide="toplevel">).
Received on Tuesday, 2 August 2011 06:46:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 04:42:43 GMT