This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
"4.4.3 The section element" http://www.w3.org/html/wg/drafts/html/master/sections.html#the-section-element
Commit pushed to master at https://github.com/w3c/html https://github.com/w3c/html/commit/423ee2376ce2e0ab6ee7d9c9631cf07c77da57de authors should provide a <hx> for a section fixes bug 23490
refer to discussion Validity constraints on <section>: http://lists.w3.org/Archives/Public/public-html/2013Mar/0129.html EDITOR'S RESPONSE: This is an Editor's Response to your comment. If you are satisfied with this response, please change the state of this bug to CLOSED. If you have additional information and would like the Editor to reconsider, please reopen this bug. If you would like to escalate the issue to the full HTML Working Group, please add the TrackerRequest keyword to this bug, and suggest title and text for the Tracker Issue; or you may create a Tracker Issue yourself, if you are able to do so. For more details, see this document: http://dev.w3.org/html5/decision-policy/decision-policy.html Status: Accepted Change Description: see commit Rationale: Authors are using section as a generic container, trying to reduce this by defining that they SHOULD provide a heading for each section.
This is a wrong fix to a non-problem. It has no practical impact except that it may confuse authors. One of the few reasons to use section elements is that they provide a way to divide a document into thematic parts even when some or no parts has a heading. Moreover, the characterization “briefly describes the content of the section” reflects a very one-sided view on the roles of a heading. For example, a novel can be divided into section elements, and they can have headings, often just numbering headings like “Chapter 1”. There is no reason to tell people not to use such headings, especially if they are just converting an existing work into HTML format. And a chapter of a novel could meaningfully be divided into sections without headings – parts that describe different courses of events, visually separated e.g. by a blank line. Note that the descriptions of h1–h6 do not say that headings describe the content of sections. They are just, well, headings. This is OK. There are many kinds of headings, and their nature is a matter of presentation style and depends on the genre – it’s not adequate to restrict headings to brief descriptions of content.
(In reply to Jukka K. Korpela from comment #3) > This is a wrong fix to a non-problem. It has no practical impact except that > it may confuse authors. One of the few reasons to use section elements is > that they provide a way to divide a document into thematic parts even when > some or no parts has a heading. Moreover, the characterization “briefly > describes the content of the section” reflects a very one-sided view on the > roles of a heading. > > For example, a novel can be divided into section elements, and they can have > headings, often just numbering headings like “Chapter 1”. There is no reason > to tell people not to use such headings, especially if they are just > converting an existing work into HTML format. And a chapter of a novel could > meaningfully be divided into sections without headings – parts that describe > different courses of events, visually separated e.g. by a blank line. > > Note that the descriptions of h1–h6 do not say that headings describe the > content of sections. They are just, well, headings. This is OK. There are > many kinds of headings, and their nature is a matter of presentation style > and depends on the genre – it’s not adequate to restrict headings to brief > descriptions of content. HI jukka, while i understand your feedback in regards to prescribing what headings are i don't understand "> This is a wrong fix to a non-problem. " why is it a non problem? And why is it wrong to make a soft requirement for the author to provide a heading?
(In reply to steve faulkner from comment #4) > why is it a non problem? Because a section element without a heading is not a problem. It is often quite adequate, as I described. > And why is it wrong to make a soft requirement for > the author to provide a heading? To people accustomed to reading RFCs, as many readers of the spec are, “should” is a serious word – and the spec defines “should” to mean what RFC 2119 says. I don’t think HTML5 should give advice on the composition of works in general. That would be way out of scope (and would often result in wrong advice). If you want to say something about headings, then this could be suitable: The content of a section element often starts with a heading or otherwise contains a heading for the section. A section may also appear without a heading
(In reply to Jukka K. Korpela from comment #5) > (In reply to steve faulkner from comment #4) > > > why is it a non problem? > > Because a section element without a heading is not a problem. It is often > quite adequate, as I described. you didm't actually describe it. > > > And why is it wrong to make a soft requirement for > > the author to provide a heading? > > To people accustomed to reading RFCs, as many readers of the spec are, I would suggest the opposite > “should” is a serious word – and the spec defines “should” to mean what RFC > 2119 says. and that is why it was used. 3. SHOULD This word, or the adjective "RECOMMENDED", mean that there may exist valid reasons in particular circumstances to ignore a particular item, but the full implications must be understood and carefully weighed before choosing a different course. http://www.ietf.org/rfc/rfc2119.txt is reasonable as the provision of a heading is recommended in most circumstances. The lack of clarity around the use of section has already resulted in widespread misuse which has had a negative effect on users.
(In reply to steve faulkner from comment #6) > > Because a section element without a heading is not a problem. It is often > > quite adequate, as I described. > > you didm't actually describe it. I gave a an example of a novel section divided into subsections. > > “should” is a serious word – and the spec defines “should” to mean what RFC > > 2119 says. > > and that is why it was used. Taken in the RFC 2119 sense, as it should, it is in not “soft”. As you quote: > the full implications must be understood and > carefully weighed before choosing a different course. How is an author even assumed to weight the implications when no rationale for the requirement is given? > is reasonable as the provision of a heading is recommended in most > circumstances. Recommended by whom, and why? It is evident that a heading is useful in many cases. So evident that it hardly makes sense to say it in normative prose (even as a “should” requirement) in a specification. In other cases, the requirement would be confusing at best, and could even make people write dummy heading content if they take the requirement seriously. > The lack of clarity around the use of section has already > resulted in widespread misuse which has had a negative effect on users. Which widespread misuse with which negative effect on users? The section element has no impact on users, really. And if some content should have a heading, then it should have a heading quite independently of use of a section element – so the context would be wrong for advocacy of headings even if we thought that such advocacy belongs to HTML5.
(In reply to Jukka K. Korpela from comment #7) > (In reply to steve faulkner from comment #6) > > > > Because a section element without a heading is not a problem. It is often > > > quite adequate, as I described. > > > > you didm't actually describe it. > > I gave a an example of a novel section divided into subsections. > > > > “should” is a serious word – and the spec defines “should” to mean what RFC > > > 2119 says. > > > > and that is why it was used. > > Taken in the RFC 2119 sense, as it should, it is in not “soft”. As you quote: it is soft compared to MUST > > > the full implications must be understood and > > carefully weighed before choosing a different course. > > How is an author even assumed to weight the implications when no rationale > for the requirement is given? that is an argument for better explanation in the spec - noted > > > is reasonable as the provision of a heading is recommended in most > > circumstances. > > Recommended by whom, and why? It is evident that a heading is useful in many > cases. So evident that it hardly makes sense to say it in normative prose > (even as a “should” requirement) in a specification. In other cases, the > requirement would be confusing at best, and could even make people write > dummy heading content if they take the requirement seriously. recommended in the spec and no it would mean that authors don't use the section element as a replacement for a div which they currently are doing. > > > The lack of clarity around the use of section has already > > resulted in widespread misuse which has had a negative effect on users. > > Which widespread misuse with which negative effect on users? > > The section element has no impact on users, really. And if some content > should have a heading, then it should have a heading quite independently of > use of a section element – so the context would be wrong for advocacy of > headings even if we thought that such advocacy belongs to HTML5. The section element maps to a region role in accessibility APIs (as required in HTML5), region roles are exposed to users. and its misuse does have a negative effect as reported by users. sections also have an effect upon the document outline and their misuse could have a negative effect upon users if any user agent implements the outline algorithm.
Although I appreciate that a "should" may have a positive impact on sectionarrhea, I confirm Jukka's stand that heading-less sections are heavily used in digital publishing (e.g. EPUB) – as I already commented last March on twitter [1], reported by Robin [2]. Would it make sense to reduce the confusion by explicitly mentioning legitimate use cases for heading-less sections, either via prose or examples ? Additionally, is it still a no-go to recommend aria-label as an alternative to headings in sections ? [1] https://twitter.com/rdeltour/status/314739301676113920 [2] http://lists.w3.org/Archives/Public/public-html/2013Mar/0143.html
(In reply to Romain Deltour from comment #9) hi Romain, > Would it make sense to reduce the confusion by explicitly mentioning > legitimate use cases for heading-less sections, either via prose or examples > ? yes that is a possibility > Additionally, is it still a no-go to recommend aria-label as an alternative > to headings in sections ? it isn't a no go note the change I made in the spec is a starting place not an end, its only an editors draft after all.
> it isn't a no go Ok, I thought it was based on that comment [1], but fine then! [1] http://lists.w3.org/Archives/Public/public-html/2013Mar/0144.html
re-opening to reflect ongoing discussion
(In reply to Romain Deltour from comment #9) > Although I appreciate that a "should" may have a positive impact on > sectionarrhea, I confirm Jukka's stand that heading-less sections are > heavily used in digital publishing (e.g. EPUB) – as I already commented last > March on twitter [1], reported by Robin [2]. > > Would it make sense to reduce the confusion by explicitly mentioning > legitimate use cases for heading-less sections, either via prose or examples > ? > > Additionally, is it still a no-go to recommend aria-label as an alternative > to headings in sections ? > > [1] https://twitter.com/rdeltour/status/314739301676113920 > [2] http://lists.w3.org/Archives/Public/public-html/2013Mar/0143.html Hi Romain, could you point to some examples of epub docs that make use of section? and related advice in epub on its use? It would be useful to get a better understanding of how it is used.
(In reply to Romain Deltour from comment #9) > Although I appreciate that a "should" may have a positive impact on > sectionarrhea, I confirm Jukka's stand that heading-less sections are > heavily used in digital publishing (e.g. EPUB) – as I already commented last > March on twitter [1], reported by Robin [2]. > > Would it make sense to reduce the confusion by explicitly mentioning > legitimate use cases for heading-less sections, either via prose or examples > ? > > Additionally, is it still a no-go to recommend aria-label as an alternative > to headings in sections ? > > [1] https://twitter.com/rdeltour/status/314739301676113920 > [2] http://lists.w3.org/Archives/Public/public-html/2013Mar/0143.html the ePub accessibility guide contains good advice on section http://www.idpf.org/accessibility/guidelines/content/about.php, lets look at how we can incorporate a reference to it to clarify requirements around use of hx
Commit pushed to master at https://github.com/w3c/html https://github.com/w3c/html/commit/91960c6f0eeb4386cb0a1c4b83c7feda16b6783e tweaked section definition see discussion on bug 23490
(In reply to github bugzilla bot from comment #15) > Commit pushed to master at https://github.com/w3c/html > > https://github.com/w3c/html/commit/91960c6f0eeb4386cb0a1c4b83c7feda16b6783e > tweaked section definition > > see discussion on bug 23490 I have tweaked it so the SHOULD refers to the identification of the theme of the section rather than should be a heading, this brings it into line with the ePub accessibility recommendations on sections http://www.idpf.org/accessibility/guidelines/content/xhtml/sections.php
(In reply to Jukka K. Korpela from comment #3) > This is a wrong fix to a non-problem. It has no practical impact except that > it may confuse authors. One of the few reasons to use section elements is > that they provide a way to divide a document into thematic parts even when > some or no parts has a heading. Moreover, the characterization “briefly > describes the content of the section” reflects a very one-sided view on the > roles of a heading. > > For example, a novel can be divided into section elements, and they can have > headings, often just numbering headings like “Chapter 1”. There is no reason > to tell people not to use such headings, especially if they are just > converting an existing work into HTML format. And a chapter of a novel could > meaningfully be divided into sections without headings – parts that describe > different courses of events, visually separated e.g. by a blank line. > > Note that the descriptions of h1–h6 do not say that headings describe the > content of sections. They are just, well, headings. This is OK. There are > many kinds of headings, and their nature is a matter of presentation style > and depends on the genre – it’s not adequate to restrict headings to brief > descriptions of content. Jukka, I have removed the offending words and have reworked the definition
Commit pushed to master at https://github.com/w3c/html https://github.com/w3c/html/commit/55c67c05f21eed81d985f7a26e9ea3b46fe915a5 changed child to descendant see bug 23490
Probably nitpicking, but if a section has a sub-section with a heading, it complies with the "descendant heading" recommendation but the top-level section is still "unidentified". Also, I w/b in favor of explicitly mentioning aria-label as an alternative way to identify the "theme" of the section.
(In reply to Romain Deltour from comment #19) > Probably nitpicking, but if a section has a sub-section with a heading, it > complies with the "descendant heading" recommendation but the top-level > section is still "unidentified". > > Also, I w/b in favor of explicitly mentioning aria-label as an alternative > way to identify the "theme" of the section. thanks Romain, not nit picking it needs to be further clarified. I will also add a note and example about use of aria-label.
(In reply to steve faulkner from comment #20) > (In reply to Romain Deltour from comment #19) > > Probably nitpicking, but if a section has a sub-section with a heading, it > > complies with the "descendant heading" recommendation but the top-level > > section is still "unidentified". > > > > Also, I w/b in favor of explicitly mentioning aria-label as an alternative > > way to identify the "theme" of the section. > > thanks Romain, not nit picking it needs to be further clarified. I will also > add a note and example about use of aria-label. So, if I understand it correctly, we agree on the fact that not always the use of a heading is appropriate for a section. However, we only must use a section when we want it to be a part of the HTML outline, even is the section is unnamed. In this bug (https://www.w3.org/Bugs/Public/show_bug.cgi?id=23545) I find a discussion about screenreaders that should skip a section if it does not have an accessible name. So, my conclusion is, that when I use a section with no heading, for example for grouping a bunch of logo's, which I still want to be part of the HTML outline, I can use ARIA-label to provide this section with an accessible name so screenreaders won't skip this.
(In reply to Willem-Siebe Spoelstra from comment #21) > However, we only must use a section when we want it to be a part of the HTML > outline, even is the section is unnamed. correct > > In this bug (https://www.w3.org/Bugs/Public/show_bug.cgi?id=23545) I find a > discussion about screenreaders that should skip a section if it does not > have an accessible name. correct, I have added advice to the spec on that: "Note:It is strongly recommended that user agents such as screen readers only convey the presence of, and provide navigation for section elements, when the section element has an accessible name." http://www.w3.org/html/wg/drafts/html/master/dom.html#sec-implicit-aria-semantics > > So, my conclusion is, that when I use a section with no heading, for example > for grouping a bunch of logo's, which I still want to be part of the HTML > outline, I can use ARIA-label to provide this section with an accessible > name so screenreaders won't skip this. yes, that is the intended outcome.
(In reply to Willem-Siebe Spoelstra from comment #21) note the current behaviour of announcing a region is onl present in one screen reader (JAWS - the most popular one) also note that the outline algorithm is not supported by any browser or AT, also see http://blog.paciellogroup.com/2013/10/html5-document-outline/ in regards to outline algorithm support.
Commits pushed to CR at https://github.com/w3c/html https://github.com/w3c/html/commit/423ee2376ce2e0ab6ee7d9c9631cf07c77da57de authors should provide a <hx> for a section https://github.com/w3c/html/commit/91960c6f0eeb4386cb0a1c4b83c7feda16b6783e tweaked section definition https://github.com/w3c/html/commit/55c67c05f21eed81d985f7a26e9ea3b46fe915a5 changed child to descendant