E-mail List Archives
Thread: Header tag
Number of posts in this thread: 10 (In chronological order)
From: Julian Orr
Date: Tue, Apr 29 2003 5:30PM
Subject: Header tag
No previous message | Next message →
I watched the video of a a screen reader user and noticed that they were
prone to use the "header tag query" action in JAWS. I was familiar with
the frame list and link list in JAWs, but not this. I am wondering if
there are any JAWS users who can tell me if, when and how they use the list
"headers" to use web pages and web applications.
My reason is that in our web app
From: James Gagnier
Date: Tue, Apr 29 2003 5:53PM
Subject: Re: Header tag
← Previous message | Next message →
MessageHi Julian:
Often the header tag is used by Jaws users to quickly access the pages headers. If the headers are coded appropriately, IE: h1 page header, h2 sub header h3 a sub header of h2 and so on, a user can quickly access the header tags to get a feel for the page and quickly access the section which is of interest.
Example:
h1 page title
h2 chapter 1
h3 section 1
h3 section 2
h2 chapter 2
h3 section 1
h4 sub-section 1
etc.
Headers offer logical structure in documents that can actually be perceived by users who access the page not only through sight but through sound as well.
James
----- Original Message -----
From: Julian Orr
To: ' = EMAIL ADDRESS REMOVED = '
Sent: Tuesday, April 29, 2003 8:24 PM
Subject: Header tag
I watched the video of a a screen reader user and noticed that they were prone to use the "header tag query" action in JAWS. I was familiar with the frame list and link list in JAWs, but not this. I am wondering if there are any JAWS users who can tell me if, when and how they use the list "headers" to use web pages and web applications.
My reason is that in our web application the page alway
From: Alastair Campbell
Date: Wed, Apr 30 2003 6:12AM
Subject: RE: Header tag
← Previous message | Next message →
On a related note, does anyone know why automated checkers such as Bobby
fail a page when headers aren't in strict assending order?
I understand why heading tags should start with H1, followed by H2 then
H3 for page title, section and subsection. What I don't understand is
why you cannot start a new sub section, going from an subsection to a
new section (e.g. H4 to H2).
James' example was:
h1 page title
h2 chapter 1
h3 section 1
h3 section 2
h2 chapter 2
h3 section 1
h4 sub-section 1
Bobby would fail this for the H2 following an H3. Is there a reason for
this, or is it a slight bug in the check?
Thanks,
-Alastair Campbell
James Gagnier wrote:
> Often the header tag is used by Jaws users to quickly access the pages
> headers. If the headers are coded appropriately, IE: h1 page
> header, h2
> sub header h3 a sub header of h2 and so on, a user can quickly access
> the header tags to get a feel for the page and quickly access the
> section which is of interest.
----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/
From: Simon Jessey
Date: Wed, Apr 30 2003 1:01PM
Subject: Re: Header tag
← Previous message | Next message →
----- Original Message -----
From: "Alastair Campbell" < = EMAIL ADDRESS REMOVED = >
Subject: RE: Header tag
> I understand why heading tags should start with H1, followed by H2 then
> H3 for page title, section and subsection. What I don't understand is
> why you cannot start a new sub section, going from an subsection to a
> new section (e.g. H4 to H2).
>
> James' example was:
> h1 page title
> h2 chapter 1
> h3 section 1
> h3 section 2
> h2 chapter 2
> h3 section 1
> h4 sub-section 1
>
> Bobby would fail this for the H2 following an H3. Is there a reason for
> this, or is it a slight bug in the check?
There is no failure here, that I can see. There should never be a reason to
jump from a lower priority header to a higher priority header. In the
example given, chapters would normally be given unique documents, so the
hierarchy would be like this:-
<h1>Page Title</h1> (mimics the <title>...</title> element)
<h2>Chapter</h2>
<h3>Section 1</h3> (if required)
<h3>Section 2</h3>
If 'sub-sections' are required, then the documents should be further
sub-divided into sections, with each section having its own page instead.
This may seem like overkill, but it is the best method of maintaining a
proper, hierarchical structure.
Obviously there are exceptions to this. An example would be a W3C technical
recommendation. These are made available as a single, enormous HTML file for
those who prefer it, but such a file jumps up and down the header hierarchy
because the original, multi-page document (which did not do this) was
combined to make the large file.
Simon Jessey
w: http://jessey.net/blog/
e: = EMAIL ADDRESS REMOVED =
----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/
From: John Foliot - bytown internet
Date: Wed, Apr 30 2003 1:27PM
Subject: RE: Header tag
← Previous message | Next message →
Simon Jessey wrote:
> There is no failure here, that I can see. There should never be a
> reason to
> jump from a lower priority header to a higher priority header.
I'm sorry Simon, but I beg to differ...
Look at it this way:
<h1>Main Title</h1>
<h2>Section 1 - Here</h2>
<h3>This</h3>
<h3>That</h3>
<h3>The Other</h3>
<h2>Section 2 - There</h2>
<h3>This</h3>
<h3>That</h3>
<h3>The Other</h3>
<h4>An Alternative Other</h4>
<h2>Section 3 - Anywhere</h2>
<h3>This</h3>
<h3>That</h3>
<h3>The Other</h3>
Is perfectly valid, legitimate and semantically correct. If a validation
tool "fails" this, the tool is flawed.
(That this much content could "probably/possibly" be spread over multiple
pages is not the point here, although it would also be a recommendation.
However, structurally this is most correct.)
JF
>
From: Simon Jessey
Date: Wed, Apr 30 2003 3:27PM
Subject: Re: Header tag
← Previous message | Next message →
----- Original Message -----
From: "John Foliot - bytown internet" < = EMAIL ADDRESS REMOVED = >
Subject: RE: Header tag
> Look at it this way:
>
> <h1>Main Title</h1>
>
> <h2>Section 1 - Here</h2>
> <h3>This</h3>
> <h3>That</h3>
> <h3>The Other</h3>
>
> <h2>Section 2 - There</h2>
> <h3>This</h3>
> <h3>That</h3>
> <h3>The Other</h3>
> <h4>An Alternative Other</h4>
>
> <h2>Section 3 - Anywhere</h2>
> <h3>This</h3>
> <h3>That</h3>
> <h3>The Other</h3>
>
> Is perfectly valid, legitimate and semantically correct. If a validation
> tool "fails" this, the tool is flawed.
>
Hmmm. That looks like it should be marked up as a list, not a series of
cascading headings. An ordered list with unordered lists in each list item
would do nicely, and would be still more semantically correct, would it not?
I can see where you are going with it, but I still think that moving up and
down heading levels feels wrong somehow. XHTML2.0 partially corrects this
problem by introducing a <section>...</section> element that partly governs
the seniority of headings. Unfortunately, I think XHTML 2.0 in its current
form is a flawed proposal for a number of reasons.
----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/
From: Simon Jessey
Date: Wed, Apr 30 2003 6:17PM
Subject: Re: Header tag
← Previous message | Next message →
----- Original Message -----
From: "Lori K. Brown" < = EMAIL ADDRESS REMOVED = >
Subject: Re: Header tag
> My assumption, reading J. Foliot's lucid example, is that there is STUFF,
> that is to say, paragraphs and things, following each of the headings. So,
> no, Simon, it's not a list.
I think perhaps you misunderstand what I meant by a list. To see an example
of how list markup can be used, see my weblog (in my signature). Each entry
is a list item, and each item can have headings, paragraphs, tables, images
and other elements. Using list markup creates a logical, hierachical
structure (with semantic richness) for something that has many sections and
subsections. Such an approach is widely used by forward-thinking web
designers because of its good structure, accessibility and easy-to-style
nature.
If you think about it, a collection of 'sections', which in turn may have
collections of sub-sections, are nothing more than nested ordered or
unordered lists. Since list markup can contain pretty much anything, it
makes sense to use it.
In cases where such an approach cannot work, my _recommendation_ is to try
and break things up into shorter documents, to _avoid_ messing with the
structured nature of headings. I have created countless pages and I have
thus far not found the need to break that heading structure in the way this
thread has described.
<< Simon, you are bending over backwards to defend an illogical and
indefensible interpretation of the standards. Separate pages for each? That
is simply bananas. If they were trying to make me roll my eyes and give up
on validation, this goes a long way toward success. Any architecture that
requires us to break things up into arbitrary tiny pieces is just crazy
talk. >>
It is unfortunate that you have chosen to interpret my suggestions in this
way. I assure you, my intention was merely to contribute to an interesting
discussion, and to give my thoughts on the perceived fault in the Bobby
tool. I am a regular user of Bobby because I like to do my best to ensure my
web pages are as accessible as possible. I consider Bobby 'approval' to be
as necessary as HTML validation, so it is important to me that the client
works correctly:-
http://bobby.watchfire.com/bobby/bobbyServlet?URL=http%3A%2F%2Fjessey.net%2Fblog%2F&;output=Submit&gl=sec508&test
Simon Jessey
w: http://jessey.net/blog/
e: = EMAIL ADDRESS REMOVED =
----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/
From: Lori K. Brown
Date: Wed, Apr 30 2003 9:32PM
Subject: Re: Header tag
← Previous message | Next message →
I know that lists can be semantically rich, I just find it completely
crazy that HTML, the semantic structure that was ORIGINALLY DESIGNED for
research paper type content, presumably headings and paragraphs, is now
being so overconstrained that its originators would get their hands
slapped for using it as they themselves originally intended. And I think
that list coding should be used for things that an ordinary person would
think make sense to be described as a list. And a long outline of text
headings and paragraphs fails this smell test, completely.
I don't understand what is gained by anyone with this strictness. It seems
pointless and off-putting, and runs counter to the common sense
inclusiveness that the web should aspire to.
LKB
> ----- Original Message -----
> From: "Lori K. Brown" < = EMAIL ADDRESS REMOVED = >
> Subject: Re: Header tag
>
>
>> My assumption, reading J. Foliot's lucid example, is that there is
>> STUFF,
>> that is to say, paragraphs and things, following each of the headings.
>> So,
>> no, Simon, it's not a list.
>
> I think perhaps you misunderstand what I meant by a list. To see an
> example
> of how list markup can be used, see my weblog (in my signature). Each
> entry
> is a list item, and each item can have headings, paragraphs, tables,
> images
> and other elements. Using list markup creates a logical, hierachical
> structure (with semantic richness) for something that has many sections
> and
> subsections. Such an approach is widely used by forward-thinking web
> designers because of its good structure, accessibility and easy-to-style
> nature.
>
> If you think about it, a collection of 'sections', which in turn may have
> collections of sub-sections, are nothing more than nested ordered or
> unordered lists. Since list markup can contain pretty much anything, it
> makes sense to use it.
>
> In cases where such an approach cannot work, my _recommendation_ is to try
> and break things up into shorter documents, to _avoid_ messing with the
> structured nature of headings. I have created countless pages and I have
> thus far not found the need to break that heading structure in the way
> this
> thread has described.
>
>
> << Simon, you are bending over backwards to defend an illogical and
> indefensible interpretation of the standards. Separate pages for each?
> That
> is simply bananas. If they were trying to make me roll my eyes and give up
> on validation, this goes a long way toward success. Any architecture that
> requires us to break things up into arbitrary tiny pieces is just crazy
> talk. >>
>
> It is unfortunate that you have chosen to interpret my suggestions in this
> way. I assure you, my intention was merely to contribute to an interesting
> discussion, and to give my thoughts on the perceived fault in the Bobby
> tool. I am a regular user of Bobby because I like to do my best to ensure
> my
> web pages are as accessible as possible. I consider Bobby 'approval' to be
> as necessary as HTML validation, so it is important to me that the client
> works correctly:-
>
> http://bobby.watchfire.com/bobby/bobbyServlet?URL=http%3A%2F%2Fjessey.net%2Fblog%2F&;output=Submit&gl=sec508&test>
>
> Simon Jessey
>
> w: http://jessey.net/blog/
> e: = EMAIL ADDRESS REMOVED =
>
>
> ----
> To subscribe, unsubscribe, or view list archives,
> visit http://www.webaim.org/discussion/
>
--
Lori K. Brown
User Interface Engineer
SiteScape, Inc.
----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/
From: Alastair Campbell
Date: Thu, May 01 2003 7:16AM
Subject: RE: Header tag
← Previous message | Next message →
Thanks for the comments everyone, I hadn't realised this was such a
controversial issue! I'm afraid this is a fairly long post, since I'm
replying to several people.
Simon Wrote:
> If 'sub-sections' are required, then the documents should be further
> sub-divided into sections, with each section having its own
> page instead.
I appreciate Simon's recommendation about chunking up pages, but I
believe there are times when a (fairly) long document is necessary &
desirable, and appropriate mark-up should help people navigate that
document.
The type of case I had in mind was a set of (racing) rules, which in use
will often be printed out, but should be a good (XHTML) web document to.
I'm a little sceptical about marking up large portions of the page as a
list. What is the point of headings and paragraphs then? Whether or not
it closely follows a book metaphor, it should provide a mechanism for
valid nesting. I can see Simon's point about lists when the context is
blog entries, but not for a document with sections and sub-sections.
The WAI isn't particularly clear what the rule on nested headings is, so
I couldn't blame Bobby's makers for not being 100% on it.
"Use header elements to convey document structure and use them according
to specification." - does anyone know where to find the details on the
nesting of headings?
HTML 4 doesn't seem to:
http://www.w3.org/TR/html401/struct/global.html#h-7.5.5
Perhaps it is just a matter of the logic applied in the Bobby test?
Bobby seems to check that each heading is the same number or one higher
than the previous one.
I would suggest that is should be:
Each heading should only be allowed to increment by 0 or 1 compared to
the previous heading, but is allowed to decrement by any number.
I hope that made sense, I'm not a mathematician or programmer!
Does anyone use heading browsing, for example in Jaws? Does it make it
more confusing to have nested headings?
I guess that it would make *more* semantic sense if these things were
actually nested, as the section tag in XHTML 2 implies, or as Patrick
suggested with div tags. However, the headings are there for current
user agents to use in many documents on the web.
Last point to Simon: Although I also use Bobby regularly, some of the
checks are not as useful as they were. For example, your page would get
a 'triple a' rating apart from two things, one of them being "Separate
adjacent links with more than whitespace."
I remember reading that no currently used user-agent has a problem with
links not separated by non-link characters. Yet we still have to deal
with this to pass a test that the general public uses (without an
advanced level of accessibility knowledge).
I could be wrong about the whitespace checkpoint (I can't find the
reference at the moment) but that and the nested headings checkpoint
seem to cause more hassle than benefit.
-Alastair
----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/
From: John Foliot - bytown internet
Date: Thu, May 01 2003 8:13AM
Subject: RE: Header tag
← Previous message | No next message
Simon,
IMHO, you are partially right but mostly wrong here.
Yes, a structure of document headings can be interpreted as a list (similar
to what I originally illustrated)... think Table of Contents. But then, as
you proceed through the "book" you are presented with chapters <h1> which
may have subheadings <h2> and even tertiary headings <h3> under some of the
sub headings <h2> within any given chapter. But then, boom, you are into
the next chapter <h1>, and a similar (or different) semantic structure takes
place for that chapter. Chapters, when presented in a Table of Contents,
may be seen as a list, but when they are then presented as "the work", are
they still list items, or do they take on a different persona, a different
semantic "occupation"?
Lists, are intended to be brief, point form style information. It doesn't
have to be just a couple of words, it can be a paragraph or two, but
generally they are short, complete thoughts; and multiple ones at that,
ordered out in either a random order (<UL>) or chronological order (<OL>),
or occasionally as a "Question and Answer, Call and Response" style listing
(<DL>)
Heading tags on the other hand can contain 10 - 20 or more paragraphs before
the next logical header appears, and it can either be a subset of the
current idea (<h3> following an <h2>), equal to the previous idea (<h2>
followed by another <h2>), or followed by a whole new concept (transition
back to <h1> for example). Each of these functions offer semantic
information beyond the fact that they are groupings of ideas, which
unfortunately is all that list mark-up can do.
Simon, "information" within a web-site can assume different roles, depending
on how and when it's being conveyed. I invite you to visit
www.dougthompson.ca On this site, each week the Councillor writes a weekly
newsletter to keep his local constituents informed of what's going on.
Pretty basic stuff. On the main page, there is a list of the three most
recent articles, actually coded as a definition list, with the "headline"
(<dt>) and then the first paragraph(<dl>) of that week's column. In the
context of what is being presented here, it *is* a list of the last three
columns. "Clicking" onto the headline however takes you to the week in
question, and there, while the "text" or "information" remains the same, the
semantic structure of that document and it's contents has changed; there are
now a series of <h3> headlines grouped under two larger <h2> headings.
Perfectly logical and, semantically correct... these aren't really list
items now are they?
Now for an added bonus... while I always extol working towards software
agnostic development, the most common screen reader in the marketplace
(JAWs) takes advantage of the <h> tags whenever a page is loaded. When a
non-sited user loads a web page into their browser, JAWs tells them 3 things
before even starting to read the page: The <title> of the document, how
many links there are on the page, and how many headers there are. Should a
document load up with numerous headers, the JAWs user has the ability to
bring up a list of all of the Headers in the document, and effectively
"scan" them to see if the information they are seeking is present, (perhaps
as a subheading in Chapter 3?). Should the information be present (or
appear close enough) the Jaws user can select that header, and the software
treats the <h> tag almost like a named anchor, taking the "reading focus" to
that point in the document, allowing them to skip over all the other stuff.
Lists alone do not provide that functionality.
And so fianlly, back to the original question. If a validation tool fails a
listing as I originaly illustrated, the tool is flawed:
> <h1>Main Title</h1>
>
> <h2>Section 1 - Here</h2>
> <h3>This</h3>
> <h3>That</h3>
> <h3>The Other</h3>
>
> <h2>Section 2 - There</h2>
> <h3>This</h3>
> <h3>That</h3>
> <h3>The Other</h3>
> <h4>An Alternative Other</h4>
>
> <h2>Section 3 - Anywhere</h2>
> <h3>This</h3>
> <h3>That</h3>
> <h3>The Other</h3>
JF
*********************
>