By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
425,910 Members | 1,036 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 425,910 IT Pros & Developers. It's quick & easy.

HTML headings article

P: n/a
I wrote an article on HTML headings:
http://codewallop.110mb.com/goodprac...adingology.htm

--
Spartanicus
Apr 8 '07 #1
Share this Question
Share on Google+
22 Replies


P: n/a
Spartanicus <in*****@invalid.invalidwrote:
I wrote an article on HTML headings:
http://codewallop.110mb.com/goodprac...adingology.htm
Another argument in favor (or if you prefer, in favour :-) of skipping
heading levels is that sometimes, skipping heading levels better reflects
the document structure.

For example, you might have sections which use H2. Some of those sections
might have subsections which use H3, while others would have no
subsections. If the content within those sections and subsections needs to
be structured into "chunks", what heading levels should be used for those
"chunks"? The "never skip heading levels" argument would have some of the
"chunks" use H3, and some of them use H4. But it can make more structural
sense to use H4 for all the "chunks", even though this skips the H3 heading
level in sections which have no subsections.
--
Darin McGrew, mc****@stanfordalumni.org, http://www.rahul.net/mcgrew/
Web Design Group, da***@htmlhelp.com, http://www.HTMLHelp.com/

"When strong encryption is outlawed, only outlaws jvyy hfr fgebat rapelcgvba."
Apr 9 '07 #2

P: n/a
Scripsit Darin McGrew:
Spartanicus <in*****@invalid.invalidwrote:
>I wrote an article on HTML headings:
http://codewallop.110mb.com/goodprac...adingology.htm

Another argument in favor (or if you prefer, in favour :-) of skipping
heading levels is that sometimes, skipping heading levels better
reflects the document structure.
"Another"? It seems that I have to scan through the cited article to see its
argument for skipping heading levels in some cases. It seems to be the
age-old "but h1 looks so BIG!!" argument in slightly modernized clothes.

The article complains about lack of arguments in specifications that
recommend that heading levels be not skipped. To me, the basic argument has
always been so evident that it need not be mentioned: you must not skip
levels for the same reason why you must not skip numbers when counting 1, 2,
3, ... Heading levels indicate logical nesting of parts of a document, and
you cannot nest something at the 3rd level without having something at the
2nd level. But people apparently often view the heading levels as "levels of
importance", and some unfortunate wordings in specifications might
contribute to this misconception.
For example, you might have sections which use H2. Some of those
sections might have subsections which use H3, while others would have
no subsections.
That's possible, though it often results from insufficient structuring or
poor division into sections.
If the content within those sections and subsections
needs to be structured into "chunks", what heading levels should be
used for those "chunks"?
"Chunk" is not an HTML concept or a document structuring concept.
The "never skip heading levels" argument
would have some of the "chunks" use H3, and some of them use H4. But
it can make more structural sense to use H4 for all the "chunks",
even though this skips the H3 heading level in sections which have no
subsections.
You seem to postulate the existence of document structure "bottom up", e.g.
so that at the lowest level we have paragraphs, they may be grouped into a
larger blocks, etc. However, the HTML structuring idea is clearly "top
down". You can't really have it both ways.

The solution is to design the structure according to the "top down" division
into parts. If you can't do this and you need headings for the "chunks",
then you might decide to violate the recommendations in HTML specifications
(and the HTML _standard_, ISO HTML ! :-) ), but it doesn't help to explain
that this is a good thing to do.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 10 '07 #3

P: n/a
"Jukka K. Korpela" <jk******@cs.tut.fiwrote in message
news:oQ*******************@reader1.news.saunalahti .fi
Scripsit Darin McGrew:
>Spartanicus <in*****@invalid.invalidwrote:
>>I wrote an article on HTML headings:
http://codewallop.110mb.com/goodprac...adingology.htm

Another argument in favor (or if you prefer, in favour :-) of
skipping heading levels is that sometimes, skipping heading levels
better reflects the document structure.

"Another"? It seems that I have to scan through the cited article to
see its argument for skipping heading levels in some cases. It seems
to be the age-old "but h1 looks so BIG!!" argument in slightly
modernized clothes.
The article complains about lack of arguments in specifications that
recommend that heading levels be not skipped. To me, the basic
argument has always been so evident that it need not be mentioned:
you must not skip levels for the same reason why you must not skip
numbers when counting 1, 2, 3, ... Heading levels indicate logical
nesting of parts of a document, and you cannot nest something at the
3rd level without having something at the 2nd level. But people
apparently often view the heading levels as "levels of importance",
and some unfortunate wordings in specifications might contribute to
this misconception.
In a non-CSS document, skipping heading levels is a way to stop the
sub-headings from being so ENORMOUS :->

--
Max Demian
Apr 10 '07 #4

P: n/a
Scripsit Max Demian:
In a non-CSS document, skipping heading levels is a way to stop the
sub-headings from being so ENORMOUS :->
Did you actually read the message that you quoted?

Maybe your comment has some subtle sarcasm in it, despite the use of an
emoticon. After all, if _skipping_ levels in the strict sense, e.g. having
h1 followed by h4 with no intervening h2 and h3, is supposed to _solve_ the
size issue, then it's really a laughable "solution". It creates a strong
discrepancy between font sizes of h1 and an element that is supposedly meant
to be logicall a 2nd level heading.

Anyway, people who surf around the net these days with a non-CSS browser or
with CSS support disabled can surely be expected to be interested in the
_content_ and logical _structure_ of pages rather than visual experience.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 10 '07 #5

P: n/a
"Jukka K. Korpela" <jk******@cs.tut.fiwrote in message
news:m9*******************@reader1.news.saunalahti .fi
Scripsit Max Demian:
>In a non-CSS document, skipping heading levels is a way to stop the
sub-headings from being so ENORMOUS :->

Did you actually read the message that you quoted?

Maybe your comment has some subtle sarcasm in it, despite the use of
an emoticon. After all, if _skipping_ levels in the strict sense,
e.g. having h1 followed by h4 with no intervening h2 and h3, is
supposed to _solve_ the size issue, then it's really a laughable
"solution". It creates a strong discrepancy between font sizes of h1
and an element that is supposedly meant to be logicall a 2nd level
heading.
Anyway, people who surf around the net these days with a non-CSS
browser or with CSS support disabled can surely be expected to be
interested in the _content_ and logical _structure_ of pages rather
than visual experience.
I was thinking of a non-CSS page.

h2 and even h3 are very large by default. Sometimes you just don't want the
font to be so large for a sub-heading, especially if it's a long one, which
would dominate the page.

I agree that using CSS is best, but for a simple, single, portable page
straight HTML might be appropriate.

--
Max Demian
Apr 10 '07 #6

P: n/a
On Apr 10, 1:37 pm, "Max Demian" <max_dem...@bigfoot.comwrote:
I was thinking of a non-CSS page.

h2 and even h3 are very large by default. Sometimes you just don't want the
font to be so large for a sub-heading, especially if it's a long one, which
would dominate the page.

I agree that using CSS is best, but for a simple, single, portable page
straight HTML might be appropriate.
Nothing prevents you using CSS in a simple, single, portable page. It
just means a style element rather than a link.

However, if you really wish to return to the markup of the previous
millenium, then put font elements inside the headings rather than
using the wrong semantics.

Apr 10 '07 #7

P: n/a
Darin McGrew wrote:
Another argument in favor (or if you prefer, in favour of skipping
heading levels is that sometimes, skipping heading levels better reflects
the document structure.
FWIW, I agree with you here.

For example

<h1= Document heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Heading for a "Case Study"

where case studies may appear within either a chapter or a subchapter. For
purposes of consistency, it may be considered preferable to have all case
studies at the same heading level.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact
Geek of ~ HTML/SQL/Perl/PHP/Python*/Apache/Linux

* = I'm getting there!
Apr 11 '07 #8

P: n/a
Toby A Inkster <us**********@tobyinkster.co.ukwrote:
Another argument in favor (or if you prefer, in favour of skipping
heading levels is that sometimes, skipping heading levels better reflects
the document structure.

FWIW, I agree with you here.

For example

<h1= Document heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Heading for a "Case Study"

where case studies may appear within either a chapter or a subchapter. For
purposes of consistency, it may be considered preferable to have all case
studies at the same heading level.
Absolutely. There are any number of circumstances where the logical
structure of a document or of information sometimes skips a level. This
is not just about consistency of appearance either, but a matter of
reflecting the logical structure of the document.

Daniele
Apr 11 '07 #9

P: n/a
Toby A Inkster wrote:
Darin McGrew wrote:
>>Another argument in favor (or if you prefer, in favour of skipping
heading levels is that sometimes, skipping heading levels better reflects
the document structure.

FWIW, I agree with you here.

For example

<h1= Document heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Heading for a "Case Study"

where case studies may appear within either a chapter or a subchapter. For
purposes of consistency, it may be considered preferable to have all case
studies at the same heading level.
Well, then: how about we split the difference. Keep the semantic logic
of heading *levels*, but bend their *presentation* to suit the content.

h1 {font-size:180%;} /* or whatever sizes we really like */
h2 {font-size:160%;} /* and bold and text-align as we like */
h3 {font-size:140%;}
h4 {font-size:120%;}
..caseStudy {font-size:120%; bottom-border:1px dotted grey;}

....
<h2>Chapter 3: A New Dawn<h2>
<p>yadda yadda</p>

<h3>The Details</h3>
<p>More blah blah with intro text for a fascinating case study</p>

<h4 class="caseStudy">Case Study: George Lucas at 43</h4>
<p>text of the fascinating study</p>

<h2>Chapter 4: Two Squared</h2>
<p>still more text</p>

<h3 class="caseStudy">Case Study: Carrie Fisher's Hair</h3>
<p>Text from this ludicrous example...</p>
etc.

This may _look_ mixed up to the fastidious reader, since the next
heading size jumps down unexpectedly, but that's what matches the
content, if you say you want all case studies to look the same.
Meanwhile, the heading level flow matches Spartanicus' strictness.

--
John
Apr 11 '07 #10

P: n/a
John Hosking wrote:
Well, then: how about we split the difference. Keep the semantic logic
of heading *levels*, but bend their *presentation* to suit the content.

h1 {font-size:180%;} /* or whatever sizes we really like */
h2 {font-size:160%;} /* and bold and text-align as we like */
h3 {font-size:140%;}
h4 {font-size:120%;}
.caseStudy {font-size:120%; bottom-border:1px dotted grey;}
Nope, still not happy with that as a solution at all. It has nothing to do
with font size, or any other aspect of how the headings look -- just a
feeling of uneasiness over some case studies being afforded a higher level
than others.

Imagine a summarising program that took a document's structure from its
headings, and allowed you to generate an even shorter summary from
trimming this heading tree down to just H1-H3. Imagine the difference
between the resultant precis using each of the two heading models
discussed here.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact
Geek of ~ HTML/SQL/Perl/PHP/Python*/Apache/Linux

* = I'm getting there!
Apr 11 '07 #11

P: n/a
Scripsit Toby A Inkster:
For example

<h1= Document heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Heading for a "Case Study"
Fine.
where case studies may appear within either a chapter or a
subchapter.
Then the division into chapters and subchapters has been made inadequately
and should be fixed. The document structure is not balanced if some chapters
contain subchapters and some don't.
For purposes of consistency, it may be considered
preferable to have all case studies at the same heading level.
But in your example, they would not be at the same heading level. If a case
study appears after a 2nd level heading without an intervening heading, it
is at heading level 3. Using markup that claims it to be at level 4 would be
semantically wrong.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 11 '07 #12

P: n/a
On Wed, 11 Apr 2007 22:09:12 +0300, Jukka K. Korpela put finger to
keyboard and typed:
>Scripsit Toby A Inkster:
>For example

<h1= Document heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Heading for a "Case Study"

Fine.
>where case studies may appear within either a chapter or a
subchapter.

Then the division into chapters and subchapters has been made inadequately
and should be fixed. The document structure is not balanced if some chapters
contain subchapters and some don't.
Why? Real life isn't necesarily that neat. If a document relates to a
real life situation, then there are quite likely to be cases where
some chapters require subchapters and some don't.
>For purposes of consistency, it may be considered
preferable to have all case studies at the same heading level.

But in your example, they would not be at the same heading level. If a case
study appears after a 2nd level heading without an intervening heading, it
is at heading level 3. Using markup that claims it to be at level 4 would be
semantically wrong.
It's level 4 with respect to the document as a whole, even if it isn't
with respect to that particular page of the document.

Mark
--
Visit: http://www.CorporateContact.info - phone and email contacts for Amazon, Paypal, eBay and lots of other hard-to-contact organisations
"A sky isn't always blue, a sun doesn't always shine. It's alright to fall apart sometimes"
Apr 11 '07 #13

P: n/a
Scripsit Mark Goodge:
>Then the division into chapters and subchapters has been made
inadequately and should be fixed. The document structure is not
balanced if some chapters contain subchapters and some don't.

Why? Real life isn't necesarily that neat. If a document relates to a
real life situation, then there are quite likely to be cases where
some chapters require subchapters and some don't.
Strangely enough, no real life case has been cited in this discussion. An
example of a page where one needs to skip heading levels might be an
interesting challenge, though probably an easy one, to anyone who wants to
refute the need.

Real life (which probably spells "pointy-haired boss" here) might prevent
the author from structuring a document properly, but this does not mean that
the resulting markup would be adequate. If you have been forced into using
illogical markup, I won't blame you, as long as you're not trying to present
it as a virtue.
It's level 4 with respect to the document as a whole, even if it isn't
with respect to that particular page of the document.
There's no level 4 in that respect, since heading levels only exist in
relation with the implied logical nesting of parts of a document. If level 4
heading does not mean a heading for something that is nested inside
something at level 3 etc., what _does_ it mean?

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 11 '07 #14

P: n/a
On 2007-04-11, Jukka K. Korpela <jk******@cs.tut.fiwrote:
Scripsit Mark Goodge:
>>Then the division into chapters and subchapters has been made
inadequately and should be fixed. The document structure is not
balanced if some chapters contain subchapters and some don't.

Why? Real life isn't necesarily that neat. If a document relates to a
real life situation, then there are quite likely to be cases where
some chapters require subchapters and some don't.

Strangely enough, no real life case has been cited in this discussion. An
example of a page where one needs to skip heading levels might be an
interesting challenge, though probably an easy one, to anyone who wants to
refute the need.

Real life (which probably spells "pointy-haired boss" here) might prevent
the author from structuring a document properly, but this does not mean that
the resulting markup would be adequate. If you have been forced into using
illogical markup, I won't blame you, as long as you're not trying to present
it as a virtue.
>It's level 4 with respect to the document as a whole, even if it isn't
with respect to that particular page of the document.

There's no level 4 in that respect, since heading levels only exist in
relation with the implied logical nesting of parts of a document. If level 4
heading does not mean a heading for something that is nested inside
something at level 3 etc., what _does_ it mean?
It could mean "an item of merely quaternary importance". If so then
there's no need to nest headings properly-- my whole document might only
deal with items of quaternary importance. The ranking is in relation to
other things generally, not restricted to the contents of this document.

This is how to make sense of Toby's point I think-- "Case Studies" are
all of the same general importance regardless of how the document is
structured.

This seems a reasonable way to understand the meaning of <h1>, <h2>,
<h3>. The alternative is to use classes, and keep heading levels for
nesting, so:

<h3 class="notVeryInteresting">Case Studies</h3>

for example. Then I can put a veryImportant section inside a
notVeryInteresting section and still get a hierarchical TOC.
Apr 11 '07 #15

P: n/a
Toby A Inkster wrote::
>For example

<h1= Document heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Heading for a "Case Study"

where case studies may appear within either a chapter or a
subchapter.
Jukka K. Korpela <jk******@cs.tut.fiwrote:
But in your example, they would not be at the same heading level. If a case
study appears after a 2nd level heading without an intervening heading, it
is at heading level 3. Using markup that claims it to be at level 4 would be
semantically wrong.
That depends on what you think the underlying structure is.

I think the underlying structure is document heading, chapter heading,
subchapter heading, case study heading, etc. However, HTML does not include
elements for these types of heading, so somehow we have to map HTML's H1-H6
headings to the underlying structure.

It seems that you think it's more consistent to preserve the hierarchy of
HTML's H1-H6 headings and to use class="caseStudy" to identify the case
study headings. Others of us think it's more consistent to use HTML's H4
heading consistently for case study headings.

To me, the class="caseStudy" approach is too close to div soup.
--
Darin McGrew, mc****@stanfordalumni.org, http://www.rahul.net/mcgrew/
Web Design Group, da***@htmlhelp.com, http://www.HTMLHelp.com/

"If you aren't part of the solution, then you are part of the precipitate."
Apr 11 '07 #16

P: n/a
On Thu, 12 Apr 2007 00:08:11 +0300, Jukka K. Korpela put finger to
keyboard and typed:
>Scripsit Mark Goodge:
>>Then the division into chapters and subchapters has been made
inadequately and should be fixed. The document structure is not
balanced if some chapters contain subchapters and some don't.

Why? Real life isn't necesarily that neat. If a document relates to a
real life situation, then there are quite likely to be cases where
some chapters require subchapters and some don't.

Strangely enough, no real life case has been cited in this discussion. An
example of a page where one needs to skip heading levels might be an
interesting challenge, though probably an easy one, to anyone who wants to
refute the need.
You're right, it is easy. Consider, for example, a product catalogue
of parts:

Level 1 - manufacturer name
Level 2 - product name
Level 3 - product variant
Level 4 - part name/number

For a product which has multiple variants or models, level 3 will be
needed. Where a product only has one model, level 3 will be
superfluous.

Or consider a tourist guide to various towns and cities:

Level 1 - name of region
Level 2 - name of city
Level 3 - subdivision of city
Level 4 - individual tourist attraction

Small towns and cities probably won't need subdividing. You can fit
all the attractions in my howetown on to one page! But if the city is
London or Berlin, then subdividing by map sector makes sense.
>Real life (which probably spells "pointy-haired boss" here) might prevent
the author from structuring a document properly, but this does not mean that
the resulting markup would be adequate. If you have been forced into using
illogical markup, I won't blame you, as long as you're not trying to present
it as a virtue.
It isn't illogical to structure a document by reference to the data it
contains. It's illogical to force a unified structure onto a document
if the data it contains isn't that consistent.
>It's level 4 with respect to the document as a whole, even if it isn't
with respect to that particular page of the document.

There's no level 4 in that respect, since heading levels only exist in
relation with the implied logical nesting of parts of a document. If level 4
heading does not mean a heading for something that is nested inside
something at level 3 etc., what _does_ it mean?
Levels don't necessarily imply nesting. They imply status - a level 4
section on one page is equivalent in status to a level 4 section on
another page, even if the first page has a level 3 section and the
second page does not.

Mark
--
Blog: http://mark.goodge.co.uk Photos: http://www.goodge.co.uk
"My shallow heart's the only thing that's beating"
Apr 12 '07 #17

P: n/a
Scripsit Darin McGrew:
Jukka K. Korpela <jk******@cs.tut.fiwrote:
>But in your example, they would not be at the same heading level. If
a case study appears after a 2nd level heading without an
intervening heading, it is at heading level 3. Using markup that
claims it to be at level 4 would be semantically wrong.

That depends on what you think the underlying structure is.
The HTML heading elements express a nesting structure. This doesn't depend
on what I think. I would not have designed a markup language that way, but I
can live with other people's design.
I think the underlying structure is document heading, chapter heading,
subchapter heading, case study heading, etc.
You can call 2nd level parts chapters and 3rd level parts subchapters; that
would match one particular way of naming parts of a document. This is
immaterial, however, since the HTML elements are independent of such naming.
If "case study heading" is supposed to mean a heading for a case study,
wherever it appears in a document's nesting structure, then it's simply
something that is not expressible in HTML.
However, HTML does not
include elements for these types of heading, so somehow we have to
map HTML's H1-H6 headings to the underlying structure.
The point is that the document should be structured so that comparable
parts, such as case studies, are at the same level of nesting, so that the
mapping is self-evident.
It seems that you think it's more consistent to preserve the
hierarchy of HTML's H1-H6 headings and to use class="caseStudy" to
identify the case study headings.
No, I didn't say so and I didn't think so. If you have decided to structure
a document so that comparable parts appear at different levels of nesting, I
will abstain from saying which markup you should use.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 12 '07 #18

P: n/a
On 2007-04-12, Jukka K. Korpela <jk******@cs.tut.fiwrote:
Scripsit Darin McGrew:
>Jukka K. Korpela <jk******@cs.tut.fiwrote:
>>But in your example, they would not be at the same heading level. If
a case study appears after a 2nd level heading without an
intervening heading, it is at heading level 3. Using markup that
claims it to be at level 4 would be semantically wrong.

That depends on what you think the underlying structure is.

The HTML heading elements express a nesting structure. This doesn't depend
on what I think. I would not have designed a markup language that way, but I
can live with other people's design.
Are you sure? The HTML specification says,

"There are six levels of headings in HTML with H1 as the most
important and H6 as the least. Visual browsers usually render more
important headings in larger fonts than less important ones."

And then remarks:

"Some people consider skipping heading levels to be bad practice. They
accept H1 H2 H1 while they do not accept H1 H3 H1 since the heading
level H2 is skipped."

It doesn't say for example, "Numbered headings denote nesting levels of
sections within a document, *not* the relative importances of each
section".

It seems the primary meaning is level of importance, and as you suggest,
that may be isomorphic with nesting level in many documents. But it is
not necessarily so.

[...]
>However, HTML does not
include elements for these types of heading, so somehow we have to
map HTML's H1-H6 headings to the underlying structure.

The point is that the document should be structured so that comparable
parts, such as case studies, are at the same level of nesting, so that the
mapping is self-evident.
Why?
>It seems that you think it's more consistent to preserve the
hierarchy of HTML's H1-H6 headings and to use class="caseStudy" to
identify the case study headings.

No, I didn't say so and I didn't think so. If you have decided to structure
a document so that comparable parts appear at different levels of nesting, I
will abstain from saying which markup you should use.
Fair enough.
Apr 12 '07 #19

P: n/a
Jukka K. Korpela wrote:
Strangely enough, no real life case has been cited in this discussion. An
example of a page where one needs to skip heading levels might be an
interesting challenge, though probably an easy one, to anyone who wants to
refute the need.
Here's a book for you: Watt & Brown "Programming Language Processors in
Java". (Good book by the way.) It's structured into chapters, subchapters,
sub-subchapters and examples. Although all chapters have subchapters, not
all subchapters have sub-subchapters. Imagine typing this document up
into an HTML document. Using the structure for which I've been arguing,
you'd have:

<h1= Book heading
<h2= Chapter heading
<h3= Subchapter heading
<h4= Sub-subchapter heading
<h5= Example heading

If you wanted to pull out a table of contents, not listing examples, then
it would be a simple matter of pulling out <h1-4>.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact
Geek of ~ HTML/SQL/Perl/PHP/Python*/Apache/Linux

* = I'm getting there!
Apr 12 '07 #20

P: n/a
Scripsit Ben C:
>The HTML heading elements express a nesting structure. This doesn't
depend on what I think. I would not have designed a markup language
that way, but I can live with other people's design.

Are you sure? The HTML specification says,

"There are six levels of headings in HTML with H1 as the most
important and H6 as the least. Visual browsers usually render more
important headings in larger fonts than less important ones."
D'Oh! I was sure that nobody would pick up that unfortunate sloppy wording,
but I forgot this is c.i.w.a.h. where many people actually check things out
and are clever.

The wording, taken literally, is absurd. The importance of headings varies
in many ways. If you have a document that describes, say, agriculture in
different countries, with country names as natural headings, the
"importance" principle would imply that headings from h1 to h6 should be
used depending on the importance of agriculture in each country, from
<h1>USA</h1down to <h6>Vatican</h6>.

So it is logical to assume that they meant something different. They just
expressed it poorly. The good old HTML 2.0 specification, a great
improvement over its successors in clarity, says:

The six heading elements, <H1through <H6>, denote section headings.
Although the order and occurrence of headings is not constrained by
the HTML DTD, documents should not skip levels (for example, from H1
to H3) - -.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 12 '07 #21

P: n/a
Scripsit Mark Goodge:
Or consider a tourist guide to various towns and cities:

Level 1 - name of region
Level 2 - name of city
Level 3 - subdivision of city
Level 4 - individual tourist attraction

Small towns and cities probably won't need subdividing. You can fit
all the attractions in my howetown on to one page! But if the city is
London or Berlin, then subdividing by map sector makes sense.
OK, it seems that I have to surrender. That's a good example, and so is the
other one, and Toby's example. What they share is an overall hierarchy, with
some branches going deeper than others, together with some lowest-level
headings, though conceivably e.g. the individual tourist attraction
descriptions themselves could have internal structure that requires
lower-level headings.

So can we conclude that heading levels _should not_ be skipped, in an
RFC'ish sense, i.e. they normally must not be skipped but they may be
skipped in exceptional situations after due consideration of the situation,
the needs and implications?

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/

Apr 12 '07 #22

P: n/a
On Thu, 12 Apr 2007 22:16:05 +0300, Jukka K. Korpela put finger to
keyboard and typed:
>Scripsit Mark Goodge:
>Or consider a tourist guide to various towns and cities:

Level 1 - name of region
Level 2 - name of city
Level 3 - subdivision of city
Level 4 - individual tourist attraction

Small towns and cities probably won't need subdividing. You can fit
all the attractions in my howetown on to one page! But if the city is
London or Berlin, then subdividing by map sector makes sense.

OK, it seems that I have to surrender.
That's a very refreshing response, for Usenet :-)
That's a good example, and so is the
other one, and Toby's example. What they share is an overall hierarchy, with
some branches going deeper than others, together with some lowest-level
headings, though conceivably e.g. the individual tourist attraction
descriptions themselves could have internal structure that requires
lower-level headings.

So can we conclude that heading levels _should not_ be skipped, in an
RFC'ish sense, i.e. they normally must not be skipped but they may be
skipped in exceptional situations after due consideration of the situation,
the needs and implications?
I wouldn't put it quite that strongly. I'd argue that heading levels
should be followed, *unless* the structure of the data is better
presented otherwise. That is, the default is to follow the levels, but
it doesn't necessarily have to be "exceptional" not to.

Mark
--
Blog: http://mark.goodge.co.uk Photos: http://www.goodge.co.uk
"Every whisper, every waking hour"
Apr 12 '07 #23

This discussion thread is closed

Replies have been disabled for this discussion.