[thelist] Coding standards.... [headers]

kasimir-k kasimir.k.lists at gmail.com
Sat Dec 9 11:12:49 CST 2006

Joel D Canfield scribeva in 09/12/2006 15:35:
>> it is fine to use H1, H3, H2 if that's how your document flows.
> how would a document flow like that, semantically?

H1 header of the first section, most important
    H3 header of subsection of the first section, not very important
       Paragraphs, some additional information...
H2 header of the second section, not as important as the first

> how would you get to
> the third most important point before the second most important point?

If the third most important point belongs together with the most 
important, as above, but of course there's countless number of other 
ways too.

> I'm not saying it's impossible, it just sounds like quite an exception.

If on a page I want to have summary of some other pages covering the 
same subject as this page, and only then start this page content, then I 
might have H5 or H6 in the summary appearing before H1.

> this is a subject I'm still pretty ambivalent about, so I'm really just
> trying to come to a definitive concept in my own head, rather than
> leaving it as random as it is right now

The specs leave this open - they only talk about different levels of 
importance (and mention that some people consider skipping levels a bad 

A common good writing practice is the inverted pyramid where you have 
only one of each H-level appearing in order: H1, H2, ... H6. Sometimes a 
n inverted tree is needed:
And yes, sometimes there are good reasons not to put the most important 
content first.

But it's not random at all, actually quite simple. You organize your 
content on a page as is good and right in that particular case. Then you 
put H1 for the most important title, where ever it may be, etc. with 
other levels.

But how the content is organized is not about coding standards but 
content creation standards. The coding standard should simply be to use 
different levels of H according to the importance of the headings.



More information about the thelist mailing list