1 webpage != 1 document

Over on the pmwiki-users mailing list, we're having a discussion about
the use of heading tags in the sidebar and document structure.  You
can read the thread at
http://thread.gmane.org/gmane.comp.web.wiki.pmwiki.user/16355.  In
short, the pmwiki-users list is trying to decide how do we keep
headings used in the sidebar from wrecking the outline structure, and
from "outvoting" the page's real name in search engine indexes.  So
far the consensus is to stop using heading in the sidebar, and fake
them with some other element.  I feel that this is a lesser evil,
rather than a semantic improvement.

As I see it, the root problem here is that the model of a what webpage
is says that it's one document.  But when did you last see a
well-designed live webpage that contained *just* one document?  If the
W3C's site was constructed like that, we could only find other W3C
pages if they were linked in the body text, because there would be no
navigation links.  Logically speaking, navigation is never the page
content proper unless the page is a sitemap.

Best practice in web design demands plenty of site-related content in
every page, such as the masthead and navigation bar(s).  There may
also be document-related secondary content, like a sidebar for a
magazine story.  Evidently, real webpages contain more than just one
document each.

Does anyone else agree that the "1 webpage = 1 document" idea is flawed?  

What if we had a way to mark content separate from the page's primary
document, so that user agents can recognize these site-related and
document-related chunks, and consider their heading structure
separately from that of the primary document?

Bronwyn

Received on Tuesday, 16 August 2005 10:17:58 UTC