Re: HTML 5 Authoring Guidelines Proposal the use of the section element and its potential impact on screen reader users

Henri Sivonen wrote:
.....

>
> > From this point of view of course it would be simpler for
> > implementors to separate the two methods either using
> > section+h or using h1-h6. Separation of different functionalities
> > is mostly a good idea to avoid more difficulties as necessary.
>
> What should happen when the UA gets a mix anyway?

That is a general problem of the HTML5 approach - if authors do
not care about 'best practice' and intended usage models this
will in general create even more problems. 
Already that some elements may change their role somehow 
from beeing an inline or block element - the rules are somehow
complicate for authors. It is obviously possible for a certain amount
of people to memorise elements to be either of the group block 
or inline, but they will not be able to memorise more complex content
models, I suspect, looking on the questions/discussions in common 
forums about HTML ;o) 

In this situation - with two completely different models to structure
content it should be no surprise for authors to get surprising or
nonsense results from the viewer if they start to mix it and it would 
be even more educational for authors, if they get different results 
for index/outline with different viewers. 
This is a big advantage of a language with underspecified or 
unspecific error handling as HTML4 - the results are different 
(but maybe still readable) and therefore authors are informed, 
that there is some nonsense in the document, while readers
still may have access to the content. If the interpretation of any 
nonsense is the same in any viewer, authors will get even more 
insensible to any content model, because it still works, what they do,
whatever they do.
Too propper defined error handling is a sweet temptation to write
even more nonsense. And maybe in a few years authors will take the 
bait of HTML5 completely to write any nonsense, if the error 
handling provides the same appearance in any viewer.

In this case, it is some intellectual requirement to know, that
section-h is a different model as h1-h6. But if it works anyway to
mix it in any combination, obviously authors will not even try to
learn something about different models, they will simply jam 
everything together as they like, because there is no problem 
appearing, if rendering of 'any nonsense' is defined in the draft ;o)
Currently HTML5 anyway has the (arcticle+)section-h1-h6 mix,
therefore obviously there is no need for authors at all to care about
a consistent model for their documents and therefore apart from 
some aesthetical persons no one will care about it...
Even that one typically might start any section with h1 and not
h2-h6 depending on other headings outside the section will be
a challenge to explain it common authors in such a way, that
they use it in a consistent way ;o)  But that is more a problem 
for authors or education, not for implementors. 
I'm pretty sure, that it is possible to define an automatic somehow 
logical algorithmus to derive a content model for any mixture, even 
if no one will understand it, if such a mixture is really used in a 
document.

Received on Wednesday, 28 November 2007 14:11:58 UTC