E-mail List Archives

Re: page should contain no more than two h1 elements


From: Karl Groves
Date: Jun 15, 2009 2:15PM

> > My own observation shows it to be
> > exactly as Jared described, but I also recognize that the
> screenreader
> > users I know are all power users. I'm not sure how they compare to
> > "normal" users.
> My observations are similar. So let's go ahead and call it best
> practice and be done with this controversy forever, eh?

Oh, I don't look at it as a "controversy", just a healthy discussion

> The point here is that "best practice" is very subjective and open to
> interpretation, particularly in web accessibility. I've found that
> most in the accessibility field seem to agree that one h1 that is the
> document title makes the most sense for most typical web pages. But
> there are also some that extend this to mean that h1 must be first or
> that h1 must be the exact same as the <title>. Where does "best
> practice" end and extremism begin?

IMO, a best practice becomes a dogma whenever it is viewed as the only
possible approach to an issue - such as those we've seen by other replies
in this thread. A quick glance at the SSB site or any of my own personal
sites shows where I sit on this issue: One H1 element on each page. That
doesn't mean, however, that this is the only appropriate approach.

> It's quite difficult to scientifically test these types of things, and
> even if you could, I think you'd find the same general findings that
> WebAIM's survey found - that screen reader users are very diverse and
> do things a lot of different ways. But this is the type of thing that
> we will be delving deeper into in a future survey.

I think it is great that you guys did that survey and hope you continue
with similar efforts. If you need any assistance translating it into
other languages (or anything else, for that matter), let me know.

Another thing that's sorely missing in this area is usability studies. As
far as I'm aware, the overwhelming majority of usability studies of
assistive technology users that exist are qualitative in nature and more
than likely going to be performed for a client (read as: not for public
consumption). Ginny Redish and Mary Theofanos did a few usability
studies that were published. Everything else that I'm aware of is
private. I've said for years that it would be great to collect real,
quantitative data on topics like this. The level of effort required might
be a bit prohibitive, though.

> PS - What a great discussion!

I agree!