David posted a second version of the same reply with lots of interesting
responses, but unfortunately it was given the wrong subject: "oops!Re:
Designing web pages for screen readers". I'm redirecting back to the
original thread, with replies to some comments I didn't get in the
earlier version.
David Poehlman wrote:
dp: the effect of this can and is being reduced by stating in text what one
will find. I often see this now on sites where they say go to your account
where you can... so one link taking you to an index with an explanation of
what is contained therein does not or does not have to deminish its
effectiveness.
A block of text describing what links you can click in the index
followed by a link to an index is worse than just providing
the links in the first place. It does not reduce the clutter on the
page, adds an extra step to navigation, wastes time, occupies bandwidth,
makes users read their destination options twice, and is not as easily
skipped as a group of links.
What I do think is a good idea on a large site is providing minimal
navigation that links to other hub pages which provide further, more
specialized navigation.
I've seen a lot of people click through nav links only to click back
realizing they went to the rong place.
Well, you were there not me, but that sounds like a result of poorly
labeled links, not the presence of links.
I do believe there are other approaches but it
really is not necessary and is downright intrusive to the experience to
clutter every page with sight nav. I want to read the story. I want to
check out, I want to interact with my shopping cart etc. I don't want to
have to hunt all over every page to do it.
Well, users with average vision don't have to hunt around since they can
get a gestalt view of the page and will ignore the bits they aren't
interested in /until they need them/. The next versions of HTML and
XHTML will hopefully make it even easier for users with visual
impairments to skip, not just over links, but directly to content.
Publishers can provide a skip link as an interim measure, if they want,
though this is most useful to mobility impaired users who have been
poorly served by user agents. Skipping over one index link or activating
a skip link is not necessarily slower than skipping over a group of
navigation links: both require one command in JAWS or Window-Eyes.
Notably, many users arrive via a search engine at the wrong page. The
experience available from the page is not necessarily the experience
they actually came for. A bit of navigation helps them get where they
want to go fast.
dp: this assumes sites need to be complex in the first place which they do
not.
I'm sure lots of websites are needlessly complicated to use. But any
sufficiently large site is either going to have a complex structure or
be a chaos. People need (and demand!) help in navigating complex structures.
http://www.useit.com/alertbox/breadcrumbs.html
dp: requires a lot of reading to understand. May be good visually but is
not necessary to achieve the purpose and drowns the auditory and braille
user in a ton of > signs.
What other way of letting people know at a glance where they are within
the site structure would you suggest?
The > signs have the correct meaning here of "greater than". See this
discussion from RNIB, who use them in their breadcrumbs:
http://www.rnib.org.uk/xpedio/groups/public/documents/PublicWebsite/public_asciiart.hcsp
Now it's within the capacity of CSS to add greater than signs to
breadcrumb items in screen media but not aural or braille media, though
Internet Explorer does not yet support this feature and little AT pays
attention to the distinction between media types. But I'm not sure that
how the breadcrumbs relate to each other would be properly communicated
if the greater than signs were removed from screen reader output. There
is some discussion of adding specific markup support to HTML and XHTML
for breadcrumbs. This would allow more customization of how breadcrumbs
are read, though it's not obvious to me what sort of customizations
people would want to do.
The drowning of the user in breadcrumbs is, however, completely
avoidable, since with supporting tools they can skip the breadcrumbs.
They're just a group of links.
I wrote:
Assuming you don't drown content with navigation, it imposes no
cognitive load on sighted users because they've learnt to simply ignore
it until they need it:
David Poehlman wrote:
dp: unless they are just starting out.
If they're just starting out, they'll need to hunt around to some degree
whatever design you adopt.
dp: or to avoid doing the thinking up front. People need to be made
to think. Most don't want or need to be played down to. I won't visit
a site that makes me feel dumb.
Let's say I want to buy some wine. If I go to a supermarket, I want to
be able to locate the wine section as quickly as possible. I don't want
to spend ages learning about the supermarket's unusual layout in order
to find the wine section, because that's of little interest to me. I
want to spend that time thinking about what wine to buy. Thus I want to
be in the wine section as fast as possible. The supermarket owner, on
the other hand, needs to balance my need to get to the wine as fast as
possible with the chance to interest me in the supermarket's other
wares. Making me learn about the supermarket's layout takes time away
that could be spent looking at the latest offers.
You said earlier that sites should be kept simple and that users
shouldn't have to hunt around too much for the experiences they came
for. I can't reconcile those ideas, with which I agree, with the
assertion here that people should be made to think their way through the
logic of one's quirky design, rather than spending the same time doing
whatever they came to the site to do.
Now my wine-buying analogy cuts both ways: in favour of both retaining
navigation and minimizing it. I'm comfortable striking that kind of
balance on a page-by-page basis. One way of minimizing navigation is to
cut it down and another is to equalize the playing field by giving all
users the ability to skip to content.
"Don't make people think" isn't about dumbing down: it's about keeping
the site's interface simple enough to give the users time and space in
which to do the interesting thinking about the site's content and
functionality. As a user, I'm all for it, and as a developer, I'm happy
to go the extra mile to accomplish it.
dp: this seems to me to point out the opposite. They don't think before
clicking because they know what ok will do not because they don't want to
think. If however, we did things better, we would provide the user with
viable alternatives.
They don't know what OK would do (well, except for making the annoying
dialog go away). We know they don't know what OK would do because if you
ask them they can't explain what the message is about or why it might be
important. If they knew what OK would do they would probably click
Cancel to avoid being phished. For a lengthier discussion, see Peter
Guttmann's "Phishing Tips and Techniques" presentation (apologies for
the PDF):
http://www.cs.auckland.ac.nz/~pgut001/pubs/phishing.pdf
Regards
--
Benjamin Hawkes-Lewis
