W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 2001

Re: [webwatch] Tabbing to Non-Actionable Elements on a Page

From: Al Gilman <asgilman@iamdigex.net>
Date: Sun, 18 Mar 2001 13:31:48 -0500
Message-Id: <Version.32.20010317113427.043306e0@pop.iamdigex.net>
Message-Id: <Version.32.20010317113427.043306e0@pop.iamdigex.net>
To: webwatch@yahoogroups.com
At 06:48 AM 2001-03-17 -0800, Kelly Ford wrote:
>Hi All,
>What are people's opinions about web designs that might allow people to tab 
>to what I'll call non-actionable elements on a web page.  I think of 
>actionable items as links and controls like edit boxes, check boxes and 
>such.  Non-actionable items would tend to be static text.

[Al, here]

** summary:

User profiling of the tab-sequence contents is an experimental topic. 
There is
no established guideline or best current practice.  It potentially competes
mindshare with other navigation mechanisms not bound to the TAB and Shift+TAB
keystrokes.  In particular inside a table you should probably be thinking
explicit two dimensional navigation with the arrow keys and not TABbing.

** more

This question has exercised the WAI, and the User Agent Working Group in
particular, a lot.  Some of the things you would like to move to are
headers or
even the starting points of functional blocks such as forms or tables even
where they have no specific header introducing them.

I haven't seen enough experimentation with working prototypes to understand
what the relationship should be between the tabbing sequence and navigating
with respect to a virtual table of contents tree structure.

In principle you could have a 'preferences' setting that would decide what
and does not make it into the tabbing sequence.  But how many users would
undertand and use this is an open question.

[starred section; may skip on first reading]

The more subtle issue is what is the relationship of objects with a "move to
me" method as compared with the more obviously active thingies.  In HTML it is
possible to create an anchor with a NAME and no HREF.  Most people would
say it
has no action.  Once you're there, there is nothing special you can do as a
function of being there.  But rather, what you can do specially is that you
go there directly by name.  This is an underutilized functionality on the Web
and an unexplored corner of the theoretical foundations (as I see it). 
ACCESSKEY creates a similar capability.  This is another way of looking at
question.  The "move to me" capabilities of objects in a page are related to
what you can do once you're there, but not synonymous with that question. 
the relationship should be is not much known beyond the idea that the usual
suspects to be treated as active elements should be on some easy to get there
list with high probability.

I regard a "tilt toward action" as an individual difference among web users. 
Even between different visits by the same user.  That is why I lean toward the
pruning rules for the tabbing sequence being controlled by personal
preferences.  Or multiple modes of navigation with TABbing being just one of
them.  Different people, and the same people at different times, will have
or less inclination to zero in on the things that they can _do_ in a page.

A wild card in this deck would be the ability to mark or recognize a
section of
a page as a "no tabbing zone -- read all contents."

[end starred section]

>Perhaps you want to give someone the ability to tab through the individual 
>data cells of a table for example.  This may help someone using screen 
>magnification or a screen reader that doesn't support table navigation 
>explore the table.

[Al, here]

Inside tables, any tabbing sequence is a poor substitute for real two
dimensional navigation.  The latter will be clear to users if bound to the
arrow keys.  Compare high-end table implementations such as in Home Page
and spreadsheet programs.

Notification on entry to and exit from tables and forms is probably a good
idea.  In other words, you don't necessarily have to stop at this boundary,
you should hear about it when you cross it.  Navigation to the start or end of
a table or form is structural navigation and should be thought about in one
context with frame and section navigation.

Also inside tables one is interested in varying levels of verboseness in
orientation.  The most obvious verb here is a "What's that?" that gets you an
expanded explanation of what is is that you are currently at.  This service is
similar to what is provided by onMouseOver effects in the GUI interaction


>I'm not sure if doing this is even valid HTML though.  I experimented with 
>putting tabindex on each cell in a table and also just the first cell in a 
>row and ZoomText, JAWS and Window-Eyes all read the text and focus seemed 
>to move to that cell.  But I don't believe this is valid HTML.
>Forgetting the validity of my specific coding for a minute, let's say I 
>have a way to make this happen.  What do people think of the idea as a 
>general concept?  Should the tab key only go to actional items or all 
>items, even if some of those items are not actionable?
>------------------------ Yahoo! Groups Sponsor ---------------------~-~>
>Make good on the promise you made at graduation to keep
>in touch. Classmates.com has over 14 million registered
>high school alumni--chances are you'll find your friends!
>To Post a message, send it to:   webwatch@eGroups.com
>To Unsubscribe, send a blank message to: webwatch-unsubscribe@eGroups.com 
>Your use of Yahoo! Groups is subject to
Received on Sunday, 18 March 2001 13:09:55 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:36 UTC