W3C home > Mailing lists > Public > wai-tech-comments@w3.org > August 2001

Re: [webwatch] Web-Based E-mail and Message Check Boxes

From: Al Gilman <asgilman@iamdigex.net>
Date: Fri, 24 Aug 2001 10:42:05 -0400
Message-Id: <200108241422.KAA6901095@smtp2.mail.iamworld.net>
To: webwatch@yahoogroups.com
At 07:51 AM 2001-08-24 , Kelly Ford wrote:
>Hi All,
>
>The typical web-based e-mail found at places like Hotmail, Yahoo, and such 
>usually includes a check box for each message.  The check box is used for 
>selecting the message for things like deleting or moving to folders.
>
>I would be interested in people's thoughts/opinions on what you'd like to 
>experience as the label for check boxes in such situations.  [snip, full
message below] 

[Al, here]

Can we broaden the question a bit?  I could use a reference point in "if I ran
the Zoo" terms for how these functions might work.

What was your all-time favorite email interface?  Is it NetTamer?  Pine?  What
is it?  Forget the Web, forget what you have learned about using contemporary
operating systems.

What would you _really_ like your mail tool to work like?

I say this because my kneejerk reaction to "how do we label the checkbox" is
that this is the wrong question, that eyes-free users shouldn't have to bother
with the checkbox, that's a graphical metaphor for something more
fundamental. 
Can we get at the more fundamental something?  If there were other stuff in
the
speech access mode so that you basically don't know or care that there is a
checkbox, what would that be?  

The dream is that you just read and manage your mail.  The checkbox is a
medium-adaptive feature for the visual interface, not the definition of the
interface.  The definition of the interface is "this message is marked for
deletion."  That is a piece of the state of the message in your session
buffer.  The key at this point is that it is the state of the message, not the
state of the checkbox, that you want to manipulate.

And you probably want a "delete this message" keystroke command and not a
"delete this message" checkbox to make the interface work well in keyboard
plus
speech.  Navigating to the checkbox is just a detour you don't need.  What
message you are talking about can be obvious.  Make the system deliver this
service.

Managing your email, that is being processed by a web service like hotmail, is
the prototyping lab of many many web services to come.

For eyes free access, do you want to focus more on the current message and do
with it what you are going to do with it?   What I would expect you are not
going to do is review the list of messages, with its marks on those to be
deleted, as a list, the way the visual user would naturally do with the
standard GUI design of this service.  

So let's make the delete that you do while you have the message selected more
sure, put in enough confirmation at this point, so that review later is not
required.  And get it over with while you have the message to be deleted in
mind.  After all, when the information gets to the server, deletion or filing
will be done on a message by message basis.  That is the level of the object
that you are manipulating.  

Pages of folder contents listings are the medium of the exchange in the dialog
only because that is natural when visually reading a half-page-sized screen. 
It's not the natural shape of the dialog in keyboard plus speech.  So let's
get
the ideal answer on the table first and see how much of that we can get rolled
into the standard definition of the Web, before we start compromising.  There
may be a better deal in this somewhere.

OK, now let me backpedal a bit.  It's fine for the checkbox to show the
"marked
for deletion" status of the mail item.  And be there for the visual user to
change this status bit with.  But the medium in which we interface to the web
service has to have multilevel scoping in how it handles verbs, multiple verb
options in any context, so that the system knows that 'delete' is at the
message level and it can have a current selection more locally on the author,
the subject, etc. but if you emit the 'delete' hotkey or shortcut, it knows to
handle this at the message level and marks it in the 'delete' checkbox without
you having to move to the checkbox to do it.

Everybody, please do give your feedback.  

There is a feature in HTML 4 called ACCESSKEY that we may wish to re-engineer
slightly in XHTML 2.0.  Maybe it can be made to do what you really want.  Tell
us what that is.  The web mail scenario is high on the list of those where we
will learn what it is we need to be doing.  

Al

At 07:51 AM 2001-08-24 , Kelly Ford wrote:
>Hi All,
>
>The typical web-based e-mail found at places like Hotmail, Yahoo, and such 
>usually includes a check box for each message.  The check box is used for 
>selecting the message for things like deleting or moving to folders.
>
>I would be interested in people's thoughts/opinions on what you'd like to 
>experience as the label for check boxes in such situations.  At present 
>most screen readers default to reading the linked item for each message 
>i.e. the name of the message sender because that's what is acted on to open 
>the message.
>
>For people using programs like HPR, Window-Eyes and JAWS, the full header 
>information is available as they cursor through the buffered view of the 
>web page.  By buffered I mean the HPR item view, Window-Eyes MSAA mode and 
>JAWS Virtual PC.
>
>If you'd like to experience something different from the actionable text 
>for each message, how would you feel about having text duplicated.  To give 
>an example based on how such might read in the decolumnized view of the 
>page with the different programs.  In this example *C indicates where a 
>screen reader would say check box and *L link.  I'm also assuming here that 
>tables are used and should the user want, appropriate column header 
>information is available with columns being status, checked status (not a 
>visible label), from, subject, date and size.
>
>Existing design
>
>New
>*C
>*L Kelly Ford
>Testing Some E-mail
>8/24/2001
>3K
>
>When the user tabs to the check box, or activates it, he or she would hear 
>Kelly Ford, unchecked or something similar.
>
>It is possible to adjust the text of the check box as the following 
>illustrations show.
>
>Example 1
>
>New
>*C Select or Deselect Message
>*L Kelly Ford
>Testing Some E-mail
>8/24/2001
>3K
>
>One tradeoff of this example is that the user no longer gets the message 
>sender name when moving to the check box with tab or pressing enter to 
>activate it from a screen reader's buffered view of the page.
>
>It is also possible to include text from the message header.
>
>Example 2
>
>New
>*C Kelly Ford, Testing Some E-mail
>*L Kelly Ford
>Testing Some E-mail
>8/24/2001
>3K
>
>But in this case, it is likely the user is going to experience the text 
>duplication in reading through the web page as I've illustrated.
>
>The ideal might be to use header commands or something similar to only have 
>message header text spoken when the check box is activated.  In such an 
>example there wouldn't be any text after the check box in the buffered view 
>of the page.  However I don't believe that this kind of design is supported 
>by the assistive technology at present.
>
>Although I'm asking my question about e-mail, the same question could be 
>asked about a number of web pages.  For example you might have a table 
>listing products from a shopping site and check boxes to indicate which 
>products you want to add to your cart.  In such an example, the check boxes 
>would typically again only speak the actionable item from each product, 
>although it would be a bit more descriptive because of the nature of the
page.
>
>Anyway, I'd be interested in people's thoughts/opinions.
>
>Thanks,
>
>Kelly
>
>
>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
<http://docs.yahoo.com/info/terms/>http://docs.yahoo.com/info/terms/ 
>  
Received on Friday, 24 August 2001 11:45:40 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:24:25 UTC