Skip Navigation
trace.wisc.edu HelpSearchBottom of Page

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Browser 'back' is_a Undo



In message <Version.32.20000206120937.01148680@pop.iamdigex.net> 
Al Gilman writes:

>>Forwarded message follows:
>>
>>Date:         Fri, 4 Feb 2000 10:57:55 +0100
>>Reply-To: Pascal MAGNENAT <pascal.magnenat@INTERACTIONS.CH>
>>Sender: "ACM SIGCHI WWW Human Factors (Open Discussion)" <CHI-WEB@acm.org>
>>From: Pascal MAGNENAT <pascal.magnenat@INTERACTIONS.CH>
>>Subject:      61% of users could not buy train tickets on the Swiss Federal
>>              Railways site
>>To: CHI-WEB@acm.org
>>
>>
>>Possible origins of poor usability:
>>
>>- unpredictable effect of the browser "back" button (many users tried to
>>delete an item thrown in the trolley [aka shopping cart] by this means)

>This is an absolutely stunning clue.
>
>In all the fighting about whether 'Back' should be a scripted button on the
>page or a browser function, I had not previously heard this.
>
>To make hypermedia browsing accessible to the iOpener customer base, we
>have to understand that "undo link traversal" is a special case of "undo"
>in the user's total experience.  We need to make sure that all these
>flavors of "undo" blend smoothly so that the user is not surprised by what
>they do.
>
>It's not just clicks that we need to sort out over the subprocess or layer
>stack to see who gets to handle them.  It's 'undos' too.  And the user
>doesn't want to know there is a difference.
>
>Al

In this case the confusion in the user's mental model is over whether 
"back" means "back in time" (i.e. undo) or "back in space" (i.e. return
to where you were).  "Undo link traversal" is clearly "back in space"
or simply "return".  The term "undo" is associated with a "do" or
"input" action, rather than a "go" or "navigation" action.  

This type of confusion arises when the web designer adds a button
labelled "back" on the page.  The user has no idea whether this has 
the same effect as the "back" button on the browser, and, if not, 
what the difference is.

If the user is to be allowed to undo as well as to return, 
there should be a specific "undo" button on the page.

As a general principle, I suggest:

"Leave the browser to do the functions it can do, and have the web 
applications only to do the functions which the browser cannot do."

Even little efforts by the web designer to "help" the user in
navigation may be misguided.  For example I have seen a "top"
button at the bottom of a page, designed to make it easier for
the user to get to the top of the page.  (The designer has
heard that users have difficulty scrolling, though the fact that
the user has got to the bottom of the page means that they have
managed to scroll down.)  The danger is that after pressing "top",
the next press of the browser's "back" button may, or may not, 
take you to the bottom of the page!  

What _should_ the browser do in this case?  

As a general rule, should the browser put links _within_ a page 
on the return stack or not?

WAI people, do the user agent guidelines have anything to say on this?

Cheers from Chiswick,

John (trying to build a better browser)
-- 
Access the word, access the world       Tel/fax +44 20 8742 3170/8715
John Nissen                             Email to jn@tommy.demon.co.uk
Cloudworld Ltd., Chiswick, London, UK   http://www.tommy.demon.co.uk