W3C home > Mailing lists > Public > wai-xtech@w3.org > March 2008

Re: [DHTML Style Guide] auto-complete widget proposal

From: Chris Blouch <cblouch@aol.com>
Date: Thu, 27 Mar 2008 12:24:58 -0400
Message-ID: <47EBCA5A.1080508@aol.com>
To: Joseph Scheuhammer <clown@utoronto.ca>
CC: Becky Gibson <Becky_Gibson@notesdev.ibm.com>, wai-xtech@w3.org

Seems reasonable. So if we are to allow large sets of possible 
auto-complete values then I suggest we will need additional navigation 
keys beyond the up-down arrow to navigate the list. Maybe some variant 
of home, end, pageUp, pageDown etc. It would be nice if pageUp/Down 
scaled according to result set size. So at a minimum I could jump 10 
entries or 10% of the list, whichever is larger. I don't think that is a 
common behavior but it seems like it would be good for long lists.

CB

Joseph Scheuhammer wrote:
> Chris wrote:
>
>> Maybe this is implementation detail but should we qualify that the 
>> list of matching items (results) must be reasonably limited?
>
> I have modified my position after thinking about this a bit more.  
> Context is important here.  If the user is expecting to be offered the 
> most likely hits based on what they've typed, then the widget should 
> offer the "top ten" hits.  If they are not satisfied with those, a 
> gesture should bring up all possible results.  A concrete example is 
> the Google search bar showing the top ten results.  If I hit return 
> without selecting any of them, I go to a page that lists all matches 
> (to be precise:  a set of pages that lists all matches).
>
> However, if the user is not expecting the most likely results, or if 
> the context entails that the most likely results are not desired, then 
> the list should show all the results.  An example here is  an address 
> form, and the widget is for the, say, the country portion of the 
> address.  The user does not expect nor want the "top ten countries" 
> (there's no such thing), but all the countries whose names match what 
> they have typed thus far.
>
Received on Thursday, 27 March 2008 16:26:16 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:51:35 UTC