W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > October to December 2010

Re: Accesskey implementations...

From: Simon Harper <simon.harper@manchester.ac.uk>
Date: Wed, 01 Dec 2010 12:01:24 +0000
Message-ID: <4CF63914.1070809@manchester.ac.uk>
To: Charles McCathieNevile <chaals@opera.com>
CC: UAWG list <w3c-wai-ua@w3.org>
OK so..

On 01/12/2010 11:02, Charles McCathieNevile wrote:
> On Wed, 01 Dec 2010 11:44:21 +0100, Simon Harper 
> <simon.harper@manchester.ac.uk> wrote:
>
>> Hi Charles,
>>
>> So I really like the look of this!
>>
>> When you say 'allow users to mark a site as 'don't bother'' have you 
>> thought about using IBM TRLs crowdsourcing social accessibility 
>> project technology -- to help or contribute to this?
>> http://sa.watson.ibm.com/
>
> Nope. That would be closer to the stuff I want to do with enabling 
> sharing of preference settings (when they are a bit more useful to 
> start with). I outlined some rough thoughts in the documentation page.
>
>> Now also have you thought about allowing an accesskey definition on 
>> the page to be a semantic keywork from an approved lexicon - such as 
>> 'save' or the like. You could then assign a key based on user 
>> preference for certain functionality?
>
> Yep. For about a decade. I think the clear answer is that where there 
> are well-known or defined rel values, it's easy to do. But spotting 
> the difference between "Save the Whales (Donate now)!", "Guardar 
> fichero", "Ahorramas", "oh no, save me", and "save preferences" is 
> somewhat tricky otherwise.
>
Maybe a taxonomy of common functionality would help here - "Save", "Save 
As", "Save Preferences" would all be in there but accesskey="Save the 
Whales" would not. I'm sure this level of common functionality would be 
doable - indeed most OS come with common functionality added to the menu 
widget. In fact it could be cooler than that, such that if Opera were 
being used on mac it would convert "edit options" to "edit preferences"? 
If I make a user assignment that "edit preferences" are always 
[designated_browser_modifier-key]+p then I would always get this.


> The big use case for the rel attribute is that it is a semantic 
> marker, with allegedly defined meanings. So you can give it the same 
> behaviour in a user agent that enables it for any page which uses it. 
> I guess that was why it went into HTML4, and why people make 
> navigation bars to use it :)
>
> In this extension I might look up rel attributes and assign specific 
> accesskeys if they happen to arise. That's probably not so hard for a 
> handful of common ones, although it might be a bit trickier to make a 
> generalisable mechanism that is usable (which is what I want in my 
> boil-the-ocean moments...)
>
> cheers
>
> Chaals
>
> Cheers
> Si.
>
> =======================
>
> Simon Harper
> University of Manchester (UK)
>
> More: http://simon.harper.name/about/card/
>
Received on Wednesday, 1 December 2010 12:01:54 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:38:42 UTC