W3C home > Mailing lists > Public > public-html@w3.org > March 2013

Re: Proposal: Provide better advice on constraint validation instructions

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Thu, 14 Mar 2013 21:11:48 +0100
To: Steve Faulkner <faulkner.steve@gmail.com>
Cc: HTMLWG WG <public-html@w3.org>
Message-ID: <20130314211148745702.ffa4b30b@xn--mlform-iua.no>
Steve Faulkner, Thu, 14 Mar 2013 18:13:31 +0000:
>> I don't understand why you keep telling me what the AX inspector says
>> when I have made I it so very clear that I fully trust that you give
>> accurate information about the AX inspector.
> I was talking about what voicoever announced not what inspector displayed

OK. Having switched VoiceOver to English, I can verify. My very bad.

> * I have explained to you about repeated text - and no repeated texts.
> you have drawn a conclusion about repeated text, not one that I share based
> on experience with observing how users interact with forms.

More attention than conclusion, IMO.

To a VoiceOver user, what comes across as most 'clean'? The title 
attribute solution currently in the text, or one or both of your 
proposed replacement examples? 

If L = 'Part number', H = the instruction text and E = cue to edit the 
text field, then the VoiceOver experiences can be summarized like so:˙

  Current spec: LHLE
Your example 1: LLHEH
Your example 2: LLEH

All the above are automatically read to the user. The difference 
between your two examples w.r.t. what is automatically read to the user 
before the edit text cue, is interesting.

Regarding help tag upon request: In the current spec example, the title 
attribute is used as help tag text. In your first example, there is no 
specific 'help tag', as VO sees it - everything is read if the user 
asks for help tag. In your second example, the aria-describedby text is 
used as help tag.

To me, the behavior that the current spec example generates, seems more 
straight forward. But of course, I don't have not observed users using 
a form. Neither do I know how other ATs present the examples. 

But, at least seen from a VoiceOver perspective, there are some 
drawbacks to the solutions you propose. Also, I’m sure many authors 
would like to have invisible help tags/descriptions.
Leif Halvard Silli
Received on Thursday, 14 March 2013 20:12:18 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:16:32 UTC