W3C home > Mailing lists > Public > public-html@w3.org > April 2009

Re: feedback on 4.10.4.2.8 The pattern attribute

From: Henri Sivonen <hsivonen@iki.fi>
Date: Thu, 16 Apr 2009 11:09:27 +0300
Cc: "Steven Faulkner" <faulkner.steve@gmail.com>, jgraham@opera.com, "HTMLWG WG" <public-html@w3.org>
Message-Id: <EAC41E43-A3AB-4FC4-BAEF-950D4F3C9B5C@iki.fi>
To: Charles McCathieNevile <chaals@opera.com>
On Apr 16, 2009, at 01:59, Charles McCathieNevile wrote:

> However, rel="help" or aria-describedBy is probably better than the  
> suggestion that the title attribute will do it, given that the title  
> attribute is a complicated mess of expectations and assuming  
> particular behaviour for it has proven to be problematic.


How could rel=help, which implies the help text would be structured  
text at another URI, be better than using a simple string from an  
attribute that Opera 9 already supports for this purpose at least  
visually? Would rel=help come with roughly similar cans of worms as  
longdesc?

If I enter an ftp URI in the address box on http://html5.validator.nu/  
in Opera 9 or 10 and hit enter, I get a visual popup that tells me the  
format of the input is wrong and then includes the value of the title  
attribute (per current spec).

Why wouldn't an HTML5 forms-aware browser expose that popup to AT the  
same way such a GUI popup would be exposed to AT in a non-browser app?  
Why should AT have to know which attribute the text came from?

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Thursday, 16 April 2009 08:10:30 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:02 UTC