W3C home > Mailing lists > Public > public-pfwg-comments@w3.org > April to June 2009

role="form" missing (closest construct: role="search")

From: Gregory J. Rosmaita <oedipus@hicom.net>
Date: Wed, 8 Apr 2009 17:50:18 +0100
To: public-pfwg-comments@w3.org
Cc: public-xhtml2@w3.org, wai-liaison@w3.org
Message-Id: <20090408164628.M16484@hicom.net>
aloha, all!

i don't know how this slipped through the cracks, but it seems to be
a rather glaring omission to me...

In re-reviewing the ARIA 1.0 Last Call draft for the umpteenth time,
it struck me that while there is a role for "search" there is no 
role for "form", as not all forms are search forms, and the "search"
role designates more than merely "here are form controls" but binds 
any explanatory and complimentary text included in the "search" 
area...

In neither Section 4.3.4. "Document Structure" [1], or Section 4.3.6. 
"Landmark Roles" [2] does there appear a "form" role

While there are specific form control roles defined by ARIA 1.0, there
is no "meta" role, "form" which can be used to bind all such elements 
(as well as any declarative FORM elements) into a cohesive section 
named "form" 

This is important to address for ARIA 1.0 because today's web is extremely
forms-driven -- e.g. networking sites, twitter, blogs, web mail clients, 
ecommerce, security prompts, CAPTCHAs, etc.  Therefore, it is necessary 
to have a "form" landmark region that provides a means of binding forms 
composed of a mixture of declarative markup and scripting into a single 
region

In contemplating the lack of a "form" role, i have convinced myself that 
the place for a "form" role is in the XHTML Vocab document, located at:

http://www.w3.org/1999/xhtml/vocab#

as "form" is a role that should properly be defined in the list of 
default roles.

Additionally, XHTML2 will be using XForms as its Forms module, so 
role="form" could serve as a flag to an application that an xforms (or 
javascript-enabled xform [3]) follows and that whatever special 
processing is necessitated should be fired and remain in effect until the 
user chooses to leave forms/xforms mode and re-enter "normal" (or 
"non-forms") mode or until the form/xform has been submitted.

the OWL for a such a landmark role named "form" would be:

   <owl:Class rdf:ID="form">
      <rdfs:subClassOf rdf:resource="#landmark"/>
      <role:nameFrom>author</role:nameFrom>
   </owl:Class>

but i think that the predefined role of "form" should be something
defined in the XHTML vocabulary document as a pre-defined role (that
is, part of the Role Module) and not "just" in ARIA 1.0

this topic was broached during the 8 April 2009 XHTML2 WG teleconference;
the discussion can be found using the following URI:

http://www.w3.org/2009/04/08-xhtml-minutes.html#item03

note that the chair of the XHTML2 WG, Roland Merrick, assigned himself 
an action item to discuss a "form" role with the XForms group -- it is 
highly advisable that any such discussion have PF participation, since
the "form" role is an urgently needed repair strategy for today's web
content, and a critical landmark for anyone attempting to use a mashed-
up form...

roland's action is documented at:

http://www.w3.org/MarkUp/tracker/actions/71

References:

[1] http://www.w3.org/TR/wai-aria/#roleatttribute_inherits

[2] http://www.w3.org/TR/wai-aria/#structural

[3] ubiquity-xforms: 
* http://code.google.com/p/ubiquity-xforms
* http://code.google.com/p/ubiquity-xforms/wiki/ImplementationStatus
Received on Wednesday, 8 April 2009 16:51:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:14:59 GMT