W3C home > Mailing lists > Public > www-forms-editor@w3.org > June 2002

[48 hour response please] Re: User understanding of what will be submitted by SUBMIT

From: Al Gilman <asgilman@iamdigex.net>
Date: Wed, 26 Jun 2002 09:52:25 -0400
Message-Id: <>
To: w3c-wai-pf@w3.org, "Steven Pemberton" <steven.pemberton@cwi.nl>, <www-forms-editor@w3.org>
Cc: <wai-liaison@w3.org>


[preliminary] This indeed looks on instant review as though it does meet the functional 
requirement we posed.

PF:  Anyone see a flaw in this reasoning?  Please respond by this time Friday if you 
see anything.


What is the baseline (draft) against which we should pose language-to-insert as edits?  
It would appear that there have been enough changes so that we wouldn't know how to 
say what needs to be said without being aware of the context in the section where it goes.

[It may go in the section on host languages which we understand to be rather new.]

[And it is not immediately obvious with this data model strength that this document
needs to say more.  The next stop may be the DOM.]


At 09:10 AM 2002-06-26, Steven Pemberton wrote:
>Thanks for the message. We were actually expecting text to include in our
>forthcoming CR spec, as agreed on March 1st at the Cannes FtF.
>However, let me address your comments here:
>> Users need to know, when they are submitting a form, what information they
>> are submitting.
>> The PF group has discussed several possible approaches:
>> 1. Requiring an explicit reference from each form control to the model
>> it is associated with. The principle is that this provides unambiguous
>> information in the DOM that assistive technologies (and simple adaptation
>> techniques) can access. See for example
>> http://lists.w3.org/Archives/Public/www-forms-editor/2002Feb/0110
>This requirement is already the case. If there is only one form, then the
>controls default to the only model. If there is more than one form, then the
>controls must explicitely reference the model.
>> 2. Requiring an explicit container for a form, as in HTML, that does not
>> allow other forms to be contained. The principle is that this makes it
>> easy to identify the things that are being submitted - not requiring real
>> processing. The original proposal was made (Member only - need permission
>> make it public) in the PF meeting of 19 June 2002 - see logs at
>> http://www.w3.org/2002/06/19-pf-irc#T17-03-27
>This is also already the case, only maybe it wasn't clear. The form is
>actually the model. It identifies explicitely what is to be submitted on the
><submission> element (was previously called <submitInfo>). If there are
>several forms, there are several models. If there are several different ways
>of submitting a form there are several <submission> elements. The <submit>
>form-control binds directly to a <submission> element in the model, so it is
>trivial to find out from there what is being submitted. What is not trivial
>is to identify what is 'hidden' in old-HTML Forms parlance (i.e. what hasn't
>been displayed to the user up to that point), and how to present it all to
>the user in an easy-to-understand way (since it is just XML data).
>> 3. Placing a conformance requirement on User Agents that they offer the
>> possibility of reviewing what information is being provided in a given
>> and to "confirm or cancel" the submission. See "doing more" point 3 at
>> http://www.w3.org/TR/UAAG10-TECHS/guidelines.html#tech-info-form-submit
>> See also (Member Only)
>> http://lists.w3.org/Archives/Member/w3c-wai-pf/2002JanMar/0254
>As I said, this is easy to do, though hard to format in a friendly way. So I
>think we fulfil all your requirements, and we only need text from you that
>satisfies what you think needs to be said.
>Best wishes,
Received on Wednesday, 26 June 2002 09:52:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:25:04 UTC