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

RE: Foreign form controls -- ambiguity in XForms 1.0

From: Micah Dubinko <MDubinko@cardiff.com>
Date: Thu, 4 Apr 2002 15:15:12 -0800
Message-ID: <E840F0B7E6189547BDB91DA8BF2228AB28C3A6@csmail.cardiff.com>
To: "'www-forms-editor@w3.org'" <www-forms-editor@w3.org>
I withdraw this Last Call comment, as the XForms Working Group has agreed to
an XForms Extension module in the teleconference of 2 Apr 2002. This change
renders the below comment obsolete.

Thanks!

.micah

-----Original Message-----
From: Micah Dubinko 
Sent: Tuesday, February 05, 2002 6:05 PM
To: 'www-forms-editor@w3.org'
Subject: Foreign form controls -- ambiguity in XForms 1.0


Beyond signature controls, others have expressed the desire to incorporate
other form controls, tree controls for instance, into XForms. Other groups,
such as VoiceXML, also wish to build specific solutions from the general
framework of XForms. There's a need to express form controls outside of the
core set defined in XForms 1.0.

Note that (X)HTML allows the <object> tag as a data-submitting form control.
This has been rarely used, but the limited data representation of HTML form
data no doubt contributed to that.

The spec is not clear on 1) whether foreign form controls are currently
allowed, and 2) if they are, what processing (if any) is applies.

The spec does clearly state:
"Note that except where specifically allowed by the Schema for XForms,
foreign-namespaced elements are not allowed as content of elements in the
XForms namespace."

Form control elements, however, can exist as content of the containing
document, not necessarily of an XForms element

<html:body>
  <xforms:input.../>
  <otherns:treecontrol../> ?
...
and thus are not covered by the above statement.

To fix this, the XForms spec should be changed in one of the following ways:

1) Clearly state that foreign form controls are not allowed.

2) Clearly state that foreign form controls are allowed (and make sure
##other is in the content models of <group>, <case>, <repeat>, etc..) and
specify concrete processing.
[note that the processing may very well be quite minimal, as is <object>
processing in HTML forms]

Given the opening paragraph, I would strongly favor the 2nd option. :-)

Thanks,

.micah
Received on Thursday, 4 April 2002 18:19:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 10 June 2009 18:12:11 GMT