W3C home > Mailing lists > Public > www-forms@w3.org > June 2005

RE: [Fwd: Re: XForms: Question about tax form example]

From: Mark Birbeck <mark.birbeck@x-port.net>
Date: Thu, 23 Jun 2005 20:23:09 +0100
Message-ID: <565B15FF-9DFA-418E-94CE-B678D5E1FF96@S009>
To: "'Klotz, Leigh'" <Leigh.Klotz@xerox.com>
Cc: "'Aaron Reed'" <aaronr@us.ibm.com>, "'Erik Bruchez'" <erik@bruchez.org>, <www-forms@w3.org>

Hi Leigh,

Hope all is well with you.

> I lobbied for allowing <label> as a first child of <group> 
> but couldn't get it allowed.

xf:label *is* allowed as the first child of xf:group according to the
schema, so perhaps your lobbying paid off. (However, the schema allows zero
or more labels...but that's another topic!)

> Also, the WAI group (and TV Raman on the Xforms WG) were 
> quite firm that label needs to be a child element of the 
> control it labels.

Yep...absolutely. And that has been re-stated in a number of recent threads,
including one running on the internal list.

> Stylistically, it would probably have been better to define 
> the label element in the XML Schema as a child element of 
> form controls rather than a top level element...

Absolutely! You are exactly right that the schemas were not that well
designed (a feature of many W3C schemas, unfortunately). The new schemas
take the approach that only elements that can exist at the top level are
defined to be top level. The schemas should therefore have very few top
level elements, since you only need xf:model, the form controls, and the
grouping constructs (xf:group, xf:repeat and xf:switch).

Also -- as I'm sure you are aware -- the other weakness of schemas that have
every element available as a top-level one, is that authoring tools then
incorrectly show all elements as being available, making it very awkward for
page designers. To get round this, tool designers often have to reconstruct
the schemas, which is a daft imposition, since it means that different
authoring tools could generate different XForms documents.

> ... and in my 
> opinion there is still room to do that in XForms 1.0 if the 
> WG believes that it is simply a schema authoring error (typo).

The schemas are being looked at again, in the light of XHTML 1.1
modularisation, and of course XForms 1.1, so there is a possibility that
they could be retrofitted. However, anyone developing an XForms processor
today would be better off implementing XForms 1.1.

Certainly the schemas we use for Sidewinder are the 'fixed' ones.

> I believe that the Mozilla implementation that allows label 
> anywhere is incorrect.

Yes. However, in reference to Erik's point that you replied to, it's
difficult to prevent, if you are not validating. It's quite an overhead to
check the structure of the document without a schema validation phase. So
whilst formsPlayer will 'technically' allow an xf:label anywhere -- just
like Mozilla -- if you use the Sidewinder Viewer as your web application
viewer, then xf:labels in the wrong place won't get through (because we
validate for XHTML+XForms+SVG+MathML). To put it a different way, it's not a
'feature', in formsPlayer, but a rapidly disappearing loop-hole!



Mark Birbeck
x-port.net Ltd.

e: Mark.Birbeck@x-port.net
t: +44 (0) 20 7689 9232
w: http://www.formsPlayer.com/
b: http://internet-apps.blogspot.com/

Download our XForms processor from
Received on Thursday, 23 June 2005 19:24:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:36:15 UTC