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

Re: XForms Reusability / Modularization (was RE: XForms WD 200208 21 - 2.1 XForms and XHTML etc)

From: joern turner <joern.turner@web.de>
Date: Sat, 31 Aug 2002 01:54:32 +0200
Message-ID: <3D7005B8.6020506@web.de>
To: "Tomayko, Ryan" <Ryan_Tomayko@stercomm.com>
CC: www-forms@w3.org

first, sorry if i've pointed out things that already have been clear to 
you, but i got the impression (please correct me) that the idea of 'one 
form for every platform' has already been given up or is considered hard 
or not to achieve with the current spec.

i absolutely agree with the thoughts about model usage - this is 
platform neutral and therefore is not problematic.

but, to take your example below, what happens if you've to change the 
order or add/delete UI controls ? does this mean that you've to change 
all four UI versions of the form ?

i think this avoidable if one tweaks interpretation of the current 
official proposals a bit: if one interprets XForms UI as a meta language 
you can provide a mapping between XForms and every platform you want to 
support. this mapping has to be done only once for each platform. so, 
what finally is left to do is to write a layout for each platform. as 
you've correctly pointed out this is a logical requirement if you 'want 
full control over the presentation'.

so, analogous to your statement about models i would say: it is not 
necessary to change all four files if you want to change the logic of 
the UI.


Tomayko, Ryan wrote:
>>>1. In order to create a reusable XForms document, the document could 
>>>assume no knowledge of the host language (profile). i.e. The XForms 
>>>document could not contain elements or attributes from the host 
>>>languages namespace.
>>i agree with your first sentence here, but disagree with the second: 
>>there's no need for a form-processor to deal with the host-language at 
>>all (see above). that's why namespaces are so cool - why not ignore what 
>>you're not interested in ?
> Yes, definitely. I should have been more clear on my statement. The
> implementation I'm working on supports HTML+XForms and could support
> WML+XForms very easily. The XForms Processor ignores all non XForms markup.
> I was speaking more toward building cross-profile *documents*, not
> cross-profile processors. There is no way to build a single generic XForms
> document and have it run on either a web browser or a wireless device. You
> have to author two documents - one for each host language. The spec doesn't
> attempt to accommodate profile independent XForms (that's a statement not a
> complaint). 
> XForms presentation needs a host language to make it meaningful but XForms
> data/logic markup (xforms:model) is meaningful without any host language at
> all. The difference is that operations on the data layer are always the same
> regardless of host language but operations on presentation are dependant on
> host language.
> Let's say I have four documents that do the same thing for different
> devices: a HTML+XForms, WML+XForms, SVG+XForms, and a HDML+XForms document.
> I need all four versions because I want full control over each presentation
> language. However, when I want to change a validation rule, I need to make
> that change in all four documents. This is unnecessary. There's no reason
> for the model definition to be specified in all four documents.
> The reason I haven't posted my concern to the www-forms-editor list is
> because XInclude will do the job today.
> <html>
> 	<head>
> 		<xi:include src="shared-model.xml"/>
> 	</head>
> 	...
> </html>
> - Ryan
Received on Friday, 30 August 2002 19:55:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:37:44 UTC