RE: 8.1.7 - range Element

-----Original Message-----
From: Jérôme Nègre [mailto:jerome.negre@e-xmlmedia.fr]
Sent: 03 September 2002 13:49
To: AndrewWatt2001@aol.com
Cc: xforms@yahoogroups.com; www-forms@w3.org; www-forms-editor@w3.org
Subject: Re: 8.1.7 - range Element



>> If that's the intended mechanism what's the purpose of an optional pair 
>> of "hint" attributes? Can the user agent ignore the "hint"?

>Since having a schema is optional, one needs a way to express the 
>bounds, hence the start and end attributes.

>> What happens if the two sources of information are inconsistent? Which 
>> value takes precedence?

>I would say that the final range is the intersection of both ranges. But 
>that's my 0.02 euro.

What if there is no intersection?

Section 4.3.16 xforms:revalidate event (for example) specifies that an instance must be validated first against any schema, and then against XForms constraints. 

So if the range set by the XForms constraint falls outside the range allowed by the schema, there will be an error, no matter how a forms engine rendered the range. The instance must be invalid against either the Schema, or the XForms constraint. In such circumstances surely the XForms form should fail to initialise (not least because the form processor cannot know how to render the range). 

Otherwise the schema constraint is applied first, and then the XForms one. So the forms engine might as well use the restriction of the schema given by the XForms constraint as its guide. 

I suppose there can only be an inconsistency if the XForms constraint specifies an invalid range.

However, what are the possibilities for having other XForms constraints which are not valid against a specified schema? How many of these would have no implication for how a form might be rendered? Given that an XForms author cannot be guaranteed to have access to every schema which may be applied to an XForms form, at what point would a form throw an exception?

(Apologies if I am simply showing how badly I have read the latest draft!)

All the best

Mark

Received on Tuesday, 3 September 2002 10:05:50 UTC