Re: XForms CR - 6.1.3/6.1.4 - Catch 22 Interaction?

In a message dated 13/11/2002 17:30:55 GMT Standard Time, 
roland_merrick@uk.ibm.com writes:


> Greetings, what you describe is the intention. e.g. I will only insist on 
> knowing the name of you first born if you have indicated that you have one. 
> 
> 
> i.e. <firstBornName> is specified as required. 
> 
> however, it also has a relevant clause which is dependent on another 
> element e.g. <hasChild/ >. 
> 
>     <firstBornName/> 
>     <hasChild/> 
> 
> <bind nodeset="firstBornName" required="true" relevant="hasChild = 1" />
> 
> If <hasChild> is false, <firstBornName> is irrelevant and is not submitted. 
> 
> 
> Regards, Roland 
> 
> 
> AndrewWatt2001@aol.com 
> Sent by: www-forms-request@w3.org 13/11/2002 15:52 
> 
> 
>         
>         To:        www-forms@w3.org, www-forms-editor@w3.org, 
> Xforms@yahoogroups.com 
>         cc:         
>         Subject:        XForms CR - 6.1.3/6.1.4 - Catch 22 Interaction? 
> 
>        
> 
> 
> 
> 
> It seems to me that there is a potential problem for the user of a form if 
> it
> is possible for required="true()" and relevant="false()" to co-exist. In 
> that
> scenario the table at the end of 6.1.4 indicates that the required by
> non-relevant form control would likely be inaccessible to the user (unless 
> he
> had diagnostic powers which it is probably unreasonable to expect).
> 
> As I read the CR it would not be possible to submit the form because
> required="true()" and the data hasn't been entered and it might not be
> possible to access the required form control because relevant="false()".
> 
> Hopefully such a situation would occur rarely, but wouldn't it be better to
> ensure that it cannot happen? It seems to me that it might be better if 
> there
> were a rule which indicated that a piece of data which is not relevant 
> cannot
> be required.
> 
> Andrew Watt 
> 

Roland,

The first part of 6.1.3 reads, "Description: describes whether a value is 
required before the instance data is submitted.".

Your comment seems to indicate that it was intended to read something 
different.

I suggest that some redrafting is needed.

Andrew Watt