Re: [LC] 11.8 The header Element (PR#80)

Hi Steven,

Thank you very much for considering this issue.  In this case, the working group
did not implement a change based on the feedback, and I will now endeavor to
explain why.

The main reason is that the WG felt that XPath-based name and value attributes
made it too difficult for the case of specifying a literal, which is easier to
understand when you recognize that the nodeset attribute is optional.

Basically, the header element was created *and then* nodeset was added to
simplify the specification of multiple headers from instance data.

So the question might then become, why can't we add name and value as literal
attributes at least.  Two reasons.  First, this variant is not useful when
nodeset is used.  Second, when there is no nodeset, it does indeed seem less
verbose to specify a header with attributes.  The problem we encountered there
was that the value attribute would then be a literal string, which is
inconsistent with its datatype of XPathExpr everywhere else in XForms.  And if
we make it an XPath, then again we would be in the trap of authors having
frustration when they forget to put the extra quotes needed to express a string
literal in an XPath.  So, in this case, the WG felt it was best to have only the
child elements name and value, both of which offer a literal content version and
a value attribute for specifying an XPath.

Please let us know if you have any further concerns about this issue.

Thank you,
John Boyer

> 
> http://www.w3.org/TR/xforms11/#submit-header
> 
> This seems over-engineered. Why not instead of
> 
>       <header nodeset="...">
> 	     <name value=".../>
> 	     <value value="..."/>
>       </header
> 
> use the simpler:
> 
>      <header nodeset="..." name="..." value="..."/>
> 
> 
> 

Received on Thursday, 26 July 2007 18:11:27 UTC