Re: XForms CR - 7.10.1 Another XPath 1.0 Conformance Question

In a message dated 18/11/2002 20:09:15 GMT Standard Time, 
Ryan_Tomayko@stercomm.com writes:


> Also, I personally don't like the term "copy" (as in "copy documents")
> because if a URI is specified instead of an inline document, the term "copy
> document" becomes misleading.

I think the term "copy document" is misleading either way. It is, if I 
understand this correctly, supposed to be a copy XPath data model. 

 It might be easier to imagine this as their
> 
> simply being separate instance documents for each <xf:instance> element and
> each of them may be initialized with either data from the source document
> (inline) or from an external location (uri). 
> 
> Another possibly misleading point might be the misconception that using
> <xf:instance/> element is equivalent to XIncluding an external resource 
> into
> the source document. My interpretation is that this is not the case.

If that interpretation is correct then do you also expect that there is no 
way, using the XForms CR, to access (say for comparison purposes) the 
original data?

If I were trying to access such an external document using XSLT 1.0 I would 
use the document() function. As far as I am aware XForms as currently drafted 
currently doesn't have that functionality. Is that correct, as you see it?

 The
> 
> instance element simply tells the processor where to get initial data for a
> certain instance document, it does not modify the source document in any
> way. 

I didn't say otherwise.

The XForms CR seems to use the term "instance document" for another subtly 
different concept - the serialized XML document which is sent by an XForms 
Processor to a server or other application back end. So, it seems to me that 
we should avoid using "instance document" for several concepts, otherwise the 
current confusion of terms will get even worse.

The reason I bring this up is that I can see where this might lead to
> 
> the idea of "copies" of instance documents. i.e. The XForms Processor runs,
> expanding external instance documents into the source document and then
> makes "copies" of each instance document. I don't conceptualize it this way
> at all. I view it as there being multiple instance documents that are
> completely separate from the source document and they may be initialized
> with XML specified inline or located externally.
> 

Well, I don't view there being multiple "instance documents" at all. I 
visualise there being multiple XPath data models (at least that is what John 
Boyer's post of a couple of days ago seemed to me to say.).

Andrew Watt