Re: Issue 252 (reopen)

Glen Daniels wrote:

>>There is yet a third way of accomplishing the same goal, by 
>>keeping the annotation component but using an attribute part 
>>of the annotation. 
>Hm.  I think you are still operating under a slight misconception.  When
>you use an extension attribute a la:
><xs:element name="picture" type="myNS:image"
>            xmlmime:expectedMediaType="image/*"/>
>That ***IS*** an annotation.  The purpose of the "attributes" part of
>the annotation schema component is precisely so that attributes from
>other namespaces have a place to appear in the component model.  So the
>element component would look something like:
Hey Glen,

I did not say the extension attribute is not an  "annotation" in my 
message. We are not disagreeing there. Lets separate the discussion of 
an annotation "element" from the annotation "component". It caused 
enough hassle on long distance telephone lines during the last f2f.

I am saying something different:  It appears that "syntactically" there 
is a third option per 3.13.2: The annotation "element" may contain 
attributes from other namespaces. This capability is also reflected in 
the schema of schemas [1] that is normative.

If my understanding serves me right, this third option is reflected the 
same way in the component model as you describe below:

>element {
>  name : "picture"
>  type : "myNS:image"
>  annotation : {
>    documentation : {}
>    appInfo : {}
>    attributes : [
>      xmlmime:expectedMediaType : "image/*"
>    ]
>  }
 From the component model perspective, it seems to me that extension 
attribute (syntactically) and the extension attribute within an 
annotation element (syntactically)  are "equivalent". Perhaps, it is 
somewhat convoluted, but this is schema. (Sorry ;-)) So, the result is 
the same, but the means is different.

>Note that the "documentation" and "appInfo" parts of the annotation
>component are empty.  Please reread the example in the schema spec [1],
>specifically the paragaph right before section 3.13.3 (along with the
>example) for more on this.
>Once this becomes clear, I think further consideration of the issue,
>except perhaps as a result of feedback from soapbuiders, becomes
Well, technically this *IS* new information as this option was not 
considered, isn't it? I trust the wg can decide what to do with it.

BTW, thanks again for sending the message and soliciting the feedback 
from soapbuilders! That addresses my concern of exploring how much 
support is out there on this topic. Lets see what we find out.





Umit Yalcinalp                                  
Consulting Member of Technical Staff
Phone: +1 650 607 6154                          

Received on Saturday, 11 September 2004 00:22:20 UTC