[GRAAP-WG] telecon on 3/14

[Karl Czajkowski]

On Mar 14, Jim Pruyne modulated:

> And, finally, after all of the confusion, the good news.  I believe I have the final draft of WS-Agreement.  It can be found at:
> 
> http://forge.gridforum.org/sf/docman/do/downloadDocument/projects.graap-wg/docman.root.current_drafts/doc6091/37
> 
> (sorry for the long URL).  If we have no complaints during this telecon, I'll submit it right afterward.
> 
> Best,
> 

One nit...

The grouping in the psuedo-schema for the Terms document in Section
4.2.2 seems a little off to me, i.e. not expressing the actual syntax
from the schema in the appendix which I think accurately captures what
we've always been doing.

With the current nesting and normal regular expression precedence, it
says that the outermost wsag:All can contain exactly one of the
compositors OR a sequence of basic terms, rather than a sequence of
compositors and or basic terms.  I think the opening brace should move
"up" to include the sequence of compositor children within the same
choice block that is already getting repeated with the "*" operator,
but the operator needs to be changed to "+" for one-or-more semantics:

<wsag:Terms>
    <wsag:All>
        {
            <wsag:All>
                wsag:TermCompositorType
            </wsag:All>       | 
            <wsag:OneOrMore> 
                wsag:TermCompositorType
            </wsag:OneOrMore> |
            <wsag:ExactlyOne>
                wsag:TermCompositorType
            </wsag:ExactlyOne> |
            <wsag:ServiceDescriptionTerm>
                wsag:ServiceDescriptionTermType
            </wsag:ServiceDescriptionTerm> |
            <wsag:ServiceReference>
                wsag:ServiceReferenceType
            </wsag:ServiceReference> |
            <wsag:ServiceProperties>
                wsag:ServicePropertiesType
            </wsag:ServiceProperties> |
            <wsag:GuaranteeTerm>
                wsag:GuaranteeTermType
            </wsag:GuaranteeTerm>     
        } +
    </wsag:All>
</wsag:Terms>

This corresponds exactly to the <xs:choice maxOccurs="unbounded">
(defaulting to minOccurs=1) that is in the schema.


-- 
Karl Czajkowski
karlcz at univa.com