onEvent inconsistency between spec & schema?

[Ron Ten-Hove <Ronald.Ten-Hove@Sun.COM>]

BPELers,

    This is just a query to see if an inconsistency we found is worth raising as an issue, or at least an editors' action item. The <onEvent> pseudo-schema reads:

     <onEvent partnerLink="NCName" portType="QName"?
         operation="NCName" (messageType="QName" | element="QName")?
         variable="NCName"?
         messageExchange="NCName"? >*

but the real schema reads:

<complexType name="tOnEvent">
    <complexContent>
        <extension base="bpws:tOnMsgCommon">
            <sequence>
                <element ref="bpws:scope"/>
            </sequence>
            <attribute name="messageType" type="QName" use="required"/>
            <attribute name="variable" type="NCName" use="required"/>
        </extension>
    </complexContent>
</complexType>

    The problem areas are in bold blue. The pseudo-schema introduces a new attribute, named element, as an alternative to messageType, and makes them optional. The actual schema makes no mention of an "element" attribute, and the messageType attribute is required.

    I recall that we discussed (on the list) such a change, in relation to issue 295, but I don't recall this being part of Mark Ford's resolution to 295. Regardless, the spec and the schema don't agree. Should we simply fix the schema (an editors' action item), or do we need to open an issue to allow the TC to decide how to change the specification to match the schema?

-Ron