I vote for adding namespace="##other" for consistency sake.
We don't yet have concrete use cases for carrying V1/V2 types as root
elements in extesions.
Regards,
Subbu
Rich Thompson wrote:
>
> I was updating the v2 schema to implement the decision to not extend the
> v1 schema and hit into a validation error (used the W3C schema validator
> against this schema for the first time). The error is that EventPayload
> is non-deterministic because the any allows for elements from the wsrp
> v2 types namespace. Solutions include dropping the NamedStringArray
> reference or restricting the any to ##other (like we do for the Property
> type). I recommend dropping the NamedStringArray element as I could
> imagine use cases for various other types we have defined, but before we
> have that discussion, are there any other options people see?
>
> for your reference, here is the current definition of EventPayload:
>
> <complexType name="EventPayload">
> <sequence>
> <element name="namedStringArray" type="types:NamedStringArray"
> minOccurs="0"/>
> <any minOccurs="0" maxOccurs="unbounded"/>
> </sequence>
> </complexType>
> <element name="EventPayload" type="types:EventPayload"/>
>
>
> Rich
