Re: essential test cases?

Dan Connolly wrote:

> While mulling over the recent proposal[2], I realized
> that the namespaces spec (ironically) doesn't currently
> specify any interactions between multiple documents, and as long
> as that remains its scope, I think the proposed wording
> is adequate.

I don't think so. You can get the two bats problem even within a single
document because of external entities.  Consider a document
http://www.example.com/catalog/catalog.xml

<!DOCTYPE catalog [
<!ENTITY bats "animals/bats.xml">
]>
<cat:catalog xmlns:cat="../2000/vocab#">
&bats;
</cat:catalog>

where http://www.example.com/catalog/animals/bats.xml

<cat:item
  xmlns:animal="../2000/vocab#"
  animal:type="bat"
  cat:type="available-for-special-order"/>

A namespace processor has to decide whether this violates the XML
constraint on attribute uniqueness or not. On the literal
interpretation it does because animal:type and cat:type have the same
local part and prefixes that have been bound to identical namespace
names (although the namespace names refer to different URIs).

I don't believe the Microsoft proposal has taken this sort of case into
consideration.

James