In article <zT25j.193237$Eg2.136955@f...>,
geoff <nospam@n...> wrote:

>The snippet below had 'po' in front of 'billTo', 'shipTo', 'order', etc. but
>they say by moving the 'po' namespace definition to the 'po:po' element, the
>child elements fall into that namespace:

That's not true, in particular this statement:

  ... the content of any namespace-prefixed element is considered to
  belong to the namespace of its parent element ...

But there's something like it that is true.  Namespaces are a notation
used to determine the interpretation of elements and attributes.  The
responsibility for doing that determination rests with the application
that's processing the file.  The obvious approach is to say that the
application interprets elements in some particular namespace (or group
of namespaces) as belonging to it, and treats others as uninterpreted
data.  But some applications treat no-namespace children of their
elements as also belonging to the application, even though they're not
in its namespace.

Presumably the applications that understand the "po" namespace in this
example use the second approach.

XML Schemas provides support for this way of doing things through the
elementFormDefault attribute.  If this is set to "unqualified" then
the schema can validate (by local element declarations) no-namespace
children of elements in the target namespace.

-- Richard
-- 
"Consideration shall be given to the need for as many as 32 characters
in some alphabets" - X3.4, 1963.