RE: Namespaces, schemas, Simon's filters.

["David E. Cleary" <davec@progress.com>]

> [1]	Is there a sufficent consensus that we could make the default
> non-use of local elements an official "best practice"?

I think there probably is, but xmlschema-dev should also be pinged on it.
There may be some points of views over there that can add to the discussion.

> [2]	Is there any consensus on when - if ever - use of local elements
> *would* be "best practice"?

I think a case can be made when the local elements are specifically related
to the context of the containing element. This matches the patern of
unqualified attributes that is commonly used today. Not only can these local
elements not be of use standing alone, they are more explicitly tied to the
element that contains them. You can almost think of them as structured
attributes. Having them being local reinforces the notion that they must be
treated/processed as a group.

David Cleary