Re: [Fwd: simple barenames for schema component designators] last call: "XML Schema: Component Designators" [abstractComponentRefs-37]

Dan Connolly writes:

>> I think the general principle of "keep the simple things simple" applies.

Are we perhaps seeing the following two desiderata somewhat in opposition?

* To XML schema, all types have equal standing, regardless of the 
namespace in which their names appear, namespaces of the base types from 
which they draw information, number of schema documents used to compose 
them, etc. From this point of view, there is architectural simplicity in 
having an identification mechanism that models that orthogonality. 

* Dan believes that for his users, the types they locally create are 
distinguished, if only by the likelihood that users will wish to refer to 
them.  Thus, for these users, the "simple case to keep simple" is to 
optimize for bare name reference to such user defined types, perhaps at 
the expense of a more orthogonal identification architecture for 
components in general.

I am neither defending nor criticizing the choices embodied in the current 
proposal from the schema workgroup, just suggesting that the above may be 
the trade-off.    Maybe we can get lucky and find a way to scale from 
using barenames to achieve the second goal, while retaining a more robust 
architecture to handle the general case. 

Noah

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Friday, 1 April 2005 22:09:31 UTC