'Re: "RE: Including schemata with duplicate referents"'

Hi,

Alessandro Triglia wrote:
[...]
> Note the use of the word "corresponding" here, which is the same word used above ("which in turn corresponds to a valid schema").
> Again, a schema is said to "correspond to" a <schema>.  This correspondence is specified under "Schema Representation Constraint: Import Constraints and Semantics" with regard to imported namespaces, and is specified under "Schema Representation Constraint: Inclusion Constraints and Semantics" with regard to inclusion.  The schema "corresponding" to a <schema> contains all the components of the imported schema.  Now, if the <schema> corresponding to that schema is **included** by another <schema>, the schema corresponding to the latter must include those components as well.
> Correct me if there is a flaw in my reasoning, but my interpretation of the Rec is that the import mechanism and the include mechanism are fully recursive with regard to what components become part of the schema.  
> (However, it is still mandatory to explicitly <import> or <include>  in order to be allowed to reference a component in a schema document.)

No objections. I hope my question was understood as how to actually
implement this behaviour in a schema processor as I'm currently not
sure what technique to use to avoid imports/includes of
identical components which are already referenced in the various
recursive schema construction steps. IOW, if including/importing I
need to add copies of components of the included/imported schema
to the including schema, but I must not include them if they
are already included; if those rejected components are already
referenced by components which _are_ included I don't know if
remapping the references to their included twins or leaving them
as 'not included but referenced', which seems memory wasting and feels
like inconsistent.

Regards,

Kasimier

Received on Friday, 5 November 2004 14:51:23 UTC