P14: questions

[Peter Denno]

Hi Jochen,

On Thursday 23 May 2002 09:06 am, Jochen Haenisch wrote:
> Dear WG11 experts,
>
> focus has recently been on P11e2 - maybe I can get some attention for P14
> ?!?
>
> We have a couple of questions:
> 1) 9.10 Rule declaration:
> Rules may be defined for use within a schema_view. What is the intention of
> this feature? Shall these rules control the population of the view_entities
> in the schema? Or shall they (also) have an impact on the referenced
> original population? The latter does in our opinion not make too much
> sense, but maybe there are examples that we have not thought of?!?
> Similar for schema_map.

Rules do not control anything. As Clause 9.10 says:
Inclusion of rule declarations has no effect on the execution model nor on the 
contents of extents.

The motivation, I think, was that by including rules, an Express-X schema 
could be 'transformed' (see the annex) into an Express schema.

Perhaps Martin Hardwick can provide more detail on this. I do not use this 
feature.

>
> 2) Reference to populations:
> Is the only way of identifying populations the FROM-clause (binding
> extent)? 

There is an EXTENT function.

>What if one wants to change a single population? Then source and
> target schema are the same. In such a case it is necessary to have names
> for source and target populations, not only for the schemas.
> Or is this out of scope for P14?



1) You can use a schema_view.

2) In a schema_map, the REFERENCE clause can declare an alias, which can be 
used to qualify identifiers (to indicate their schema and population).

REFERENCE FROM src : schema_a AS SOURCE;
REFERENCE FROM trg : schema_a AS TARGET;

MAP my_map AS tp : trg.person;
 FROM sp : src.person;
 SELECT
   tp.name := sp.name;
END_MAP;

However, what I just demonstrated isn't exactly in scope for Express-X. The 
scope statement includes the following:

"mapping of data governed by one version of an EXPRESS schema to data governed 
by another version of that EXPRESS schema, where the two schemas have 
different names;"

The reason for the above is that Express-X is about declaring relationships 
between schema (as another item of the scope statement asserts) . It doesn't 
make sense declare relationships between a schema and itself. 

There is no rule on the reference clause stating that schema shall have 
different names. I think this is going to be a place where tools (tools that 
want to do more than declare relationships between schema)  are going to 
ignore the stipulation in the scope statement. After all, moving data from 
one population to another isn't within scope either!


-- 
Best Regards, 
- Peter 

Peter Denno 
National Institute of Standards and Technology, 
Manufacturing System Integration Division, 
100 Bureau Drive, Mail Stop 8260             Tel: +1 301-975-3595 
Gaithersburg, MD, USA 20899-8260          FAX: +1 301-975-4694