[wg11] EXPRESS Section 11 question

[Alan R Williams]

> Date: Mon, 24 May 2004 11:39:05 -0400
> From: Ed Barkmeyer <edbark at nist.gov>
> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) 
Gecko/20030624 Netscape/7.1 (ax)
> X-Accept-Language: en, fr, de, pdf, it, nl, sv, es, ru
> To: Alan R Williams <alanrw at cs.man.ac.uk>, SC4 WG11 <wg11 at steptools.com>
> Subject: Re: [wg11] EXPRESS Section 11 question
> X-Spam-Score: -4.9 (----)
> X-Scanner: exiscan for exim4 (http://duncanthrax.net/exiscan/) 
*1BSHXP-0007dE-MB*ugwjiX0khPQ*
> 
> Alan,
> 
> I wrote:
> >>It is my position that the text of Part 11 is slightly in error, and the 
> >>clause should be corrected to read:
> >>"An entity which is either explicitly REFERENCE'd or implicitly
> >>interfaced shall only be instantiated to play the role described by an
> >>attribute of an instantiation of an entity in **, or interfaced into, ** 
> >>the interfacing schema."
> 
> You wrote:
> > I think that only sort-of works.  If you had a set of mutually referencing 
> > instances of implicitly interfaced entities, then that would meet the above 
> > critera, but would probably be wrong.
> 
> Possibly.  But they wouldn't be implicitly interfaced unless at least 
> one of them is also referenced by an explicitly interfaced entity.

If we have two schemas:

SCHEMA marriage_model;

  ENTITY wife;
    spouse : husband;
  END_ENTITY;
  
  ENTITY husband;
    spouse : wife;
  WHERE
    mutually_married :
      wife.spouse :=: SELF;
  END_ENTITY;

END_SCHEMA;

SCHEMA children_model;

  REFERENCE FROM marriage_model (wife, husband);

  ENTITY child;
    mother : wife;
    father : husband;
  END_ENTITY;

END_SCHEMA;

It's an awful model :-)  Anyway, from the viewpoint of children_model we could 
legitimately have
  #1 = wife (#2);
  #2 = husband (#1);
  #3 = child (#1. #2);

IIUC your correction would also allow
  #4 = wife (#5);
  #5 = husband (#4);

and they would be legal as wife and husband are interfaced into the schema.  I 
think it needs something like:

"An entity which is either explicitly REFERENCE'd or implicitly
interfaced shall only be dependently instantiated.  A dependently instantiated 
entity instance must play the role described by either an
attribute of an instantiation of an entity in  
the interfacing schema or an attribute of another dependently instantiated 
entity instance."

The definition is recursive and so should work back up an attribute chain.

> I suppose there could be some other means of implicitly interfacing 
> them, e.g. by explicitly REFERENCEing a FUNCTION or RULE that involves 
> one of those data types.
> 
> > Just to check, you do think an s1.e1 would be OK if it was referenced by an 
> > s1.e2 that was referenced by an e3.  Yes?
> 
> Yes.

Good (or bad from my implementing colleague's point of view).

[snip]


> -Ed
> 
> -- 
> Edward J. Barkmeyer                        Email: edbark at nist.gov
> National Institute of Standards & Technology
> Manufacturing Systems Integration Division
> 100 Bureau Drive, Stop 8264                Tel: +1 301-975-3528
> Gaithersburg, MD 20899-8264                FAX: +1 301-975-4694
> 
> "The opinions expressed above do not reflect consensus of NIST,
>   and have not been reviewed by any Government authority."

Alan

Alan Williams, Room IT301, Department of Computer Science,
University of Manchester, Oxford Road, Manchester, M13 9PL, U.K.
Tel: +44 161 275 6270      Fax: +44 161 275 6280