[regrep-cc-review] BCC/ACC Inheritance and Registry

["CHIUSANO, Joseph" <JCHIUSANO@lmi.org>]

Tim,

 

As the updates to the CC specification are near completion, I have been reviewing past UBL LCSC e-mails in an effort to identify any potential features that we can consider incorporating in the OASIS/ebXML Registry V3 architecture, as part of the Core Components Review effort.  This excerpt from the e-mail below caught my attention:

 

<Excerpt>

a) simply add a couple associations to fig 6-1: BCC->BCC called "based on" and ACC->ACC called "based on" -- that would give us the ability to derive ACC's from other ACC's and likewise for BCC's.  If we take Eve's CCT/RT analysis to heart, we'll need at least BCC inheritance soon.

<Excerpt>

 

I would like to please ask the following questions:

 

(1) I am not aware of any further action on the concept of BCC/ACC inheritance - can you please tell me if there has been some? 

 

(2) (assumes answer to #1 is "yes") Knowing what inheritance is in the object-oriented realm, what would it specifically mean in this sense?  That is, have any "rules" been applied to allow differentiation between inheritance and a completely new entity (BCC or ACC) - e.g. how much of the "base" entity must be retained in order for the new entity to be considered "inherited from" the base entity rather than a new entity, etc.

 

(3) (assumes answer to #1 is "yes") Do you feel that it would be valuable to incorporate this feature in the Registry architecture - i.e. the ability to associate BCC's with the BCC's they are "based on" (and same for ACC's)?  If so, would it naturally extend to BBIE's and ABIE's?

 

Thanks in advance,

Joe

**************************************************************************
  Joseph M. Chiusano
  Logistics Management Institute
  2000 Corporate Ridge
  McLean, VA 22102
  Email: jchiusano@lmi.org
  Tel: 571.633.7722
**************************************************************************

-----Original Message-----
From: Tim McGrath [mailto:tmcgrath@portcomm.com.au]
Sent: Monday, May 13, 2002 5:25 AM
To: ubl-lcsc@lists.oasis-open.org
Subject: [ubl-lcsc] [Fwd: Bill's comments for Mike dependencies paper.]

I am forwarded a message from Bill Burcham regarding the work Mike is doing on dependencies and business rules
 (hope you don't mind bill).

Can we build our rules as 'subtypes'?  Any thoughts?

-------- Original Message --------

<snipped>

 

Also I heard something about Mike making a proposal regarding "enhancing the metamodel" (perhaps adding associations and also adding subtyping through inheritance).  Is that right?  The association thing is the next level of sophistication beyond our "property" thing.  We talked a bit about subtyping -- but recommended expunging it for now since as it was applied (BCC->RT->CCT) it was too limited.  I've been thinking we'll have to add it back in soon.  A couple ways come to mind:

 

a) simply add a couple associations to fig 6-1: BCC->BCC called "based on" and ACC->ACC called "based on" -- that would give us the ability to derive ACC's from other ACC's and likewise for BCC's.  If we take Eve's CCT/RT analysis to heart, we'll need at least BCC inheritance soon.

 

b) don't add it to the metamodel at all, but handle it solely throught "context methodology" a.k.a. "magic"

 

Actually I think we'll need a combination of (a) and (b) eventually: you in LCSC will want to capture type-subtype relationships, and CM will want to make "contextualized components" subtypes of their "base components" (which might also actually be "contextualized" from some other "base").