Dan,
Actually, I understand that IBM actually owns a large number of DUNS
numbers for different parts of the enterprise. They are all "IBM" (unless
some are "Lotus", "Tivoli", etc.).
Naming an MSH is a whole other ball of wax. An MSH is not a pile of
circuits nor some number of lines of code. An MSH is really a data
structure that supports some part of the enterprise that is behind it. The
same circuits and lines of code might implement many MSHs. It isn't obvious
to me that any of the identifiers, endpoint addresses, or names that we
have been discussing is appropriate for labelling a single MSH. A single
PartyId might be served by many MSHs, so PartyId doesn't sound like the
right answer. In other cases, a single MSH might serve multiple PartyIds.
The MSG team would have to define what constitutes an MSH. Before defining
what constitutes an MSH, the MSG team would have to decide what purpose is
served by identifying an MSH since that may determine the nature of the
identifier.
One possibility (jumping over the requirements discussion that should
preced this statement) is that an MSH is identified by the concatenation of
the PartyId and CPAId.
Regards,
Marty
*************************************************************************************
Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown Hts, NY 10598
914-784-7287; IBM tie line 863-7287
Notes address: Martin W Sachs/Watson/IBM
Internet address: mwsachs @ us.ibm.com
*************************************************************************************
Dan Weinreb <dlw@exceloncorp.com> on 10/25/2001 01:45:10 PM
Please respond to "Dan Weinreb" <dlw@exceloncorp.com>
To: Jzheng@vitria.com
cc: Martin W Sachs/Watson/IBM@IBMUS, dmoberg@cyclonecommerce.com,
ebxml-cppa@lists.oasis-open.org
Subject: Re: [ebxml-cppa] Proposed schema changes, plus illustrative
examp le
Date: Thu, 25 Oct 2001 10:04:22 -0700
From: Jean Zheng <Jzheng@vitria.com>
Actually, in a case like this, don't you really want to know which MSH
the exceptions were from? Marty has explained to me that a PartyId
doesn't actually identity an MSH; there might be 10,000 MSH's within
the IBM corporation, but there might be just be one PartyId, for
"IBM". For the purposes you're discussing here, I would think that
what's needed is a way to name a particular MSH, not just the name of
the corporate entity that owns the MSH.
(Marty, if I am in any way misrepresenting your position, please correct
me.)
-- Dan
However, I'm afraid I didn't make a clearer statement regarding
PartyName.
I was not implying that all information that could be in PartyRef must
be
duplicated in CPP/A. I simply want a human readable description of
PartyID
to accompany PartyID. And I think that could be useful. For example,
to
response to a request from Party A, Party B can either send a Response
or a
Exception or a Signal. As Party A, if I want to see a list of
Exceptions I
have received in the past hour, which one of the following list will
seem a
better option?
List1:
Party2321212
Party4545644
Party91018-3
or
List2:
Compaq
Mom-Pop-at-Corner
Dell
Again, from the technical point of view, it doesn't make any difference.
Because underneath the display, record is probably all identified by
PartyID. It just seems nicer to be able to display list2. ebXML is,
after
all, meant for Business-To-Business applications, which implies,
hopefully,
the large portion of the user will be Business analysts rather than IT
people.
> MWS: Since PartyRef contains the PartyName, there is no reason to add
Can I make that assumption that PartyRef always contains PartyName?
PartyRef points to a Dtd or Schema that can contain any number of
elements.
It is not clear to me that it will definitely contains PartyName. Maybe
I
should make that assumption? If so, shouldn't we make that a mandate
for
the Dtd/Schema that PartyRef points to?
I do realize that having PartyName element or not won't make or break
our
functionality. But I just want to clarify the intention of this
request.
Thanks for bearing with me! :)
Jean
>