Title: Message
Tony,
Your
interpretation is still correct: isReplacedBy points from the "old" entity to
the "new" one. I did not suggest to add even another tModel that does the same
vice versa.
What I
suggested is that we grandfather the categorization of isReplacedBy although we
now know better (it should be a category system). If we decide otherwise, i.e.
saying that it is correctly categorized as an identifier system, we are running
into other problems, including the categorization of the owningBusiness tModel,
which is then wrong (it IS a category system but SHOULD then also be an
identifier system).
Claus
-----Original Message-----
From: Rogers, Tony
[mailto:]
Sent: Donnerstag, 6. M酺z 2003
01:14
To: Von Riegen, Claus; Andrew Hately;
Subject: RE: [uddi-spec] Update to
CR010 for identifier definition
Actually, it suggests that we should have a tModel called, perhaps,
"Replaces" - this would be put on a new tModel to identify it as one which
replaces an old tModel, and give the old tModel's key as its
value.
This
assumes that I haven't misunderstood "isReplacedBy" - if it is intended to label
the new tModel with the key of the old one it replaces, then I have to say that
the name is confusing. "isReplacedBy" sounds like it is added to the old tModel,
specifying the key of the new tModel that this old one
isReplacedBy.
Either
way, I have no problem with this being an identifier (now that we have the
cardinality resolved), because we're identifying one tModel in terms of another
one to which it is related by way of replacement.
Tony
Rogers
-----Original Message-----
From: Von
Riegen, Claus [mailto:]
Sent: Wednesday, 5
March 2003 23:54
To: 'Andrew Hately';
Subject: RE: [uddi-spec] Update to
CR010 for identifier definition
Andrew,
Thanks, I very much like the new glossary definitions. They certainly
clarifies UDDI's meaning of identifiers and categories.
Although I clearly don't want to restart the whole discussion again,
the new definition still does not allow isReplacedBy to be an identifier
system.
Since the core definition of an identifier system
remains the same as "Any value set intended to be used to
identify the entities
in which it is referenced.", there is strictly
speaking no possibility for an entity to be identified with another
entity's identifier. In other words, it is not the question whether a value
from a given value set can represent an identifier or not, it is a
question in which context and to which type of entity it is applied to (this
is part of the corresponding value set's
definition).
Taking the isReplacedBy example, we have an entity X and a
separate entity X1, which logically replaces X. The entity key of X1 can
certainly be used to identify X1 since it is actually X1's unique identifier
within the given registry. But when X1's key is referenced in entity X, it
still idenfies X1 - and not X. This is the question of the context I mentioned
above: one can certainly use entities' identifiers and
reference them in another entity, but this reference would be
placed in a categoryBag rather than in an
identifierBag.
Please note that I can live with the current situation, that is,
grandfathering the isReplacedBy tModel's categorization as an identifier
system, although we now know better. Another option would be to relax the
glossary definitions even further by saying, for example, "Any value set
intended to be used to identify entities
[Ed. note: X1 from the example] in entities [Ed. note: X from the example] in which it is
referenced." But my gut feeling is that this somehow even complicates the
interpretation - and we would also have to update the sections
where identifierBag and categoryBag are introduced in chapter
3.
Claus
-----Original Message-----
From: Andrew Hately
[mailto:]
Sent: Dienstag, 4. M酺z 2003
22:45
To:
Subject:
[uddi-spec] Update to CR010 for identifier
definition
Please
review the attached update.
Andrew Hately
IBM Austin
UDDI Development, Emerging
Technologies
Lotus Notes: Andrew Hately/Austin/IBM@IBMUS
Internet:
(512) 838-2866, t/l
678-2866
