Garret,

my .02

1. Indicator. Type. Both OAGi and ATG2 have translated the CCT
Indicator. Type into the xsd:boolean built-in type. Furthermore,
the list of valid values has been reduced to "false" or "true"
(other values allowed by xsd:boolean are "TRUE", "FALSE", "0",
"1"). The reasoning is that since each of these sets of values
does not change the semantics of what is being communicated,
there should not be multiple representations for the same
semantic concept. The pair "true", "false" was selected because
it was felt that it best conformed to existing usage of
booleans. It has been argued that "0", "1" would be a more
international representation since it is not language specific.
What do you think?

NOTE: It has also been argued that Indicator. Type should be
anything with two, and only two valid values which would prevent
it from being defined as xsd:boolean. I'd like to set aside that
argument for now and assume we are only dealing with xsd:boolean.


<AJS>"0" and "1" are certainly more international that "true" and "false"
and if we are going to restrict the standard it should be to "0" and
"1".</AJS>

2. Identifier Lists. The ATG2 NDR has introduced the concept of
"Identifier Lists". Line 530 states "In those cases where
identifier schemes mimic the functionality of code lists,
reusable identifier lists schema modules will be created to
minimize the impact of identifier list changes on root and other
reusable schema. Each reusable identifier list schema module
will contain enumeration values for codes and code values".
Should OAGIS implement Identifier Lists? If so, what rules or
guidelines should apply in order to determine if something is an
Identifier List or not?

3. Code vs Identifier. I put this one last so you wouldn't
immediately delete this email. Do we now have some general
consensus about how to distinguish codes from identifiers and
vice versa? Has anyone taken all of the material written to date
and summarized them into a clear set of rules/guidelines?

<AJS>2 and 3 seem to go together and get into the same old discussion
between codes and identifiers. I don't think anyone will ever agree on
what is a code and what is an identifier, so opening the discussion again
does not seem to be the right thing to do. I have disagreed since day 1,
and have been out voted each time -- I have given up on this. The concept
of identifier lists is just another code list/identifier discussion. If an
identifier needs an identifier list, shouldn't it just be a code list?
This is a never ending battle :-( </AJS>


Regards,

Alan Stitzer
AVP, Mgr Global Standards
Marsh USA, Inc
253 Coconut Pt
Jupiter, FL 33458
USA

office: 561 743 1938
mobile: 561 371 5339
fax: 561 743 1993



garret.minakawa@o
racle.com@Interne
t To
OAGI-CoreComponents@yahoogroups.com
11/03/04 01:53 PM @Internet
cc

Subject
[OAGI-CoreComponents] FEEDBACK
REQUESTED: Misc CC Implementation
Issues










1. Indicator. Type. Both OAGi and ATG2 have translated the CCT
Indicator. Type into the xsd:boolean built-in type. Furthermore,
the list of valid values has been reduced to "false" or "true"
(other values allowed by xsd:boolean are "TRUE", "FALSE", "0",
"1"). The reasoning is that since each of these sets of values
does not change the semantics of what is being communicated,
there should not be multiple representations for the same
semantic concept. The pair "true", "false" was selected because
it was felt that it best conformed to existing usage of
booleans. It has been argued that "0", "1" would be a more
international representation since it is not language specific.
What do you think?

NOTE: It has also been argued that Indicator. Type should be
anything with two, and only two valid values which would prevent
it from being defined as xsd:boolean. I'd like to set aside that
argument for now and assume we are only dealing with xsd:boolean.

2. Identifier Lists. The ATG2 NDR has introduced the concept of
"Identifier Lists". Line 530 states "In those cases where
identifier schemes mimic the functionality of code lists,
reusable identifier lists schema modules will be created to
minimize the impact of identifier list changes on root and other
reusable schema. Each reusable identifier list schema module
will contain enumeration values for codes and code values".
Should OAGIS implement Identifier Lists? If so, what rules or
guidelines should apply in order to determine if something is an
Identifier List or not?

3. Code vs Identifier. I put this one last so you wouldn't
immediately delete this email. Do we now have some general
consensus about how to distinguish codes from identifiers and
vice versa? Has anyone taken all of the material written to date
and summarized them into a clear set of rules/guidelines?

For example, I think there is general agreement that Codes are
"relatively" static and are "usually" maintained by some
authorized agency. They are a shortened/encoded/abbreviated
representation of some "thing", normally text, and usually of the
same consistent length within a code list. If we can capture
these attributes/characteristics as a set of rule, guidelines, or
some kind of checklist, I'm sure it would help a lot of people
trying to implement CCTS.

Thanks,

garret


[attachment "C.htm" deleted by Alan Stitzer/NYC-NY/US/Marsh/MMC]
[attachment "Garret.Minakawa.vcf" deleted by Alan
Stitzer/NYC-NY/US/Marsh/MMC]