Re: [Fwd: [Fwd: RE: [ubl-lcsc] [roe] OAG comment: CountryIdentificationCodecan be a data type of Country Code]]

[Garret Minakawa <garret.minakawa@oracle.com>]

While I agree that Fred's solution better fits the CCTS model, I think
there's a problem when you implement this solution in XML Schema. 
Specifically, it is related to the use of derivation by extension or derivation
by restriction.

I see Fred's solution as a derivation by restriction model which is technically fine according to XML Schema.  The problem is when derivation by restriction is actually implemented, it breaks the inheritance relationship between elements/types based on other types.  Therefore, changes to a "base" type will not be inherited by an element/type based on the "base" type if derivation by restriction is implemented.  This is a fairly well known issue with XML Schema and I'm sure lots of other people have already explained this problem much better than I ever could so I'm not going to get into any additional detail.

Many organizations (including OAGI) prohibit or discourage derivation by restriction because of this limitation regarding inheritance.  I did not see anything in UBL NDR that specifically addresses this but I'm pretty sure derivation by restriction is not used in any of the UBL xsd files.

I'm definitely not an XML Schema guru so there may be other ways to work around this problem (or an organization may decide that inheritance is not that important).  However, this is some of the background that went into the OAGI Core Component WG decision to implement CountryCode as a Data Type.

Regards,

garret
 

"Monica J. Martin" wrote:

Earlier in the process of developing the CCSD draft primer, a similar
question was raised.  Fred vanBlommestein created a short position paper
of options to consider which I have attached with his agreement.  Please
feel free to direct any questions, as Fred had detailed discussions on
this subject during previous UN/CEFACT meetings.

Thanks to Fred for his assistance (Fred, I believe this is the paper,
but I really had to search to find it from back in December 2002).

Monica

-------- Original Message --------
Subject:        [Fwd: RE: [ubl-lcsc] [roe] OAG comment:
CountryIdentificationCode can be a data type of Country Code]
Date:   Wed, 09 Jul 2003 16:06:58 -0600
From:   Monica J. Martin <monica.martin@sun.com>
To:     fjb <fjb@berenschot.com>, Monica Martin <Monica.Martin@sun.com>

I think your paper addressed the additional layer of aggregation to
handle the possibility of different attributes for country did it not, Fred?

Should I forward it or would you like to answer?

-------- Original Message --------
Subject:        RE: [ubl-lcsc] [roe] OAG comment: CountryIdentificationCode
can be a data type of Country Code
Date:   Wed, 9 Jul 2003 13:41:12 +0200
From:   Stig Korsgaard <STK@Finansraadet.dk>
To:     'Tim McGrath' <tmcgrath@portcomm.com.au>,
ubl-LCSC@lists.oasis-open.org

I have from the beginning wondered why we did that. There seems to be no
meaning in having the ABIE. On the other hand there is meaning in having a
different representation of Country in terms of either text or code but
rarely them both at the same time!

I am not sure though that the OAGIS solution agrees with CCTS - suggest we
investigate this further before making the final decision.

Best Regards

Stig Korsgaard
M.Sc.E Standardisation Coordinator
Tel:    +45 3370 1083
Cell:   +45 2121 8234
Mail:   stk@finansraadet.dk

Danish Bankers Association
Amaliegade 7
DK-1256 Copenhagen K
Tel:    3370 1000
Fax:    3393 0260
mail@finansraadet.dk
www.finansraadet.dk

-----Original Message-----
From: Tim McGrath [mailto:tmcgrath@portcomm.com.au]
Sent: 4. juli 2003 04:23
To: ubl-LCSC@lists.oasis-open.org
Subject: [ubl-lcsc] [roe] OAG comment: CountryIdentificationCode can be
a data type of Country Code

To facilitate debate and consensus on the reviewed items, I am posting
relevant issues as discussion points.  The intention is that after a one
week period the team shall resolve these issues.  For example, this
issue will be resolved at the call on Friday July 11th.

Item 4.

We created the ABIE Country and its CountryIdentificationCode to
facilitatet re-use in several structures.  This meant that Country as an
aggregate has only one property, IdentificationCode.

OAGIS has created a reusable Country Code data type and implemented a
Country data element of type=Country Code.  This seems to be a much
cleaner approach and accomplishes exactly the same thing.

Does anyone object to this approach?  It may also have implications for
other 'one-property' ABIEs.

--
regards
tim mcgrath
phone: +618 93352228
postal: po box 1289   fremantle    western australia 6160

You may leave a Technical Committee at any time by visiting
http://www.oasis-open.org/apps/org/workgroup/ubl-lcsc/members/leave_workgrou
p.php

You may leave a Technical Committee at any time by visiting http://www.oasis-open.org/apps/org/workgroup/ubl-lcsc/members/leave_workgroup.php

  ------------------------------------------------------------------------
                               Name: associations3.doc
   associations3.doc           Type: WINWORD File (application/msword)
                           Encoding: base64
                    Download Status: Not downloaded with message

  ------------------------------------------------------------------------
You may leave a Technical Committee at any time by visiting http://www.oasis-open.org/apps/org/workgroup/ubl-lcsc/members/leave_workgroup.php

begin:vcard 
n:Minakawa;Garret
tel;fax:650.506.7812
tel;work:650.506.3412
x-mozilla-html:FALSE
url:www.oracle.com
org:Oracle Corporation;e-Business Suite, ATG
adr:;;500 Oracle Parkway, M/S 6op6;Redwood Shores;CA;94065;USA
version:2.1
email;internet:garret.minakawa@oracle.com
title:Director, e-Business Integration
fn:Garret Minakawa
end:vcard