RE: [Ltru] Preliminary Investigation into Application of ISO 11179

"Debbie Garside" <> Wed, 28 June 2006 07:44 UTC

Received: from [] ( by with esmtp (Exim 4.43) id 1FvUiR-0002eD-VE; Wed, 28 Jun 2006 03:44:03 -0400
Received: from [] ( by with esmtp (Exim 4.43) id 1FvUiQ-0002e8-SQ for; Wed, 28 Jun 2006 03:44:02 -0400
Received: from ([]) by with esmtp (Exim 4.43) id 1FvUiP-0007E3-Jr for; Wed, 28 Jun 2006 03:44:02 -0400
Received: from (localhost.localdomain []) by ( with ESMTP id k5S7htAN019347; Wed, 28 Jun 2006 08:43:55 +0100
Received: from DebbieLaptop ( []) (authenticated bits=0) by ( with ESMTP id k5S7hncc019253; Wed, 28 Jun 2006 08:43:54 +0100
Message-Id: <>
From: Debbie Garside <>
Subject: RE: [Ltru] Preliminary Investigation into Application of ISO 11179
Date: Wed, 28 Jun 2006 08:43:50 +0100
MIME-Version: 1.0
X-Mailer: Microsoft Office Outlook, Build 11.0.5510
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
In-Reply-To: <>
thread-index: AcaaRtGC6+ZxHRxlRvygUz+eA4jaagAPctTw
X-Spam-Score: 2.1 (++)
X-Scan-Signature: dd055ca905b7a8538e016a7989511901
Cc: 'Doug Ewell' <>, 'LTRU Working Group' <>
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Language Tag Registry Update working group discussion list <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
Content-Type: multipart/mixed; boundary="===============1501619607=="

Hi Karen
Thanks for your input.  I need to do a little more reading before I respond
so it may be a few days :-)
I think there are two ways of looking at 11179: 1. As a standard to assist
in the design of a Meta-data Registry 2. As a standard to apply to an
existing Meta-Data registry.
I have been looking at what is as a bare minimum absolutely necessary in
order to apply ISO 11179 to the LTRU Registry.  I have read the whole
standard but I need to do further depth reading to see what is required for
the Registry to be conformant.
There are one or two things that you have mentioned that do not fit with my
interpretation but I would not want to say more until I have studied it
best regards


From: [] 
Sent: 28 June 2006 01:06
To: Debbie Garside
Cc: 'Doug Ewell'; 'LTRU Working Group'
Subject: Re: [Ltru] Preliminary Investigation into Application of ISO 11179


Did you review all of ISO 11179 or just 11179-6? 

There is a lot more to ISO 11179 than just the administrative practices (-6)
and the previous parts discuss a hierarchical metadata model not mentioned
in your review below. I think you're confusing the terms "value" and
"representation" and the other sections provide clarity on this.  The top of
the hierarchy is the Data Concept, which is an abstract description
independent of its representation -- a pure semantic layer. 

The hierarchy, as I understand it, would be something like this: 

Data Concept 
        Language Code: A standardized code used to identify a particular
language or dialect. 

Data Elements [Data Concept + Representation Class] related to "Language
Code" data concept = 

1.	ISO 639-1 Language Code:  A two-letter code assigned by the ISO
639-1 standard to identify a particular language. 

2.	ISO 639-2/B Language Code 

3.	ISO 639-2/T Language Code 

4.	ISO 639-3 Language Code 

5.	ISO 639-6 Language Code

Value Domain 
        Each data element in this case has a "Value Domain" and the Value
Domain contains the individual values such as "en-US". 


For the Language Code data concept, the ISO 11179 structure seems relevant
and useful. But when we look at some of the other concepts, it seems less
useful and perhaps problematic: 

Data Concept = Script Code 
Data Element = ISO 15924 Script Code 

Data Concept = Country Code 
Data Element = ISO 3166 Country Code 

Note that "Description" or even "Subtag Description" is too vague by ISO
11179 rules (subjective judgment, but there are a lot of examples that
support this view) and I think you would need to break these out as: 

1.	En-US Language Name 

2.	Fr-FR Language Name 

3.	En-US Script Name 

4.	En-US Country Name


These are unique data elements relating to several data concepts, so the
current model with its "Description" field would need serious revision to be
compliant, I think. 

I'm not opposed to further discussion of this moving forward. I only
question how valuable this is for a standard that has so few data elements.
It is a good thing that you're familiar with the section of the standard
I've spent the least time reviewing.  :) 

Best regards, 

Karen Broome
Sony Pictures Entertainment

"Debbie Garside" <> 

06/27/2006 01:22 AM 

"'LTRU Working Group'" <> 

'Doug Ewell' <> 

[Ltru] Preliminary Investigation into Application of ISO 11179


Findings of a preliminary investigation into the application of ISO 11179 to
the RFC3066bis Registry

Initial investigations suggest that ISO 11179 can be applied to the Registry
at a base level for very little cost.  The main cost is in mapping the ISO
11179 terminology to the existing Registry terminology and a number of
additional data elements would be required.  The Registry already
incorporates a system of metadata elements that are consistent with the
model presented within ISO 11179.  

In particular the value of the following aspects of ISO 11179-6 should be

The attributes registration authority identifier (RAI), data identifier
(DI), and version identifier (VI) constitute the international registration
data identifier (IRDI). At least one IRDI is required for an administered

Data identifiers are assigned by a Registration Authority; data identifiers
shall be unique within the domain of a Registration Authority. 

Requirements for a Registration Authority, and a discussion of the IRDI,
appear in ISO/IEC 11179-6.

As each Registration Authority may determine its own DI assignment scheme,
there is no guarantee that the DI by itself will uniquely identify an
administered item. For example, if two authorities both use sequential
6-digit numbers, there may be two administered items with the same DI's;
however, the administered items will almost certainly not be the same. 

If one administered item appears in two registers, it will have two DI's.
Therefore, both the DI and the RAI are necessary for identification of an
administered item. 

If particular attributes of an administered item change, then a new version
of the administered item shall be created and registered. The registrar
shall determine these attributes. In such a case, a VI is required to
complete the unique identification of an administered item.

For further guidance, see ISO/IEC 11179-6. 

An IRDI can serve as a key when exchanging data among information systems,
organizations, or other parties who wish to share a specific administered
item, but might not utilize the same names or contexts.

ISO/IEC 11179 does not specify the format or content of a unique DI.

The IETF (or LTRU) would need to apply for an International Code Designator
(ICD) - a four integer code; this coupled with the organization name as well
as a "department" identifier (OPI) becomes the IRDI e.g. 1234.IETF.LTRU.
The ICD would be registered by the RA of ISO/IEC 6523 Organization Codes as
Registration Authority Identifier which is currently BSI.  

Implications for the LTRU Registry
The DI (or UI - Unique Identifier) cannot be the Subtag as there are already
conflicting Subtags within the registry (e.g. cy/CY).  It is more preferable
that the unique identifier be the chosen language/country/script name
(please note, this is not the preferred name).  This would fit with the
current ISO 639-3, -5 and -6 models and open the Registry to adoption by
meta-data knowledge grids. (I will take a good look at the naming
conventions within ISO 3166-1 at a later date but prior to publication of
FDIS 3166-1).  

Anomalies within ISO naming conventions of standards issued prior to the
adoption of ISO 11179 can be dealt with on a case by case basis via set

The Subtag would become a "Representation" with the name being the unique
"Data Identifier". This would involve having a "Primary Description" which
would form the DI.

In reviewing the "Required Metadata Attributes" for a "Preferred Standard"
Status administered item, preliminary investigations reveal no serious
additional requirements other than those already mentioned here. Some
manipulation and interpretation of registry data and standard mandatory
requirements would be required but no difficulties are envisaged. I would
refer the WG to ISO/IEC 11179-6:2005(E); Table B-8 (p.34)

The ISO 11179 model allows for there being conflicting codes between
different meta-data registries in conformity with ISO 11179; that is part of
the conceptual model.  ("in conformity with" is correct - there are
essential parts of the standard).

In essence, the ISO 11179 meta-model supports linkage to other ISO 11179
conformant meta-data registries thus facilitating data exchange/interchange
whilst giving the LTRU Registry ownership of the data elements contained
therein - they become LTRU elements giving room for manoeuvre should ISO get
it wrong.

This will make language tags more meaningful in the future.  The key word
here is "linkage".  ISO 11179 conformant meta-data registries facilitate the
creation of knowledge grids, grid computing and the semantic web!

At first glance the cost/value ratio favours ISO 11179; there appears to be
very little cost yet the true benefits of future interoperability and data
exchange are unknown.  

It is recommended that further investigation be conducted before application
of ISO 11179 can be discussed at WG level. 

Further benefits with regard to data interchange should be explored.

It is further recommended that the "investigation into application of ISO
11179 Meta Data Registries to the Registry and its registration procedures
be conducted by nominated members of the WG with a view to application" be
added to the new LTRU Charter.  

Best regards

Debbie Garside

Ltru mailing list

Ltru mailing list