ACTION Required: FW: INCITS-L8-INTEREST: Re: Preliminary comments from O

Kathryn Breininger
Manager, Release & Delivery Services
CIMS - Center for Information Management Services

Kathryn,

I forwarded the OASIS Reg/Rep comments directly to the SC 32 Secretariat for inclusion with other comments received for the ballot on 11179-3.

L8 discussed the comments you sent and had some responses, mostly seeking some clarification. Kevin Keck wrote up the following. The “I” refers to Kevin after discussion with L8.

If you have any additional responses, we will consider them during the ballot resolution meeting in Jeju, Korea, scheduled for the week of June 22.

Thanks again.

Bruce

----Sent by--------------------------

Bruce Bargmeyer

Lawrence Berkeley National Laboratory

University of California

1 Cyclotron Road, MS 50B-2231A

Berkeley, California 94720

Tel: +1 510-495-2905

Fax: +1 510-486-4004

email: bebargmeyer@lbl.gov

From: Breininger, Kathryn R [mailto:kathryn.r.breininger@boeing.com]

Sent: Friday, May 01, 2009 9:25 AM

To: BEBargmeyer@lbl.gov

Cc: ebXML Regrep

Subject: RE: FW: [regrep] FW: Comments solicited -- ISO/IEC CD2 11179-3 andISO/IEC 19763 Part 3

Hi Bruce,

Thank you again for inviting us to comment on the ISO 11179-3. The TC met yesterday and finalized the comments, which are included below. Please let us know if you have any questions or any comments on our comments!

Overall, we find that the spec is a very well put together. Here are some comments we have compiled so far:

Technical Comments:

5.1.3 Contact, 5.1.5 Individual: Person <=> Contact model mismatches
- Suggest aligning with regrep with a common Party class that Organization and Person extend
- Add address, phone etc. to Party
- Direction of relationship between contact_info and Individual is not intuitive. An individual has contact_info and not the other way around.Consider reversing the relationship
- Take away title from Person and instead make it an attribute of association with an organization (titles or roles are in the context of a relationship with some organization)

The term Individual was used in preference to Person because in WG1 (of ISO/IEC JTC1 SC32) Person is defined in the legal sense, which actually corresponds to what OASIS calls a Party.

There also seems to be confusion reflected here about the meaning of the Contact class, which is not simply a record structure for contact information, it instead represents a *role* of a Person (Individual) within some Organization, i.e., is (logically) an association between a Person (Individual) and an Organization. It is for this reason that title is not an attribute of Person (Individual), it is an attribute of Contact. It is also intentional that phone, etc. are not attributes of Person (Individual) either, but rather of Contact. The US TAG has discussed changing the name of the Contact class to Representative, in part to hopefully make this more clear. I suggest we include that suggestion in our ballot comments for CD2.

I do not feel strongly about whether to include a Party superclass in the 11179-3 metamodel, but it was decided previously by the US that such a class was not needed for the purposes of 11179-3, and thus imposing a requirement that implementations provide such a supertype was not justified. Absence of the supertype in the 11179-3 metamodel should not preclude implementations from having such a supertype, the issue is whether it should be required as a matter of conformance.

6.1.2.2 Scoped_Identifier:
Suggest simplifying identifier scheme. Consider providing an example that maps to a URN naming scheme

I agree that an example would be very helpful. I will propose a ballot comment to provide at least one, in an annex.

6.1.2.4 Slot: Thanks for the good alignment here

No comment called for on ballot, but thank you.

6.2 Designation and Definition region: This clause is very difficult to follow. Its not clear what a Designatable_Item is. Suggests providing examples and clearer definition
- designation_sign attribute is particularly not clear

I will propose a ballot comment calling for some examples in clause 6.2.2.3.1.

I will also propose a ballot comment suggesting a better definition of designation_sign in clauses 3.4.45 and 6.2.2.3.2.1. Working new definition: sign denoting the designatable item, as represented by a designation.

7.1 Registration metamodel region: Consider aligning this section with ISO 19135

Could you be more specific about what parts of 19135 and 11179-3 ought to be (better) aligned?

I would note that 19135 includes 11179-3 in its bibliography and cites it as the source of two definitions (for identifier and registry). Registration in 19135 appears to have departed from 11179-3 in two main ways:

1) A distinction is made between a Register Owner and a Register Manager, whereas 11179-3 only defines a Registration Authority. Is this distinction relevant to 11179-3?

2) 19135 formalizes a notion of a hierarchical register. Is this important to also formalize in 11179-3?

8.1.2.1: Concept System section is difficult to understand. Suggest clarifying text and adding examples

Examples are provided in Annex F. I will propose a comment to insert a forward reference from clause 8 to that annex, directing readers to look there for examples.

For clarifying text, more specific comments are requested about what requires clarification.

8.2.2 Classes in the Classification region:Good alignment in Classification region

No comment called for on ballot.

RegRep and 11179 have a very different model for handling language specific content. Perhaps this is an area where we can collaborate to achieve better alignment

I think that's a great idea, yes. But it sounds like you'd prefer to put this down as a future task, not a change to be made to 11179-3 at this time, so there is no comment being called for here on the CD2 ballot.

Does the spec have something analogous to ebXML RegRep RegistryPackage? If not, consider adding it as we have found it very useful.

It currently does not, but provision of some type of grouping mechanism has been suggested before, by experts in the US TAG. I think what has been lacking is an articulation of requirements to be met by such a mechanism.

We were unable to find a place in the spec where Association support and Association metamodel was described. Consider defining a clearer Association metamodel

In some ways the Relation and Link parts of the Concept System metamodel could be considered a specialization of Association as defined in Reg/Rep, scoped to apply only to Concepts (of which many, but not all, of the Data Description classes are subtypes). One difference, however, is that Links in 11179-3 may be of higher arity than 2. Another is that Links do not have one of their ends as the "source"—the "source" of a Link is a concept system(s), rather than one of the ends of the Link.

Another element in the 11179-3 metamodel that highly resembles Reg/Rep Associations is concept system reference (clause 8.1.3.3), which is binary, is directional, and is required to be (sub)typed (it is marked as abstract). An implementer of 11179-3 who wished to do so on top of an implementation of Reg/Rep would be smart to implement concept system reference as a subclass (either explicitly or implicitly) of Association.

Introduction of the very abstract Association facility from Reg/Rep into 11179-3, though, would I think be strongly resisted by the 11179 community at large, because it would be unclear what unmet need such a generic facility would serve, and would raise legitimate concerns about how such things could be interpreted consistently when encountered in a metadata registry.

We were unable to find a concept equivalent to Repository or RepositoryItem as defined in RegRep. Is this intentionally out of scope? If so, please clarify in 1. Scope section.

I think the closest thing in 11179-3 currently is Reference_Document. 11179-3 does not specify whether the document itself should or should not be stored within the registry, but CD2 does provide an uri attribute by which the document might be accessed, wherever it is stored.

I am happy to propose a comment to state this explicitly in the scope section.

Examine spec for forward references and minimize whenever possible. An example is that of Designatable_Item

I don't think I would describe this problem as one of forward references, but I hear again frustration about the description of Designatable_Item. This should be captured in a comment, even if we have no specific recommendation to offer as a proposed solution.

[end of technical comments]

General Questions:

Is there a comments list where we can send any future comments?
What public mailing lists can one signup to to stay informed of progress of the spec?
Would it be possible for us to get feedback on RegRep 4 specifications from ISO 11179 spec team?
Would it make sense to have a formal liaison between our two groups? RegRep TC feels that would be good. As a first exercise the liaison could define a cross-walk / mapping between the concepts of the two specs
Consider adding a reference to OASIS ebXML RegRep 3.0 specifications in Bibliography section

Thanks again to for soliciting our inputs on this good work. We look forward to seeing the next version of the spec and to continued collaboration between our respective teams to achieve closer alignment in our specs.

Sincerely,

Kathryn

Kathryn Breininger
Manager, Release & Delivery Services
CIMS - Center for Information Management Services

MC 62-LC
425-965-0242 desk
425-512-4281 cell
425-237-4582 fax

From: Bargmeyer [mailto:bebargmeyer@lbl.gov]

Sent: Tuesday, April 21, 2009 3:25 PM

To: Breininger, Kathryn R

Subject: RE: FW: [regrep] FW: Comments solicited -- ISO/IEC CD2 11179-3 andISO/IEC 19763 Part 3

Katharine,

Thanks for your work on making comments. April 30 will be fine. If you have some comments earlier, that would also help. It would give some time for getting them ready for discussion.

Thanks.

Bruce

----Sent by--------------------------

Bruce Bargmeyer

Lawrence Berkeley National Laboratory

University of California

1 Cyclotron Road, MS 50B-2231A

Berkeley, California 94720

Tel: +1 510-495-2905

Fax: +1 510-486-4004

email: bebargmeyer@lbl.gov

From: Breininger, Kathryn R [mailto:kathryn.r.breininger@boeing.com]

Sent: Monday, April 20, 2009 8:47 AM

To: BEBargmeyer@lbl.gov

Cc: Farrukh Najmi

Subject: RE: FW: [regrep] FW: Comments solicited -- ISO/IEC CD2 11179-3 andISO/IEC 19763 Part 3

Importance: High

Hi Bruce,

Please hold off on redistributing the draft comments. We are worked on these during our meeting, and have modified some. We will send you a complete set of comments that have been blessed by the whole group no later than April 30th, earlier if possible. Will that still fit within your time frame?

Kathryn Breininger
Manager, Release & Delivery Services
CIMS - Center for Information Management Services

MC 62-LC
425-965-0242 desk
425-512-4281 cell
425-237-4582 fax

From: Bargmeyer [mailto:bebargmeyer@lbl.gov]

Sent: Saturday, April 18, 2009 10:32 AM

To: Breininger, Kathryn R; 'Farrukh Najmi'

Subject: RE: FW: [regrep] FW: Comments solicited -- ISO/IEC CD2 11179-3 andISO/IEC 19763 Part 3

Kathryn and Farrukh,

I very much appreciate the comments you put together. I would like to distribute them to SC 32/WG 2. Can I do so? I’d like to post the comments so that people can see and comment on them.

Do the comments have the blessing of the whole group?

There is a discussion place for the SC 32/WG 2 standards that are under development. It is: issue.metadata-standards.org. Anyone can read the issues. You have to sign up to be able to post comments, but registration is open to all.

If you set the search there to 11179-3 ED3, you will get all of the issues and comments that are posted there. It is OK for issue management, but not so good for getting a comprehensive picture. Issue 422 – 450 are issues inviting comment on the current draft by clause. However, the prior issues cover individual items in more detail.

If you want, I can ask to have you subscribed to receive a message when something is posted there. The messages give a bit of information, but have a link to the comment. It goes in bursts, lots of messages sometimes, then lulls.

There is also an email reflector, 'incits-l8-interest@incits-l8.org'. If you like, I will request anyone (in the US) to be put on that. Quite a lot of messages on that one, some of which are relevant to 11179. There is an international counterpart: sc32wg2-interest@metadata-standards.org, which does not get a lot of traffic. Let me know if you or others want to subscribe to these.

Thanks.

Bruce

----Sent by--------------------------

Bruce Bargmeyer

Lawrence Berkeley National Laboratory

University of California

1 Cyclotron Road, MS 50B-2231A

Berkeley, California 94720

Tel: +1 510-495-2905

Fax: +1 510-486-4004

email: bebargmeyer@lbl.gov

From: Breininger, Kathryn R [mailto:kathryn.r.breininger@boeing.com]

Sent: Thursday, April 16, 2009 7:43 AM

To: Farrukh Najmi; ebXML Regrep

Subject: RE: FW: [regrep] FW: Comments solicited -- ISO/IEC CD2 11179-3 andISO/IEC 19763 Part 3

Importance: High

Thank you Farrukh for your thoughtful review. We will discuss during our meeting today, edit if needed, and if all agree will send comments to Bruce.

Kathryn Breininger
Manager, Release & Delivery Services
CIMS - Center for Information Management Services

MC 62-LC
425-965-0242 desk
425-512-4281 cell
425-237-4582 fax

From: Farrukh Najmi [mailto:farrukh@wellfleetsoftware.com]

Sent: Thursday, April 16, 2009 5:56 AM

To: ebXML Regrep

Subject: Re: FW: [regrep] FW: Comments solicited -- ISO/IEC CD2 11179-3 andISO/IEC 19763 Part 3

Team here are my revised comments on this spec...

Overall, it is a very well put together spec. The way to think of ISO

11179 is that it describes a registry/repository in even more abstract

terms than ebXML RegRep. In the ideal world ebXML RegRep standard simply

provides a concrete binding for ISO 11179.

However, in the real world the specs have been done by different people

with limited alignment. Historically, we have taken 11179 specs as input

and tried to align with them as best as possible. Complicating things

are other competing specs like ISO 19135 for Registration Procedures.

I think is is a very commendable thing that the 11179 fold and

specifically Bruce Bargmeyer have taken the time to get our input. We

should ask for their input in our latest specs as well. The comments

below are based on a very quick review.

5.1.3 Contact, 5.1.5 Individual: Person <=> Contact model mistmatches
- Suggest aligning with regrep with a common Party class that Organization and Person extend
- Add address, phone etc. to Party
- Take away title from Person and instead make it an attribute of association with an organization (titles or roles are in the context of a relationship with some organization)
5.1.14 Registration_Authority_Identifier: Why have separate attribute for registration_authority_identifier. Better to representregistration_authority via an Organization and use orgs identifier
6.1.2.2 Scoped_Identifier: Suggest defining a URN naming scheme instead of current spec
6.1.2.4 Slot: Thanks for the good alignment here
6.2 Designation and Definition region: This clause is very difficult to follow. Its not clear what a Designatable_Item is. Suggets providing examples and clearer definition
7.1 Registration metamodel region: This section should be aligned with ISO 19135
7.1.2 Registration Record, Stewardship Record, Submission_Record: RegRep TC needs to see if these are relevant to our Registration Procedures work
7.1.6.1 attachment: This is so much better a name than RepositoryItem (sigh: why did we not think of it)
8.1 8.1 Concept System region: ClassificationScheme <=> Concept_System, ClassificationNode <=> Concept is another terminology mis-alignment. Perhaps that is OK since 11179 is meant to be more generic than ebXML RegRep
8.1.2.3 Assertion: Need more examples or clearer description of how Assertions play a role in a concept system
8.1.2.4.1 Description of Relation: Need more examples or clearer description of how Relations play a role in a concept system
8.2.2 Classes in the Classification region:Good alignment in Classification region
9 Binary_Relations Package: This fuctionality needs to be studied for relevance in ebXML RegRep
10 Data Description Package: This fuctionality needs to be included in a future version of ebXML RegRep
10.4 Measurement region: This fuctionality needs to be included in a future version of ebXML RegRep
Does the spec have something analogous to ebXML RegRep RegistryPackage? If not consider adding it
There does not seem to be anything analogous to RegRep InternationalString/LocalizedString or how to do internationalization of content. For an international standard this is important to include. Consider aligning with RegRep
I was unable to find a place in the spec where Association support and Association metamodel was described. Did anyone else find it?
Is there a comments list where we can send any future comments?
What public mailing lists can one signup to to stay informed of progress of the spec?

Thanks again to Bruce and 11179 team for soliciting our inputs.

Lets discuss these comments later today in our meeting.

Regards,

Farrukh

Web: http://www.wellfleetsoftware.com

Breininger, Kathryn R wrote:

Please note: this is one agenda item I want to be sure we have time to discuss, so will be placing it at the top of our agenda. Please review materials (see below) prior t o our meeting.

Regards,

Farrukh

Web: http://www.wellfleetsoftware.com

--------------------------------------------------------------------- To unsubscribe from this mail list, you must leave the OASIS TC that generates this mail. Follow this link to all your TCs in OASIS at: https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php

regrep message