[DDI-users] Data Element Discussion - Mantis Issue 543 and related [SEC=UNCLASSIFIED]
Wendy Thomas
wlt at umn.edu
Mon Mar 18 13:33:22 EDT 2013
Hi Al,
There were a couple of points raised with the originally proposed
DataElement in 3.2PR. First that it did not allow for a CodeListReference
but instead allowed for a DataElementRepresentation of type "Code" which
essentially indicated it was enumerated but didn't say with what. Our
initial position was that the DDI DataElement was a "stripped down"
codification of 11179 and we didn't provide extensive detail. It did
support the use of a Value Domain specification in the form of a
CategorySchemeReference (0..1). The DataElementRepresentation was a
specified list of options reflecting those noted in the earlier edition of
11179-3 (plus other...I think...I'd need to check).
The argument was made that a) "Code" should not be on this list and b)
CodeListReference should be allowed
[I think the additonal points regarding the support of the direct use of
DataElementConcept are not contensious]
After reading the changes made in the 2012 edition of 11179-3 I felt that
the differences between the use of a specific CodeList which provided
enumerated representation (permitted values and link to specific
concepts/category for each) and the use of "code" in a described
representation was clearer.
What I essentially threw out there was where we would go if we wanted
something broader than the very sparse view originally provided. This
seemed to be what was being requested...i.e. the ability to specify an
enuemrated representation. So if I was specifying that, why not more
specification to the described?
I think you are correct that the first question we need to answer is what
the function of DataElement is in DDI.
This could range from:
very narrow (remove the option for a CodeListReference and retain the
DataElementRepresentation (as either an internal enumeration or an external
Controlled Vocabulary...as edition 2012 leaves this pretty open)
full-fledged (11179-3 with all the bells and whistles except for the
administrative content)
I think questions on the implecations for inheritance by the Variable need
to be answered for either extreme as it has already been raised "why have
use the ConceptReference in a Variable if you've referenced a DataElement?"
The first quesion that needs to be answered is "What is the purpose of a
DDI DataElement?
I am currently agnostic on this and would like to hear from others.
Best regards,
Wendy
On Mon, Mar 18, 2013 at 12:27 AM, Alistair Hamilton <
alistair.hamilton at abs.gov.au> wrote:
> Hi Wendy
>
> The following are not comments, but questions (so ABS can progress to
> comments).
>
> I remember talking with Arofan many months ago and he mentioned Data
> Elements were being added to DDI 3.2 to capture more "timeless" (eg don't
> change from reference period to reference period) aspects of Variables.
>
> Heather has noted that the proposed specification for Data Element has
> "grown" between the PDF included in the initial release of PR documentation
> (Doc1) and the latest document (Doc2) you circulated. This growth does not
> appear to be simply a symptom of the Doc2 proposal aligning with Ed 3 of
> 11179 (for example, ValueDomain, which is in Doc2 but not Doc1, exists in
> Ed 2 of 11179 also.)
>
> Whether the Doc2 proposal is overkill depends on why Data Element is being
> added to DDI and the expected use of it within DDI. The Doc1 specification
> would not allow a "fully functional" 11179 Data Element to be defined
> within DDI. The Doc2 specification comes a lot closer to permitting that.
>
> Then again, we still have VariableRepresentation available for a variable.
> Presumably providing relatively detailed Value Domain information for the
> DataElement (eg a CodeListReference) associated with a Variable is not an
> alternative to populating VariableRepresentation for the Variable itself?
>
> I quite like the Doc1 approach (and even the DataElementConcept approach
> in 3.1) because if an implementer has a store outside DDI that defines and
> manages DEs (DECs) fully in accordance with 11179 they can reference these
> information objects from within DDI without importing too many 11179
> constructs/structures into DDI itself.
>
> If the intent were to be able to "carry around" more 11179 content within
> DDI, however, then the Doc2 approach supports that.
>
> There could be some merit to the Doc2 appropach - although it seems to
> create risks of
>
> - duplication (in different forms) of information between the
> structural definition of the DE and the structural definition of the
> Variable
> - the need for practice guidelines on, eg,
> - the level of consistency expected between the "overlapping"
> structural information for CodeListReferences and VariableRepresentations
> - the need for agencies who do focus on DataElements to populate
> the Variable object to a level were it will be interoperable with other DDI
> implementations
>
>
> Was the intent of implementing DataElement in 3.2 that DDI should be able
> to act as a "lite" implementation of that region of the 11179 metamodel or
> was it more that content in DDI should be able to integrate with an
> external implementation of that region of the 11179 metamodel?
>
> A "sleeper" issue also may be that the DE proposal (even in the Doc1
> specification potentially) carries some information that (if DDI aligns
> with 11179) really belongs on an updated version of the DEC object within
> DDI rather than (only) being built into the new DE object.
>
> Cheers
>
> Al
>
> [image: Inactive hide details for Wendy Thomas ---15/03/2013 12:10:47
> PM---Dear Dan G. Al, Heather, Arofan, Jeremy, Dan, Tim, Heather,]Wendy
> Thomas ---15/03/2013 12:10:47 PM---Dear Dan G. Al, Heather, Arofan, Jeremy,
> Dan, Tim, Heather, and Sophia, As a follow-up to the earlie
>
> From: Wendy Thomas <wlt at umn.edu>
> To: "DDI Structural Reform Working Group." <ddi-srg at icpsr.umich.edu>,
> Data Documentation Initiative Users Group <ddi-users at icpsr.umich.edu>,
> Alistair Hamilton <alistair.hamilton at abs.gov.au>, Heather Purcell <
> heather.purcell at abs.gov.au>, Tim.Dunstan at statcan.gc.ca, Dan Gillman <
> gillman.daniel at bls.gov>, "Kuan, Sophia [USA]" <kuan_sophia at bah.com>,
> Date: 15/03/2013 12:10 PM
> Subject: Re: Data Element Discussion - Mantis Issue 543 and related
> ------------------------------
>
>
>
> Dear Dan G. Al, Heather, Arofan, Jeremy, Dan, Tim, Heather, and Sophia,
>
> As a follow-up to the earlier documents trying to frame the issues
> around DataElement I have read through the new edition of
> ISO/IEC11179-3:2012(E) and provided a proposal for reworking
> DataElement. While I don't think this is its final form, I thought
> that having something concrete to react to would be useful. Please see
> the new documents attached to Mantis Issue 543.
>
> DataElementProposal_20130312.docx discribes the reasoning behind the
> outline of the proposed structure
> DataElementProposal.txt is a draft of an XML schema representation
> (note I realize there are some style issue in this rendering..its
> primarily to provide a draft to work with)
>
> The workplan is as follows:
>
> Provide feedback via email over the next week.
> Identify at the end of that time if we need a conference call. If so
> set one up for the following week.
>
> Please note that many of us are involved in NADDI the first week of
> April. As we have other issue areas that many of you are also
> interested in I am hoping to get this one moving along. The intent is
> not to cut short discussion, but to allow us to focus on a topic at a
> time. I appreciate your attention to this. Hope to hear responses over
> the next week.
>
> Primary questions to answer include:
>
> Does this address the issues identified in the document provided earlier?
> Is it compliant with ISO/IEC 11179-3:2012(E)?
> Does it meet the needs of DDI?
> Is it overkill?
>
> I expect a lively discussion :-)
>
> Wendy
>
>
>
> On Sun, Feb 24, 2013 at 5:20 PM, Wendy Thomas <wlt at umn.edu> wrote:
> > Dan G., Al, Heather, Arofan, Achim, Jeremy and Dan,
> >
> > I have tried to frame the issues regarding the introduction of a
> > DataElement in DDI based on earlier discussions and a review of
> > ISO/IEC 11179-3. You will find a document on Mantis Issue 543 entitled
> > Purpose of DataElement in DDI.docx. Please review this document and
> > make comments back to the list. There are a number of issues in 3.2PR
> > that need in-depth discussion. DataElement is the first of these. Our
> > intent is to begin an on-line discussion among interested parties,
> > attempt to find an area of concensus and then set up conference calls
> > to finalize the details. Please note that this email has been copied
> > to the DDI-SRG list and DDI-Users list in order to inform others of
> > this discussion and to identify persons who should be involved further
> > in the discussion.
> >
> > To DDI-SRG and DDI-User list members: If you have a vested interest in
> > how this new section of DDI is addressed in 3.2 PR, please let me know
> > and I will add you to the contact list. Note that we are dealing with
> > fine tuning what was proposed in DDI 3.2PR and therefore you will need
> > to have a familiarity with ISO/IEC 11179-3 and/or the maintenance of
> > DataElements for these discussions to be of benefit to you and to DDI.
> > Having said that, I wish you to know that I am very happy to address
> > any specific questions or related issues you may have with this
> > discussion of DataElements.
> >
> > Regards
> > Wendy
> >
> > --
> > Wendy L. Thomas Phone: +1 612.624.4389
> > Data Access Core Director Fax: +1 612.626.8375
> > Minnesota Population Center Email: wlt at umn.edu
> > University of Minnesota
> > 50 Willey Hall
> > 225 19th Avenue South
> > Minneapolis, MN 55455
>
>
>
> --
> Wendy L. Thomas Phone: +1 612.624.4389
> Data Access Core Director Fax: +1 612.626.8375
> Minnesota Population Center Email: wlt at umn.edu
> University of Minnesota
> 50 Willey Hall
> 225 19th Avenue South
> Minneapolis, MN 55455
>
>
> -----------------------------------------------------------------------------------------------
>
> Free publications and statistics available on www.abs.gov.au
>
>
>
>
--
Wendy L. Thomas Phone: +1 612.624.4389
Data Access Core Director Fax: +1 612.626.8375
Minnesota Population Center Email: wlt at umn.edu
University of Minnesota
50 Willey Hall
225 19th Avenue South
Minneapolis, MN 55455
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.icpsr.umich.edu/pipermail/ddi-users/attachments/20130318/9f171291/attachment-0001.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/gif
Size: 105 bytes
Desc: not available
Url : http://lists.icpsr.umich.edu/pipermail/ddi-users/attachments/20130318/9f171291/attachment-0001.gif
More information about the DDI-users
mailing list