It boils down to being clear and consistent about precisely what the identifier identifies.
steve
From: gbergcross=***@***.***-groups.org [mailto:***@***.***-groups.org] On Behalf Of Gary
Sent: Tuesday, April 19, 2016 9:24 AM
To: Greenberg,Jane ; Data Fabric IG ; Data Foundations and Terminology IG ; Research Data Collections WG
Cc: reaganwmoore ; Keith Jeffery ; Jett Jacob ; Schwardmann’ ‘Ulrich ; fbaumgardt
Subject: [rda-datafabric-ig][rda-dft][rda-collection-wg] Some thoughts on “Data Aggregations” terminology & concepts
There may be useful analogy to a physical sample that is collected.
You have that object at birth with an id, but over time you can add to this
and take away from it. The original ID serves a purpose but you may need an ID for the current state
of the object at any time there has been change.
Gary Berg-Cross, Ph.D.
***@***.***
​​
​
http://ontolog.cim3.net/cgi-bin/wiki.pl?GaryBergCross
Member, Ontolog Board of Trustees
Independent Consultant
Potomac, MD
240-426-0770
On Tue, Apr 19, 2016 at 12:04 PM, Greenberg,Jane wrote:
Reagan, this is a really nice outline.
All – I am on this list, but not sure exactly why? I think I was part of the original Data Fabric IG?
In any event, I’m jumping in below, with a comment that I think is implicit in the ongoing discussion, but want to emphasize one of Reagan’s below points—about entities can change over time.
A collection—viewed as either an object or collection of object (however you like it), is also a convenient grouping, and, yes, can be marked theoretically — at a point in time, although it is not necessarily stagnant.
Think most folks are on board with this, but it is worth restating b/c it impacts the identifier discussion and when is the collection a new ‘work’ meriting a different identifier.
Best wishes, jane greenberg
Jane Greenberg, Ph.D., A. B. Kroeger Professor,
Director, Metadata Research Center
College of Computing & Informatics
Drexel University
cci.drexel.edu/mrc/
(***@***.***)
215.895.2490
From: on behalf of reaganwmoore
Date: Tuesday, April 19, 2016 at 9:35 AM
To: Gary , Keith Jeffery , Data Fabric IG , Data Foundations and Terminology IG , Research Data Collections WG
Cc: Jett Jacob , Schwardmann’ ‘Ulrich , fbaumgardt
Subject: [rda-datafabric-ig][rda-dft] Re: [rda-datafabric-ig][rda-dft][rda-collection-wg] Re: [rda-datafabric-ig][rda-collection-wg] Some thoughts on “Data Aggregations” terminology & concepts
Gary:
We focus on collections that enable research collaborations. The types of entities we register into a collection include:
* sub-collections
* files
* soft links to other collections (micro-service structured objects)
* soft links to objects in external repositories (micro-service structured objects)
* database queries (micro-service invocation)
* workflows (workflow structured objects)
* sensor data streams (micro-service structured objects)
Many of these entities drive the dynamic execution of a procedure to realize the desired entity. Thus we use a collection to organize both data files and procedures.
A focus on publication instead of collaboration could use the definitions that are being proposed. The digital object is then static, invariant, and a persistent identifier has meaning. In a research collaboration, the entities change over time and are tracked by versions.
* A soft link can reference an entity that changes over time.
* A query issued to a database can be invariant, but the result set may change each time.
* A sensor data stream always has new data from the most recent observation. The stream itself may be identified, but the contents are not static.
* If we change the input parameters for a workflow, the result will change when the workflow is executed. Thus a workflow structured object has to associate the workflow with the input and the output.
Reagan
From: on behalf of Gary
Date: Monday, April 18, 2016 at 12:09 PM
To: Keith Jeffery , Data Fabric IG , Data Foundations and Terminology IG , Research Data Collections WG
Cc: Jett Jacob , Schwardmann’ ‘Ulrich , fbaumgardt
Subject: [rda-datafabric-ig][rda-dft][rda-collection-wg] Re: [rda-datafabric-ig][rda-collection-wg] Some thoughts on “Data Aggregations” terminology & concepts
Keith,
In short I think the answer is yes, that collections such as generated by queries can be parts of DO collections.
I would go back to Reagan Moore’s early observation (to effect if I understand right) that ” “Digital collections implement arrangement by a community for organizing their digital entities..”
They are then aggregations of interest defined by communities.
Now this ultra flexibility may, I have a sense of worry without having figured this out, overload the role of digital object since it seems to have taken on the idea that
it can be anything that is digital and identified. We are then relying on provenance info to tell us something about how the DO is created. It may be from parts of queries etc. that tells us the nature of the collection.
Is that a problem??
Gary Berg-Cross, Ph.D.
***@***.***
​​
​
http://ontolog.cim3.net/cgi-bin/wiki.pl?GaryBergCross
Member, Ontolog Board of Trustees
Independent Consultant
Potomac, MD
240-426-0770