The initial dire note about the migration has been reduced, Anna has a short term solution for concerns about ordering but this meeting is to help address the longer term concerns.
A great deal of the boring back end technical details are not usually discussed, but there are aspects of it that can reach up to affect issues.
- Currently the focus is to have data in Hydra followed the linked data model, this requires RDF modeling of metadata about objects.
- RDF models data in a flat model without hierarchy.
- Upstream developers have done additional work with RDF that makes a ordering hierarchy difficult.
Does the order of multi-valued fields have value?
The main concern is if the current fields that allow multiple values have some intellectual value or meaning to their order. If so, does this meaning need to be preserved?
Two sample objects in Hydra were chosen, a stamp and a galvanometer.
- Galvanometer (Museum): The order of entries in the medium field does have meaning. Medium is listed in order of the amount of material used in the object.
- Stamp (Archives): Subjects for the stamp do not have a meaningful order.
Review of Fields and Discussion
The initial method of sorting for user view to be proposed was an alphasort (alphabetical sort).
Alphasorting data has the following benefits:
- Easy to implement
- Does not require data model changes
Alphasorts have the following potential downsides:
- Order based meaning may be lost
A review of the fields was conducted with Hillary Kativa, Stephanie Lampkin, Sarah Newhouse, Patrick Shea, Ashley Augustyniak speaking as to how they viewed the various fields.
- Subject: This field did not have anyone mention that the order is meaningful. Patrick specifically called out that his order is not hierarchical, and Sarah mentioned adding a forgotten subject to the bottom later. Ashley was going to talk to Andrea to see if any of the cataloging assumed a hierarchy to the data that was not remembered at the time.
- This field is probably fine with alphasort.
- Medium: The order does matter to this field for the Museum objects. Stephanie explained that the order of medium entries represents percentage of object composition. This is valuable for storing and caring for the item.
- A question was raised if this data is best kept in order in Hydra, or if it is of value in the CMS but not in Hydra.
- Currently this field should have an ordered entry to maintain the hierarchy of data.
- Maker: There was a short discussion if the order of the maker is meaningful to show the relative contribution that different makers had on a work or object. It might be meaningful, but a sample object that was considered handled either order of the entries fine.
- This field is probably fine with alphasort.
- Genre:
- Creator: Nobody expressed any issues with this field.
- This field is probably fine with alphasort
- Author: This field might have some concerns, an example of scientific journals was brought up where the order of authorship has meaning. Patrick mentioned that those are not a major component of intended Hydra material.
- If Hydra will automatically generate citations, the order of authors will affect citations. If this matters enough to be worth the effort modeling things was not discussed.
- This field is probably fine with alphasort, unless citation concerns change things.
- Interviewer: This field is primarily used by Oral History data. A meeting is needed to determine if order here is meaningful. There is a desire to have oral histories provided with citation information, and this will affect citations.
- Status pending discussion
- Title:This field does need modeling changes. Anna mentioned a primary title and alternate titles set of fields. Nobody expressed any concerns with this plan.
- This field will need modeling changes with a single value title field and a multi-value alternate titles field that can be alphasorted.
- Description: This field currently is multivalued. Order does matter here since paragraphs/divisions are handled by entering data in different sections.
- A modeling change was proposed by Anna where Description would be a single value field with the ability to handle paragraph breaks.
- This covered the needs to use the field.
- A modeling change was proposed by Anna where Description would be a single value field with the ability to handle paragraph breaks.
- Series Arrangement: This field, used for Archival Material, does have value in the order of entries. Sarah noted that the notation used to express this (Series, Sub-Series) is already alphabetical however.
- An alphasort would preserve the ordering data in this field.
- Anna mentioned it would be more technically correct in a data modeling sense to turn this into multiple fields (Series, Sub-Series). This matters for data exporting since the sorting is at the display level not the triplestore level.
Exceptions to various fields were mentioned as possible, though unlikely, and the general consensus was that rare exceptions might be best handled with information added to the description field to clarify matters.