Satisfy data requirements for relevancy customization work

[nikdragovic]

Mark's initial recommendations:

50% of the subset should represent "typical" works from the collection 50% of the subset should represent edge-cases which might include:

• items with unusually long titles
• items with unusually long descriptions
• items with many keywords / subjects
• items with no keywords / subjects
• items with metadata missing in other key fields
• items with unusual metadata
• items with the word "Emory" in fields other than title and description

Via Collin, this needs to happen in test.

As of Feb. 12, there are 1718 objects in arch and 283 in test. Types that will need to be tested:

• Collection objects
• Parent hierarchical objects
• Child hierarchical objects
• Non-hierarchical objects

We will need to test objects from all collections (with the possibility of excluding yearbooks). Possible need to test all targeted fields?

List of stable objects Current sort behavior

[nikdragovic]

We've decided to go with a teachout on this, since no changes seem to be required at this time.