This profile is derived from the profile of a real collection: http://bindings.lib.ua.edu/
This collection consists of approximately 5400 book cover objects, which are stored in Fedora using an atomistic model: the parent object contains descriptive and administrative metadata about the book, and the child objects contain images (one image per child) and technical metadata about the image.
For the sake of simplicity, I've flattened the collection into compound objects: every object contains both the descriptive and administrative metadata, and the image content streams and technical metadata. I've left out the METS structmap, RELS-EXT, RELS-INT, and DC datastreams, to give a picture of the content.
| Parameter | Suggested values | Notes | |
|---|---|---|---|
| Number of objects | 10,0006000 | ||
| Average number of datastreams per object | 4see note | 1 - 4 binary TIFF datastream , 3 XML metadata datastreams: average is 2.19 1 - 4 XML MIX technical metadata streams, one for each image: average is 2.19 1 MODS datastream 1 PREMIS datastream Total average: 4.19 datastreams per object | |
| Standard deviation of datastreams per object | TBD (see note)TIF datastream not included in calculation | TIFF datastreams: .674 XML MIX datastreams: .674 | |
| Average size of datastream | see note | TIFF: TBDXML: TBD 16461985 bytes (16 MB) MODS: 2794 bytes MIX: 2804 bytes PREMIS: 817 bytes | |
| Standard deviation of datastream size | see note | TIFF: TBD XML: TBD : 14259792 bytes (14 MB) MODS: 474 bytes MIX: 15 bytes PREMIS: 200 bytes | |
| Average number of properties per node | n/a | 1 | fcr:accessroles property for authorizationno properties set |
| Standard deviation of properties per node | n/a0 | ||
| Is datastream data random? | Yes | Each datastream is unique; there are no constant value datastreams. |