Caveat Lector

NCSU: working with geospatial data

Repository requirements
- dim archive, possible future access
- minimal repository imprint on data (import/export should be repository-agnostic; avoid dealing with repository’s organization schemas)
- simple digital curation functions (checksums, structured metadata index)
- be able to exchange data
- leverage existing tech investments
- OSS with active community

Automation
Python wrappers for:
- antivirus (ClamAV)
- file compression
- JHOVE
- GIS data handling

Other tricks
- NOID for identifiers
- Metadata capture system that minimizes user error

Data tripping about from here to there, lots of transformations to write.

Extra-repository AIP management (make sure to be able to recreate everything if DSpace dies). Built their own tool to do this.

Five DSpace instances running on two Tomcats and a single PostgreSQL; each has its own space in their 15 TB assetstore.

Upcoming projects
- enhancements to current system (XTF for search, inter-archive exchange)
- Digital Collections Repository (special collections, faculty)
- Data repository (scientific data, statistical resources)