Roadmap to Digitization notes from ALA

On Saturday afternoon, I attended this session which - while billed as "a roadmap for digitization for smaller institutions" still taught me a lot. The speaker was Jacob Nadal, the Preservation Officer at the UCLA library.

There was a lot of background information on file formats & metadata. The parts that seemed most pertinent to our current questions/concerns about digital collections at the State Library included the following recommendations/best practices (though there is no specific order here):

1. Digitization efforts require considerable technical expertise from an IT perspective, so it should be managed by an IT professional
2. Avoid organizing your files by the online presentation or by projects (because digital objects will be reused in ways you cannot yet anticipate)
3. Avoid naming with filenames - use a numeric pattern/accession # (or if it's materials from archives, if there is a collection #/ serial #, use that and you can tack onto the end _Master.tif). If you NAME files, you are adding descriptive metadata into the naming scheme for the file, which is a bad idea (see again the note on digital objects being reused in ways you cannot yet anticipate).
4. Use XML for metadata (because it is a stable format for digital preservation) - can be any schema - but this is what added the essential piece for digital library functionality; you should include metadata specialists from the outset of a digitization project
5. VRA is a good XML schema for digital objects with visual characteristics.
6. Good enough is better than highly detailed/perfectionistic approach to object and its metadata
7. There are certification requirements that "trusted" digital repositories should meet (see OAIS - RLG prepared the certification criteria) & that repositories should meet even if they are not officially certified as "trusted". OCLC's digital archives, and a number of other digital repositories are certified)
8. For digital preservation, many grassroot groups form their own private, local LOCKSS networks - the software is free
9. Digital asset management is probably overkill (& added complexity) for most libraries. It would include managing metadata & materials, but it adds considerably more functionality than that. Digital asset management systems are used by advertising companies, for example, who have to monetize digital assets & control them more strictly for IP & versioning purposes. (I specifically asked about this due to our interest in this type of system and that he hadn't mentioned it in the presentation).
10. Public access to digitized materials isn't the same as digital preservation. He uses Wordpress for the site he runs to offer people public access to digital collections. (Greenstone & Omeka were also mentioned for presentation layer, but he didn't go deeply into this area).
11. For color calibration, you should buy a spectrometer - it's just a couple hundred $ and lasts several years. Trying to get color calibration by eye is a huge mistake. In fact, a study showed that color gets worse the more an "expert" in imaging looks at it - so use the device, it's impartial. Calibrate at least 1 x / week.

His presentation is now up on his site at http://www.jacobnadal.com/84.
A key resource he pointed people to was PREMIS - http://www.loc.gov/standards/premis/ - (PREservation Metadata: Implementation Strategies).