Digitization Project With an Incomplete Index
An archive is digitizing 8,000 case files from a social services agency. The paper files have a typed index card for each case, but about 15 percent of the cards are missing or illegible. The vendor proposes scanning the files in box order and creating a searchable index from OCR of the case documents themselves. The archive's finding aid currently describes the series only at box level. A researcher has asked for all cases involving a particular family name, and the archive cannot yet answer reliably. The project budget covers scanning and OCR but not manual re-indexing of the 1,200 problem cases. The agency that created the records is still the legal owner and has not approved public release of the digitized images.
1. How should the archive weigh the vendor's OCR-index proposal against the access goal?
Read the worked answer
The archive should recognize that OCR of case documents can supplement but not replace the intellectual control the index cards provided. OCR will produce text, but it will not reliably reconstruct the missing metadata fields (names, dates, case numbers) that the cards carried, and OCR errors in handwritten or poor-quality documents will create false negatives. For a researcher seeking a family name, false negatives are a serious access failure.
The archive should scope the OCR as a discovery aid, not as the authoritative index, and should prioritize manual work on the 1,200 problem cases where the card is missing or illegible. If the budget cannot cover all of them, the archive should sequence: index the problem cases first, or at minimum create a documented list of unindexed cases so searches can be qualified. Presenting OCR output as a complete index would misrepresent coverage and mislead researchers.
2. The creating agency has not approved public release. What should the archive do about access during the project?
Read the worked answer
The archive should separate preservation digitization from public access. Scanning for preservation and internal description can proceed if the agreement with the agency permits it, but public release of images requires the owner's approval. The archive should confirm the terms in writing before any images are published, and should not let a researcher's request pressure it into releasing material the agency has not cleared.
Operationally, the archive should keep digitized images in a restricted environment, document the access status of each series, and tell the researcher what is available and what is pending. The archive should also clarify whether the agency's approval covers metadata and finding-aid descriptions, which may be releasable even when images are not. The judgment is to advance the project without creating an unauthorized release.