CR 1.8 UCSD CHECKSUMTYPE CV list addition

Approved

The University of California, San Diego Libraries’ is currently in the process of depositing digital content for preservation in the UC Libraries Digital Preservation Repository (DPR). The DPR is maintained by the California Digital Library (CDL) to enable the long-term storage of digital objects.

We have asked that the CDL/DPR provide checksum validation, and validation is now a service provided by them to ensure that what we send there is actually what we think we sent there. The METS schema does not support CRC32 as a checksum type, but the CDL does as it is one of the codes used by JHOVE. The CDL has committed to supporting the following checksum types for validation: MD5 | CRC32 | SHA-1

We have consistently captured CRC32 for our 250,000+ digital objects and because CRC32 is a fairly common method for checksums, we hope that you agree that it should be supported.

The CDL has instructed us to embed CRC32 values in mets:file in the following manner:

<mets:file ID="ark-20775-bb3345557m-1-1" MIMETYPE="image/tiff" CHECKSUMTYPE="CRC32" CHECKSUM="0fb6e126" SIZE="30283702" CREATED="1999-07-22T06:42:30" ADMID="ADM1 ADM2" GROUPID="GID1">

<FLocat xlink:href="http://libraries.ucsd.edu/ark:/20775/bb3345557m/1-1.tif" LOCTYPE="URL"/>

</mets:file>

Please note that our checksum are indeed captured in PREMIS in our METS files. I suspect that not all UC institutions make the use of PREMIS in their METS files and that is the reason they’ve implemented the use of checksum and checksumtype in the mets:file. Thus, for DPR checksum validation purposes, we repeat what is in PREMIS in mets:file

Solution:

Element

CHECKSUM
CHECKSUMTYPE must contain one of the following values:
CRC32
HAVAL
MD5
SHA-1
SHA-256
SHA-384
SHA-512
TIGER
WHIRLPOOL

Thus, we kindly request that CRC32 be added as an optional attribute for checksum algorithms.

Requested by UCSD, Gabriella Montoya, [email protected]