We generate an MD5 hash for all files created as part of the
preservation transfer process. We have learned to create this "checksum"
immediately after a file is created after a colleague's experience with
a hard drive with corrupt sections. The MD5 hash is created through our
audio technical metadata collector (ATMC)when parsing the file to
collect and store metadata about it. ATMC is software developed as part
of the Sound Directions project and will eventually be released
publicly. We also use proprietary software called FastSum to verify
checksums for individual files and batches of files as we don't have
this capability within ATMC yet. We also generate checksums for our
AES31-3 ADL files which carry edit and marker metadata. We currently
store the MD5 files in the same folder as the audio and ADL files but
will probably not keep them when files move to preservation packages and
to long-term storage since ATMC stores the checksum value. But, we're
not quite there yet so we'll see.
The ARSC pre-conference workshop in Seattle will explore these and many
other issues relating to preservation in the digital domain. Further
information at http://www.arsc-audio.org/workshop2006.html
Mike
----------
Mike Casey
Associate Director for Recording Services
Archives of Traditional Music
Indiana University
(812)855-8090
Co-chair, ARSC Technical Committee
www.dlib.indiana.edu/projects/sounddirections/