Page 61 - Building Digital Libraries

CHAPTER 3

(a surprisingly common request by donors is to suppress items forever) or
simply confuse people with too many different statements—which prevents
the rights from functioning as needed.
Rights and access management is a complex topic discussed in chapter 8.

Protecting the Integrity of Resources

Ensuring the integrity of resources is an essential part of acquiring and
processing items. This issue is related to but separate from preservation,
which is discussed in chapter 4. The idea of preserving integrity is simple—
namely, ensuring that the resources and the content they contain are still
usable after being added.
Operationalizing integrity is less straightforward because the goal of
integrity is to replicate an experience with a resource rather than preserve
a digital artifact as one would preserve a physical resource. Digital items
differ from physical items in that the former don’t exist in the normal sense
of the word. A file typically takes the form of magnetic or electrical charges
and often doesn’t even occupy a contiguous area on a disk or in memory.
There is no such thing as moving a file—it can only be copied from one
place to another. Each time this happens, the bitstream is broken apart and
modified as it is encrypted, mixed with transport protocol information, and
recombined when rendered through software. By definition, any software
used to view a document formats the information it is given.
To function, all repositories require that individuals have appropriate
permissions to add and modify objects at certain points in the process. The
ingestion process needs to allow legitimate reformatting and modification
of resources while preventing illegitimate modification.
Many repositories employ checksums to help ensure the integrity of
resources. However, the following should be kept in mind about checksums:

• Checksums only help identify modified files. They cannot
help repair modified files except for purposes of identifying
other copies that have not been modified.
• If checksums are kept with the objects, they only help iden-
tify in ad ver tent modification, since those who intentionally
modify files can simply provide new checksums.
• Any legitimate modification of a file, such as creating a
preservation copy or modifying the internal metadata,
requires generating a new checksum.
• Checksums provide no additional value if files are trans-
ferred to an area where they are protected from modifica-
tion, and de letion and transfers are validated with digital
signatures.
• Regularly comparing checksums incurs significant disk
activity that impairs performance.
46