Page 150 - Building Digital Libraries

Metadata Formats

importing tends to be problematic because it assumes that metadata is
clean, has been delineated correctly, and is laid out in the file system in a
very specific structure. Digital library systems like DSpace and Fedora that
provide native support for METS objects greatly simplify the import process
by utilizing a highly structured metadata framework in which all metadata
about the object is self-contained.

METS Today
There was a time when METS was the gold standard for encoding structural
data within a digital repository, but that sentiment is quickly changing.
While METS still has a place within the digital library landscape, and likely
will for some time, a transition has begun to move to more lightweight,
semantically relevant data. While METS provides a good deal of flexibility
and structure, as an XML language, it is very expensive to develop against.
Just as systems have moved to embrace linked data and semantic web prin-
ciples, a shift is occurring within the computer science landscape, moving
away from XML processing to more lightweight languages—particularly
JSON. We can see this shift occurring within the digital library and digi-
tal humanities communities, as more and more systems move away from
directly supporting METS and instead support formats like IIIF. Like METS,
IIIF provides methods for creating and defining the structure of an object,
but it does so utilizing JSON, a lightweight format that can be used by devel-
opers, and can be easily manipulated with most mainstream programming
tools. This allows data to be unlocked and easily consumed by those outside
of the library, making it an attractive data format.

IIIF

The International Image Interoperability Framework (IIIF) is a grassroots
community standard that was originally developed to provide a set of best
practices for sharing and interacting with images. Today, IIIF currently
provides multiple “metadata” streams, including the original specifica-
tion for image display and manipulation, but it has expanded to include
specifications for defining the presentation of an option, authentication
methodology, search, and annotation. When compared to the other primi-
tives highlighted in this chapter, IIIF is a bit of an outlier. Like METS, it’s a
structure that is largely machine-generated—in that the development and
utilization of IIIF is largely system-generated in order to support the shar-
ing of works.

History
The history of IIIF is tied up in the complicated history of displaying
images within digital library systems. Digital library systems have tra-
ditionally developed their own internal image viewers, utilizing propri-
etary formats (like MR.SID) or readers to interact with content. While
JPEG2000 (an open-wavelet image format designed to support advanced
135