Page 119 - Building Digital Libraries

CHAPTER 5

not just the structure, but the content of the data provided. However, for
individuals working as developers of library systems, JSON will likely be
the primary currency that they will use to move data between various parts
of the application. This doesn’t mean that the XML won’t be visible or uti-
lized; it just means that XML won’t be the standard currency by which data
is transferred within an application. And this transition at the developer
level from XML to JSON will be welcome, since it follows the predominant
patterns for application development outside of the library community,
expanding the potential pool of tools and expertise that can be called upon
as cultural heritage institutions develop new applications and platforms.

Data Manipulation

It would be impossible to write about the full breadth of topics surround-
ing data manipulation. When working with metadata, transferring data
between systems, and upgrading digital platforms, data manipulation will
always be required. Even when restricted to digital libraries, the topic covers
so many different formats, technologies, and languages that it is easy to be
overwhelmed. It’s a big topic, and there are books specially dedicated to
13
the topics and techniques that are required to efficiently move data between
formats and systems, as well as to identify and understand the challenges
associated with the manipulation of data.
Given the breadth of the topic of data manipulation, how does one even
prepare for the future to meet this inevitable challenge? The best advice is
to build up a data toolkit; that is, to identify a set of languages, tools, and
strategies that one might use to perform specific data manipulation tasks,
and become familiar with them. The truth is that many developers, metadata
specialists, and others will have specific tools and languages that they cannot
live without. And just like operating systems and browsers, the advocacy for
one tool or language can take on a kind of religious fervor. But the reality
is that data manipulation, like any task, can be done in a wide variety of
ways and through a wide range of workflows. There really isn’t a best way
or tool to do it. There are certainly many wrong ways, but no one best way;
this means that for the most part, the tools or languages that one chooses
to learn are often less important than one’s ability to understand and master
the resource. The following subsections describe just a small selection of
potential tools and services that libraries should consider evaluating as they
develop their own personalized toolkits.

Programming Languages

Programming language selection often comes down to one of two things:
personal choice or organizational support. By and large, most program-
ming languages provide the same toolkits and functionalities. Differences

104