Page 137 - Building Digital Libraries

CHAPTER 6

Many technical limitations exist that make MARC a poor choice for
a digital repository. First and foremost, as a metadata schema, MARC is
virtually unknown outside of the library community. This fact limits the
number of available partners or vendors that can be used to create or sup-
port systems utilizing the record format. Moreover, it limits data-sharing to
primarily the library community. Second, the record-size restrictions limit
the amount and type of bibliographic content that can be provided. This
is generally quite obvious when one is looking at a MARC record derived
from a more robust XML language; the fields dealing with descriptions and
abstracts are often truncated or broken into multiple related fields in order
to overcome the field length limits. Within the XML formats to be described
later in this chapter, records can be created containing a much richer set of
data, including large data elements like tables of contents, abstracts, sum-
maries, or even the full text of an item. In addition, within XML formats,
rich content like images and documents can be embedded into the metadata
themselves, turning the metadata records into information objects.

MARC21XML

As one might guess, MARC21XML is a lossless XML representation of
MARC, with one notable exception. Unlike MARC, which supports the
MARC-8 character set, MARC21XML exclusively supports the UTF-8
character set, meaning that character data in MARC-8 must be rectified
to UTF-8. MARC21XML was first proposed by the Library of Congress
in 1994 as one potential method for migrating MARC data into an XML
schema. However, MARC21XML has found only sparse use within the
library community—and primarily as a data crosswalking mechanism for
MARC between other metadata schemas. For example, MarcEdit, a free
8
application written by Terry Reese, utilizes MARC21XML as the control
metadata schema when moving data between differing library metadata
schema. Figure 6.3, for example, is a MARC21XML representation of the
sample record found in figure 6.1. This representative record was generated
using MarcEdit.
Looking at figure 6.3, one can see that MARC21XML is an XML-based
metadata format; because of this, few of the technical restrictions relevant to
MARC are inherently present with MARC21XML. For example, MARC21-
XML has no predefined record size limit, meaning that users can embed
large data elements into the XML record. Moreover, like all XML-based
metadata schemas, MARC21XML can have data-like images or documents
embedded directly into the metadata.
While MARC21XML does not contain the same limitations of MARC at
the technical level, these limitations do exist. However, they are limitations
imposed due to the requirement that MARC21XML be a lossless represen-
tation of MARC. For MARC21XML to be utilized as initially conceived,
it must inherit the same technical limitations as the MARC framework if
compatibility between the two formats is to remain.

122