CNIB Meta Data

From zedwiki

Background

CNIB developed its own set of meta data for intenal use while designing a master xml storage format from which to produce braille, DAISY 2.02 and CBA E-text files.

Many existing meta data standards were looked at during development, but no single standard could be found to accomplish what was needed. In some cases, particularly with copyright information, braille requirements did not match the intended use of fields in existing standards, making it impossible to correctly implement.

Because of the highly grab-bag approach that was going to be necessary, the schema that we implemented does not directly use elements from any other standard, but has kept the naming of elements and their usage as close to Dublin Core and MODS as possible, to facilitate transformations to formats where that meta data is a requirement.

Another consideration during development was where to store the meta data. As meta data is information about the work, embedding it in the content file was seen as a less elegant solution, as it requires reading the content file in order to find out about it. A future goal is also to be able to generate the meta data as the cataloguing records are created, to avoid duplication and errors having editors input the data again once the xml has been generated. The meta data was conseqeuently split off into a separate file, which in turn allows us to store more information about the content file's format.

Structure

Our meta data is split into two sections:

1. information about the work and its publisher; and
2. information about the producer and the master file.

The information about the work and publisher is the more widely recognized of the two, and contains elements for the title, subtitle, edition, series, print page range(s) reproduced (extent), creator(s) (authors, illustrators, etc.), copyright and publisher.

The section for the producer and master contains institution-specific and technical information, such as the type of work ('general', 'cookbook', etc.), creation date, mime type, language, encoding, as well as the internal accession codes and isbns for the output formats.

Example

A typical meta data file looks like the following:

<?xml version="1.0" encoding="utf-8"?>
<meta xmlns="http://www.cnib.ca/XML/2008/meta">
	<source>
		<title>Dragon Moon</title>
		<extent>1-328</extent>
		<creator>
			<name role="author">Carole Wilkinson</name>
		</creator>
		<copyright>Copyright © 2007 by Carole Wilkinson</copyright>
		<publisher>
			<name>Hyperion Books for Children</name>
			<location>New York, New York</location>
			<identifier type="isbn13">9781423111436</identifier>
		</publisher>
	</source>
	<producer>
		<name role="accessible format producer">CNIB Library</name>
		<master>
			<type>general</type>
			<date scheme="yyyy-mm-dd">2008-10-25</date>
			<format>application/xml</format>
			<language scheme="ISO 639">en</language>
			<encoding>utf-8</encoding>
		</master>
		<output type="braille">
			<id>BR74682</id>
		</output>
	</producer>
</meta>