[H O M E : Emblem Project Utrecht]

Emblem Project Utrecht

EPU Implementation of The OAI-PMH

1 Introduction

This document describes the way the Emblem Project Utrecht has implemented the Open Archives Initiative, a protocol for Metadata Harvesting.

Our implementation is experimental, and is meant as a first step towards exposing our emblem data to the wider world. More specifically, we tries to integrate our data with those of other emblem digitisation projects, most notably the German Emblem Book Database at the University of Illinois, Urbana-Champaign.

This document does not discuss the Open Archives Initiative or the OAI-PMH protocol. More about the protocol can be found on the OAI websiteLinks to page outside of EPU. The document also does not discuss the Dublin Core Initiative’s set of metadata. More about Dublin Core Metadata Initiative (DCMI) can be found at the DCMI websiteLinks to page outside of EPU.

THe EPU’s implementation of AOI-PMH uses unqualified Dublin Core metadata. This is just a first step towards a more structured exchange of metadata. An XML Schema, which will facilitate OAI-PMH compliant exchange of detailed emblem metadata, is under development at the Herzog August Bibliothek in Wolfenbüttel. The Emblem Project Utrecht is a partner in the 'Open Emblem' Initiative which coordinates the emblem metadata integration efforts being done in individual emblem digitisation projects.

The main problem in implementation is that it never clearly chooses between providing metadata for the emblem, or for the modern version. We provide, for instance, two publisher elements: a publisher for the original book, and a publisher of our own edition. There may even be a third publisher element, if our own version is based on a modern edition of the book (as in the case of Cats or Hooft).

The first section of this document discusses how we use the 15 Dublin Core elements. The second section discusses the technical implementation.

Table of contents

2 Dublin Core Elements

2.1 General Remarks

The form of the personal names occurring in the Dublin Core (DC) elements is the one used in the EPU bibliography.

Dates given, use the contents of the value attribute of the date-element, if present, or else the element’s text content.

Where multiple DC elements of the same type were needed, multiple corresponding XML elements to hold the information, were used. This was done, rather than use a single XML element as well as comma’s or semi-colons to distinguish the individual DC elements.

2.2 Title

We filled the DC title-element with the list of titles for the emblems and preliminary texts, this is discussed inTitles. This list generally contains a lower-case version of the motto.

2.3 Creator

We created a DC creator-element for each of the original authors of the book. We do not distingiush between the primary author (such as Vaenius), the secondary author (such as, Alonso de Ledesma, Amoris divini emblemata), or the artist of the pictura. These elements are created for each of the texts in a book, even if they are not emblem texts. We may refine this procedure in the future.

For the liminary texts, an additional creator-element is used to hold the contents of the signed or docAuthor-elements, if present.

2.4 Subject

At present, the only DC subject-elements that we created are those for the pictorial motifs.

2.5 Description

A DC description-element is created for each of the emblem texts. Each quotation, epigram, or motto is represented in a separate description-element. At present, we have ignored the bibliographical references given by the emblem author or by the modern editors. Each description-element carries a xml:lang attribute to represent the language of the text. The language codes are based on ISO 639-1Links to page outside of EPU.

2.6 Publisher

A DC publisher-element is created for each of the publishers relevant to the emblem, as represented on the EPU site. So the original publisher, the EPU itself, and possibly the publisher of a modern edition on which the EPU has been based, all have elements.

2.7 Contributor

At present, the DC contributor-element is used to acknowledge the contribution of modern translators (Dutch or English). For 17th century translators, we use the creator-element.

2.8 Date

The way we use the DC date-element corresponds to the way we use the publisher-element. The date of publication of each of the relevant editions is given.

2.9 Type

We created DC type-elements with values for text and image, this was taken from: Links to page outside of EPU).

2.10 Format

We created DC format-elements with values text/xml, text/html and image/jpeg; these are the formats which are accessible from a resource, that the identifier-element connects to.

2.11 Identifier

The emblem URL is used as contents for the DC identifier-element.

2.12 Source

We created two types of DC source-element. One is filled using just the call number of the copies, which form the basis for our editions. The second form contains a very simple bibliographic description.

2.13 Language

For each language used in the text, a DC language-element is created (using ISO 639-1Links to page outside of EPU codes).

2.14 Relation

We created DC relation-elements to relate the emblem to the full book, for each image file (facsimile) related to the present emblem, and for each related emblem. URL’s are used to identify related emblems, which may be from the same book, from elsewehere on the EPU site or from elsewhere on the Internet.

We are still discussing the feasability of adding relation-elements for other sources, such as, quotations from the Bible or Ovid.

2.15 Coverage

Not in use.

2.16 Rights

Not in use.

3 Technical Implementation

The EPU’s technical implementation of OAI is very simple. For each emblem or liminary text, a file is generated which contains the metadata for that text. We use the XMLFileLinks to page outside of EPU OAI software written by Hussein Suleman at Virginia Polytechnic Institute and State University, to provide the Open Archive data provider functionality. XMLFile uses the generated files as the metadata source. The generated files are stored in directories whose names correspond to the book abbreviations we use at EPU (ca1627, va1608, etc). The directory names double as OAI set names.