Muninn Documents Ontology Specification - 0.01

The Muninn ontology to specify and work with documents.

Working Draft — 19 February 2012

This version:

http://rdf.muninn-project.org/ontologies/documents-0.01.html (owl)

Latest version:

http://rdf.muninn-project.org/ontologies/documents.html (owl)

Last Update: 0.01

Date: 19 February 2012

Authors:

Robert Warren, The Muninn Project

This work is licensed under a Creative Commons / Attribution-ShareAlike 3.0 Unported License.

Abstract

This specification defines a set of classes and properties used to represent document and their structure.

Status of this Document

This document is based on the current practices used by The Muninn Project to represent documents and digitally scanned images within the catalog.

Table of Contents

1. Introduction
   1. Ontology Objectives
   2. Ontology Limitations
2. Graves ontological constructs at a glance
3. Ontology use cases
4. Cross-reference for Graves classes and properties
5. Conclusion and future work

Introduction

This project was created as a result of some of the data and modeling problems encountered in the Muninn WW1 Project. The ontology records the bibliographic, provenance and digitization infromation of archival documents. It provides classes that are subclasses of FOAF for better compatibility while adding some needed functionality to manage documents.

The major differences with core FOAF classes has to do with properties that support document pages, digital representation of each of the pages, digital rights support and support for forms.

Ontology Objectives

The initial design objectives for the ontology were:

1. The ability to markup a document with the minimum amount of structure required for analysis.
2. Provide some provenance hooks for the source archive and re-distribution infromation.
3. Provide linkages between the original document and its transcribed and transformed form.
4. The ability to identify certain pages of the document as a filled out form.

Ontology Limitations

The ontology is a limited representation of the full document structure in that it represents the physical parts of the document and not its logical organization. The properties for a page reflect this in its previous_page and next_page organisation.

Ontological constructs at a glance

Class organization revolves around a Document class that is the union of the FOAF Document class and the Creative Commons Works class for maximal compatibility. A Page class provides a container for each page of the document who's digitized image is in turn provided by an Image subclassed from the FOAF Image class. The Form class is another subclass of a Document (as some forms tend to be full blow documents themselves) but with additional properties intended to document their use and issuing organization.

The ontology makes extensive use of subclassing because of the different layers of a document whose different representation require multiple layers of meta-data. For example, which some works are clearly in the public domain some archival organization claim copyright on the photograph of the document itself. This allows the representation of situations where the contents of the document have been transcribed / OCR'd from a source image that is not publically available.

Documents, Document Models and Representations.

The markup described within this ontology is a document which models a physical document whose physical pages are represented by a set of images. To avoid confusion we use the term modeled document to reference the complete physical document while a representation references a single digital image of a part of the physical document.

Document copyright, access control and licensing

The handling of multiple sources of documents sometimes requires different levels of access controls for different representation of the document. It is common for a historical document to be owned by a museum, its contents to be out of copyright while the images of the document be copyrighted by the photographer.

This ontology provides the combined properties of FOAF, CC Works and some properties borrowed from Dublin Core. Together they provide a powerful description markup for the different parts of the document. The initial properties on the document are:

• publisher The publisher of the document.
• author The author of the document.
• source The source of the document in terms of a ressource or an organisation.
• date_published The date of publication.
• date_created The date on which the document was created.
• date_copyrightedThe date on which the document was copyrighted; see the property notes.

These properties are the cornerstone for the definition of the ownership (and authority) over the document. The date information is important in that it defines the begining of the countdown clock after which the copyright expires.

Within the Creative Commons Works class the following properties are available:

• jurisdiction The jurisdiction under which this works falls. There is currently no guidance as to what a jurisdiction really is.
• licenseThe license under which this document is distributed. Currently the known instances of the License classes are:
  • Reproduction
  • Distribution
• legalcodeA link to a document explaining the legal status of the document.
• prohibitsA series of prohibitions for this document.
• permitsA series of permissions for this document.

The CC Work class makes use of these last two properties which are inverse of each other. The ontology also has an additional property derived from Dublin Core accessRights which combines both properties into a single directive.

Currently the access control and rights properties of the ontology are only used to markup the information. However, a future direction is the use of this information to infer using the reasonner the appropriate distribution and access rights based on the date, location and original licensing of the document.

Documents ontology overview

A few examples are presented here

3.1. Example

Here is a very basic document describing a foo:

        <documents:Image rdf:ressource="http://rdf.muninn-project.org/ww1/2011/11/11/image/34">
         <documents:size_bytes>39475</documents:size_bytes;>
         <documents:x_pixels>256</documents:x_pixels>         
         <documents:y_pixels>256</documents:y_pixels>         
         <documents:format>image/png</documents:format>         
         <documents:thumbnail>
          <documents:Image rdf:ressource="http://rdf.muninn-project.org/ww1/2011/11/11/image/34/thumbnail"/>
         <documents:thumbnail>
         <!-- Wikimedia isn't on dbpedia -->
         <documents:source rdf:ressource="http://commons.wikimedia.org/wiki/File:Crystal_Project_My_documents.png"/>
         <documents:date_retrieved rdf:datatype="http://www.w3.org/2001/XMLSchema#date>2012-02-12<documents:date_retrieved>
         <documents:creator rdf:ressource="http://dbpedia.org/page/Everaldo_Coelho"/>
         <documents:license rdf:ressource="http://rdf.muninn-project.org/ontologies/documents#lgpl"/> 
         <documents:date_published rdf:datatype="http://www.w3.org/2001/XMLSchema#date>2007-06-16<documents:date_published/>
         <documents:description>An icon from the Crystal Project icon theme.</documents:description>
         <!-- Webserver will serve copy of picture to anyone. -->
         <documents:accessRights rdf:ressource="http://creativecommons.org/ns#Distribution"/>
        </documents:Image>
      

Muninn Specific Implementation details

The Muninn RDF makes extensive use of header content negotiation in HTTP headers to provide the end user with exactly the type of content that is required. There are a few extensions that are available to force certain behaviours on the part of the server.

RDF representation only

While a sparql server will allways return the appropriate content to a sparql client, it is sometimes convinient to force the server to send only the RDF metadata contents.

      http://rdf.muninn-project.org/ww1/2011/11/11/image/34/about.rdf
     

Image in original format

While a sparql server will allways return the appropriate content to a sparql client, it is sometimes convinient to force the server to send only the RDF metadata contents.

      http://rdf.muninn-project.org/ww1/2011/11/11/image/34/image
     

Thumbnail in original format

A 130x100 thumbnail image can be requested in the original format by requesting this specific name.

      http://rdf.muninn-project.org/ww1/2011/11/11/image/34/thumbnail
     

Image in specified format

The image can be requested in a specific graphics format by requesting it using the appropriate extension (eg: image.jpg for JPG, image.png for PNG, etc...). This conversion is done automatically if the http headers request a specific format.

      http://rdf.muninn-project.org/ww1/2011/11/11/image/34/image.ext
     

Thumbnail

A 130x100 thumbnail image can be requested in a specific graphics format by requesting it using the appropriate extension (eg: image.jpg for JPG, image.png for PNG, etc...). This conversion is done automatically if the http headers request a specific format.

      http://rdf.muninn-project.org/ww1/2011/11/11/image/34/thumbnail.ext
     

Cross-reference for classes and properties

Conclusion and Future Work

The ontology is a limited representation of an archival document.