Developing a Toolkit for Digital Epigraphy

Developing a Toolkit for Digital Epigraphy

Hugh
Cayless

UNC Chapel Hill
hcayless@email.unc.edu

2003

University of Georgia

Athens, Georgia

ACH/ALLC 2003

editor

Eric
Rochester

William
A.
Kretzschmar, Jr.

encoder

Sara
A.
Schmidt

This paper will present the results of work being done at UNC Chapel Hill to
support the emerging EpiDoc standard for encoding epigraphic texts. The EpiDoc
standard is based on the Text Encoding Initiative guidelines (). Information on this collaborative effort,
which includes scholars and humanities computing experts from North America and
Europe, is available at . The
mission of the project is to develop "a software and hardware-independent
interchange specification for scholarly and educational editions of inscribed
and incised texts in Greek, Latin and other languages emanating from the ancient
Greek, Roman and nearby civilizations." The Chapel Hill team has focused on
building tools to facilitate epigraphers' work on creating EpiDoc texts and
managing the storage and presentation of those texts. The paper will begin with
a brief outline of the EpiDoc guidelines with examples of their implementation
and a discussion of some of the problems inherent in digital epigraphy. Then the
various tools developed by the Chapel Hill team will be demonstrated.
The editing and publication of inscribed documents has a rich history, which
includes the adoption of widely accepted standards for "marking up" texts.
Published Greek and Latin epigraphic texts are almost universally presented
according to the "Leiden convention,"developed in 1931 at the 18th International
Congress of Orientalists by papyrologists. The Leiden convention employs various
symbols to present information about the condition of the inscribed text and
editorial supplements and comments. For example, parentheses surround editorial
expansions of abbreviated words, so "AVG" on the support would be expanded as
"Aug(ustus)" by the editor. In EpiDoc, the same text would be written
"Aug<expan>ustus</expan>." Because of the widespread
adoption of the Leiden convention, the form of published texts (allowing for
some local variations) is remarkably consistent. Because of this consistency, it
is possible to write a parser that can translate a digitized inscription into a
data structure (such as XML). The Chapel Hill team has developed such a parser,
named the Chapel Hill Epigraphic Text Converter (CHET-C). This tool parses texts
marked up according to the Leiden convention and outputs EpiDoc XML. Both MS
Access and Java versions are being developed.
Digital texts in polytonic Greek pose problems of their own because of the
various methods which evolved to encode them prior to the advent of Unicode. One
of the earliest of these was Beta Code, which utilizes 7-bit ASCII to represent
Greek, Latin, and Aramaic texts, as well as various epigraphical and
papyrological sigla. In addition, a number of Greek fonts, each employing its
own specialized encoding, were developed over the years to handle the problem of
typing and viewing Greek with computers. There exist already various programs
and filters which can handle shifting the encoding of texts from one form to
another (the Greek texts on employ
such filters, for example). There was still a need for a more flexible system to
handle encoding shifts, however, and so the Chapel Hill team (in collaboration
with the Stoa Consortium, ) has developed
Java-based software that performs this function and which can easily be plugged
into the larger toolkit framework. The transcoder classes can perform encoding
shifts on text within specific elements of an XML file, and are also designed to
be modular and easily extensible to handle languages other than Ancient
Greek.
The team has also developed a web-based framework for the presentation of EpiDoc
texts. The framework has been built with Apache Cocoon () and is hosted by the Stoa. The
"Epidocinator" as it is called, can dynamically transform EpiDoc documents
against a variety of XSL stylesheets. Both documents and stylesheets may be
either on site or remote. The framework will also perform validation and error
reporting on EpiDoc texts. The transcoder classes will be accessible via a
Cocoon Transformer, so that Greek text can automatically be shifted from any
supported source format to any supported result format. It will employ the Java
version of CHET-C in a custom Cocoon Generator or Transformer to allow users to
generate EpiDocs from already-digitized inscriptions. The framework will be
packaged as a Web Archive (WAR), so that interested developers can easily deploy
their own digital epigraphy frameworks.
Finally, the team is also working to integrate the Java tools with jEdit, a
Java-based text editor with a very robust plugin architecture (). The goal is essentially to create an
Integrated Development Environment (IDE) for epigraphers which editors will be
able to use in the electronic publication of EpiDoc texts. All of the tools and
stylesheets developed for use with EpiDoc will be available as jEdit plugins and
editors will also have available jEdit's XML and XSLT plugins, which provide
useful functions like code-completion, tag creation help (based on a DTD),
validation, error reporting, and transformation. The combination of the web
framework and the text editor will provide epigraphers with a cross-platform,
Open Source solution for creating and publishing inscriptions.
This paper will present the results of work being done at UNC Chapel Hill to
support the emerging EpiDoc standard for encoding epigraphic texts. The EpiDoc
standard is based on the Text Encoding Initiative guidelines (). Information on this collaborative effort,
which includes scholars and humanities computing experts from North America and
Europe, is available at . The
mission of the project is to develop "a software and hardware-independent
interchange specification for scholarly and educational editions of inscribed
and incised texts in Greek, Latin and other languages emanating from the ancient
Greek, Roman and nearby civilizations." The Chapel Hill team has focused on
building tools to facilitate epigraphers' work on creating EpiDoc texts and
managing the storage and presentation of those texts. The paper will begin with
a brief outline of the EpiDoc guidelines with examples of their implementation
and a discussion of some of the problems inherent in digital epigraphy. Then the
various tools developed by the Chapel Hill team will be demonstrated
The editing and publication of inscribed documents has a rich history, which
includes the adoption of widely accepted standards for "marking up" texts.
Published Greek and Latin epigraphic texts are almost universally presented
according to the "Leiden convention,"developed in 1931 at the 18th International
Congress of Orientalists by papyrologists. The Leiden convention employs various
symbols to present information about the condition of the inscribed text and
editorial supplements and comments. For example, parentheses surround editorial
expansions of abbreviated words, so "AVG" on the support would be expanded as
"Aug(ustus)" by the editor. In EpiDoc, the same text would be written
"Aug<expan>ustus</expan>." Because of the widespread
adoption of the Leiden convention, the form of published texts (allowing for
some local variations) is remarkably consistent. Because of this consistency, it
is possible to write a parser that can translate a digitized inscription into a
data structure (such as XML). The Chapel Hill team has developed such a parser,
named the Chapel Hill Epigraphic Text Converter (CHET-C). This tool parses texts
marked up according to the Leiden convention and outputs EpiDoc XML. Both MS
Access and Java versions are being developed.
Digital texts in polytonic Greek pose problems of their own because of the
various methods which evolved to encode them prior to the advent of Unicode. One
of the earliest of these was Beta Code, which utilizes 7-bit ASCII to represent
Greek, Latin, and Aramaic texts, as well as various epigraphical and
papyrological sigla. In addition, a number of Greek fonts, each employing its
own specialized encoding, were developed over the years to handle the problem of
typing and viewing Greek with computers. There exist already various programs
and filters which can handle shifting the encoding of texts from one form to
another (the Greek texts on employ
such filters, for example). There was still a need for a more flexible system to
handle encoding shifts, however, and so the Chapel Hill team (in collaboration
with the Stoa Consortium, ) has developed
Java-based software that performs this function and which can easily be plugged
into the larger toolkit framework. The transcoder classes can perform encoding
shifts on text within specific elements of an XML file, and are also designed to
be modular and easily extensible to handle languages other than Ancient
Greek.
The team has also developed a web-based framework for the presentation of EpiDoc
texts. The framework has been built with Apache Cocoon () and is hosted by the Stoa. The
"Epidocinator" as it is called, can dynamically transform EpiDoc documents
against a variety of XSL stylesheets. Both documents and stylesheets may be
either on site or remote. The framework will also perform validation and error
reporting on EpiDoc texts. The transcoder classes will be accessible via a
Cocoon Transformer, so that Greek text can automatically be shifted from any
supported source format to any supported result format. It will employ the Java
version of CHET-C in a custom Cocoon Generator or Transformer to allow users to
generate EpiDocs from already-digitized inscriptions. The framework will be
packaged as a Web Archive (WAR), so that interested developers can easily deploy
their own digital epigraphy frameworks.
Finally, the team is also working to integrate the Java tools with jEdit, a
Java-based text editor with a very robust plugin architecture (). The goal is essentially to create an
Integrated Development Environment (IDE) for epigraphers which editors will be
able to use in the electronic publication of EpiDoc texts. All of the tools and
stylesheets developed for use with EpiDoc will be available as jEdit plugins and
editors will also have available jEdit's XML and XSLT plugins, which provide
useful functions like code-completion, tag creation help (based on a DTD),
validation, error reporting, and transformation. The combination of the web
framework and the text editor will provide epigraphers with a cross-platform,
Open Source solution for creating and publishing inscriptions.
All of the tools described above will be demonstrated briefly, using texts marked
up by various projects that have adopted the new standard. Then I will conclude
by discussing how the EpiDoc Collaborative plans to proceed in developing the
standard and the supporting tools. The project's use of open standards and Open
Source development tools provides a useful model for similar types of scholarly
publication, one which could be picked up and extended for other purposes
without a great deal of effort.
Digital texts in Greek pose problems of their own because of the various methods
which evolved to encode them prior to the advent of Unicode. One of the earliest
of these was Beta Code, which utilizes 7-bit ASCII to represent Greek, Latin,
and Aramaic texts, as well as various epigraphical and papyrological sigla. In
addition, a number of Greek fonts, each employing its own specialized encoding,
were developed over the years to handle the problem of typing and viewing Greek
with computers. There exist already various programs and filters which can
handle shifting the encoding of texts from one form to another (the Greek texts
on employ such filters, for
example). There was still a need for a more flexible system to handle encoding
shifts, however, and so the Chapel Hill team (in collaboration with the Stoa
Consortium, ) has developed software in Java
that performs this function and which can easily be plugged into the larger
toolkit framework. The transcoder classes can perform encoding shifts on text
within specific elements of an XML file, and are also designed to be modular and
easily extensible to handle languages other than Ancient Greek.
The team has also developed a web-based framework for the presentation of EpiDoc
texts. The framework has been built with Apache Cocoon () and is hosted by the Stoa. The
"Epidocinator" as it is called, can dynamically transform EpiDoc documents
against a variety of XSL stylesheets. Both documents and stylesheets may be
either on site or remote. The framework will also perform validation and error
reporting on EpiDoc texts. The transcoder classes will be accessible via a
Cocoon Transformer, so that Greek text can automatically be shifted from any
supported source format to any supported result format. It will employ the Java
version of CHET-C in a custom Cocoon Generator or Transformer to allow users to
generate EpiDocs from already-digitized inscriptions. The framework will be
packaged as a Web Archive (WAR), so that interested developers can easily deploy
their own digital epigraphy frameworks.
Finally, the team is also working to integrate the Java tools with jEdit, a
Java-based text editor with a very robust plugin architecture (). The goal is essentially to create an
Integrated Development Environment (IDE) for epigraphers which editors will be
able to use in the electronic publication of EpiDoc texts. All of the tools and
stylesheets developed for use with EpiDoc will be available as jEdit plugins and
editors will also have available jEdit's XML and XSLT plugins, which provide
useful functions like code-completion, tag creation help (based on a DTD),
validation, error reporting, and transformation. The combination of the web
framework and the text editor will provide epigraphers with a cross-platform,
Open Source solution for creating and publishing inscriptions.
All of the tools described above will be demonstrated briefly, using texts marked
up by various projects that have adopted the new standard. Then I will conclude
by discussing how the EpiDoc Collaborative plans to proceed in developing the
standard and the supporting tools. The project's use of open standards and Open
Source development tools provides a useful model for similar types of scholarly
publication, one which could be picked up and extended for other purposes
without a great deal of effort.