The Wellcome Arabic Manuscripts Project

The Wellcome Arabic Manuscripts Project
Brey, Gerhard, Centre for Computing in the Humanities, King's College London, UK, gerhard.brey@kcl.ac.uk
This poster and software demonstration will present the Wellcome Arabic Manuscripts project. The aim of this project was to create a freely available online catalogue of ca. 500 mainly medical manuscripts written in Arabic that are preserved in the Wellcome Library (London, UK). Apart from the actual manuscript catalogue the outcomes of this project that are of particular relevance to the Digital Humanities are:

a TEI/ENRICH based XML schema adapted to meet the requirements of cataloguing Arabic manuscripts and in particular to accommodate detailed codicological and textual descriptions
an open source, web-based, customizable cataloguing tool
an online research tool that gives users free access to a wealth of metadata and digitized page images of the manuscripts in this collection
high-quality digital images of each manuscript page linked to rich descriptive metadata.
The project was partly funded by the Wellcome Trust and partly by a grant from the UK's JISC Islamic Studies Programme [Henshaw, 2009]. It is a collaboration between three institutions:

Wellcome Library, London, UKhttp://library.wellcome.ac.uk/
Bibliotheca Alexandrina, Alexandria, Egypthttp://bibalex.org/
King's College London, London, UKhttp://www.kcl.ac.uk/cch/
The Wellcome Library provided the cataloguing expertise and methodology, the digital images, and overall project management, while the web based tools were developed and are hosted by the Bibliotheca Alexandria. King's College London (Centre for Computing in the Humanities) managed the technical requirements specifications for the cataloguing tool, and adapted the TEI/ENRICH XML modelhttp://library.wellcome.ac.uk/arabicproject.html.

The cataloguing methodology and the approach to this project was guided by Wellcome's tripartite approach to cataloguing Oriental manuscripts. This approach suggests that a manuscript should be considered as a product of craftsmen, authors and readers and therefore its production, intellectual content and subsequent use. The description of a manuscript is carried out having these three aspects in mind, viewing the manuscript:

as a museum object (palaeography, codicology)
as an intellectual creation (texts)
under historical user aspect (provenance, owners, editors, etc.)
As a reflection of this approach the manuscripts were catalogued in much more detail than a typical library manuscript catalogue, particularly in the areas of the materiality of the manuscript and the description of the textual content. In practical terms this approach -- together with the newly created tool described below -- allowed the cataloguing tasks to be allocated according to competence and practicality. The detailed codicological description, best done on the actual physical object, was carried out by conservators at the Wellcome Library, whereas the detailed description of the textual content (down to the transcription of chapter headings) was undertaken by cataloguers at the Bibliotheca Alexandrina.

A key objective of the project was that the encoding model should follow established standards to ensure interoperability of the manuscript catalogue and to make it extendable and as flexible as possible. A model based on TEIhttp://www.tei-c.org/ and the format used by ENRICHhttp://enrich.manuscriptorium.com/ (a European manuscript cataloguing project) was chosen [Pierazzo, 2010]. The basic TEI/ENRICH model had to be adapted and customized to accommodate the needs of the tripartite approach. It had to be ensured, for example, that the model is compatible with standards such as Ligatushttp://www.ligatus.org.uk/, an emerging standard for the detailed description of book binding features. Other extensions to the model were needed to enable the description of the very detailed physical features from a conservator's point of view, such as flaps, endbands, or covers. From a more palaeographical perspective various features relating to the scribe had to be added, for example an element to describe the Mistara, an impression on the paper achieved by applying a kind of stamp that indicated the lines a scribe would then write on. Other important palaeographical features that had to be represented were the coefficients calculated by using the "Pace" method, a system devised by Nikolaj Serikoff [Serikoff, 2001] to measure certain features of an Arabic script, such as angles of letters, or the ratio of connected and unconnected letters on a line. These features taken together are quite unique for a scribe or scribal school and help to situate a manuscript chronologically and geographically. In order to adequately represent the intellectual content of the manuscript, further adaptations were made. In order to fully represent compound Muslim names, for example, fields for their constituent parts (e. g. patronymic, honorific) were introduced. Additional incipit-like elements were also needed to hold those formulaic passages at the beginning of Arabic manuscript texts (the Basmallah, or Tahmid that superficially look like invocations, but that also tell about the subject area of the text to follow, or the origin of the author.

In addition to the cataloguing effort by the cataloguers at the Bibliotheca Alexandrina, all of the technical development was undertaken by the Egyptian partners. The two outcomes of this substantial development effort are the web based cataloguing tool for data input and administration by Wellcome and Bibliotheca Alexandrina cataloguers and the online research tool that enables scholars and the wider public access to the rich material via browsing and searching.

The web-based tool was designed to allow the cataloguing of manuscripts into valid TEI XML files without prior specialist XML knowledge. This was achieved via the development of a Schema driven editor (SDXE) together with a "configuration grammar", the XML Skeleton Annotations (XSA) [Aboulnaga, 2010].

The XSA system automatically builds JSF (JavaServer Faces) based XML editors. These editors in turn produce schema compliant XML files that follow a certain XML skeleton. To enable users to author such XML files, the system generates web forms with fields for each data holding XML element in the skeleton. It does this by reading an XML Skeleton Annotations file (XSA.xml) that contains definitions for each location in the XML skeleton including a label, a help text, authority lists, user access rights, and various other information used by the system to generate the web forms. The XSA system uses the Schema Driven XML Editor to generate schema compliant XML files. The steps necessary to generate a website using XSA are as follows:

authoring of an XSA.xml configuration file
optional creation of facelet templates for look and feel
creation of a blank XML record template
The central component of the cataloguing tool is the XSA configuration file. By changing this file it is possible to adapt the cataloguing tool for any other similar manuscript cataloguing tasks.

The web based research tool, i.e. the online manuscript catalogue is directed towards both specialist and non-specialist use. This means that it has to provide functionalities that address scholarly users, palaeographers, conservators, but also a wider audience whose specialist fields lie in other areas. The research tool therefore offers entry points into the repository via multiple levels and access routes, such as browsing (alphabetic or faceted) or searching (simple or advanced) for a wide ranging set of criteria. The advanced browsing and searching mechanisms take into account features peculiar to manuscripts in Arabic, for example search by stem or search by root, differentiated again by normalized character forms or defective (i. e. dotless) character forms.

During the poster session we will give a live demonstration of the web based cataloguing tool and the online research tool. This will showcase the key features of both applications and highlight the problems that were encountered and how they were overcome

Contributors to the Project
Wellcome Library
Dr. Richard Aspin
Dr. Christy Henshaw
Dr. Nikolaj Serikoff
Bibliotheca Alexandrina
Prof. Magdy Nagi
Dr. Noha Adly
Younos Aboulnaga
King's College London
Simon Tanner
Dr. Elena Pierazzo
Gerhard Brey
References:
Aboulnaga, Y. 2010 “XSA - XSD to configurable JSF UI, ” software package, (link) documentation wiki, (link) October 2010

Henshaw, C. 2009 “JISC funding for the Wellcome Arabic Manuscript Cataloguing Project, ” Wellcome Library Blog, 14 August 2009 (link) October 2010

Pierazzo, E. 2010 “On the Arabic ENRICH schema, ” Wellcome Library Blog, 27 August 2010 (link) October 2010

Serikoff, N. 2001 “Image and Letter: "Pace" in Arabic Script (a Thumbnail Index as a Tool for a Catalogue of Arabic Manuscripts. Principles and Criteria for its Construction), ” Manuscripta Orientalia, 7 56-66