Digital Data — Ur of the Chaldees: Making a Virtual Vision Possible

William B. Hafford

(See email contacts page for the author's email address.)

---

A two-year extension of the Ur Project officially began in July of 2013 with lead funding continuing from the Leon Levy Foundation. We have now embarked on the program designed in the first year, hiring extra hands and taking on additional assistance from volunteers and students. Our ultimate goal is to make a complete online research tool for scholars and the public alike containing all original documentation and modern observations on material from the ancient city of Ur. Our focus for the time being is on material stored in Philadelphia and in London at the two sponsoring institutions of the original 1922-1934 excavations, the Penn Museum and the British Museum.

This task involves a great deal of data and has its inherent difficulties, but as always we are approaching it with innovation and verve (for more on the first year of the project, see Hafford, W.B., "Digital Data — Ur of the Chaldees: A Virtual Vision of Woolley's Excavations," CSA Newsletter, XXVI, 1; April, 2013: at csanet.org/newsletter/spring13/nls1301.html; and the ASOR blog entry for June 2013 at asorblog.org/?p=4763). Now that we have digitized the original catalogue cards concerning more than 20,000 objects and organized them in a searchable open-source database, we are proceeding to the creation of the front end presentation of that data. There is still more to be done, however, such as connecting field photos, modern photos, archival documents, and much more to the catalogue data. The digitization of all of this information is underway; plus, some of it has already been gathered and resides in other databases in the museums. We do not want to become a separate data silo; instead of copying and storing modern images and descriptions of the same objects Woolley described, we are finding ways of connecting directly to collections management systems already existing (such as the Penn Museum’s KE-EMu: penn.museum/collections/). In this way, when new information is added to the collections data, the most up-to-date will always be made available on our site as well.

When the image access system and a few other interconnection search methods are functioning, the landing page will be made public. We hope this will happen within the next few months. Even then the website will remain a work in progress as we add to the data and connections available, and as we obtain feedback from users to improve the site overall. Our working assumption has always been that our progress should be as transparent as possible, with the understanding that our audience will be able to assist us in making the site better even in the early stages.

An important step in this process has been in hiring a programmer and database manager, Sasha Renninger. Hers is a critical position since the data being gathered at the Penn and British Museums will all be connected through the systems she develops. The institutions jointly agreed that the programmer would be based in Philadelphia, taking trips to visit London as needed and holding frequent online meetings. Meanwhile, we have constructed separate teams to digitize artifacts and original records from Ur housed in their respective institutions: conservators examine artifacts for current condition, beginning treatment on those pieces in most need; photographers record every angle of artifacts to allow virtual examination; archivists scan original photos and documents for historical and archaeological research. The database formed of Woolley’s original field catalogue data is the core of the project, but it must call on imagery and analysis, bibliographies and publications, research and original notes where applicable. It must be able to link to a variety of sources within our site and potentially beyond it. At the heart of larger-scale interconnections lies our use of Open Context (developed by Eric Kansa originally at the Alexandria Archive Institute http://opencontext.org). We are adapting the source code of this data presentation and publication site to our specific needs so that it will work with the Ur schema. The overall process simplifies data storage and delivery, moving from the concept of a relational database to subject and variable connections that can be displayed in both human-readable and machine-readable formats. This latter allows for linking across the web with other open and freely accessible information expressed in similar machine-readable fashion.

The relationships used by Open Context are partially informed by ArchaeoML (developed by David Schloen of the University of Chicago Online Cultural Historical Research Environment). Open Context uses its structural relationships for expressing objects and locations in a broad and flexible manner. For instance, one of the most important needs for archaeological data is to express one volumetric location contained within, neighboring, or partially contained within, another. What the actual unit is called makes no difference (it could be a pit, a trench, a grave, a room, a building, or any other term). Similarly with objects; they may be classified in larger or smaller groups and the relationships of groups may be defined. Specific terminologies and relationships for Ur derive from the site and its excavation as informed by an understanding of Mesopotamian culture and history; for other regions, the names of categories would be different, but relationships could be similar. For our data, Woolley’s original terminology is completely preserved, but it is also structured in a way that should make grouping for modern research more efficient, informative and comparable in the bigger picture. This will make our site part of the growing Linked Open Data on the web (for an excellent discussion of the power of Linked Data, see the TED talk by the creator of the World Wide Web, Tim Berners Lee: www.ted.com/talks/tim_berners_lee_on_the_next_web.

Archaeological data is particularly complex, with groups of artifacts being connected spatially and temporally as well as functionally, morphologically and/or aesthetically. Archaeologists do not utilize a single set of standardized descriptors for their sites, their artifacts or their archives; thus, flexible interconnections are vital. Even within Ur, designators for areas, types for artifacts, and house numbers changed over the twelve years of excavation. As we make field notes and artifact records digital, all of the iterations must be brought back together so that complete and accurate research can be conducted.

 

The variety of information available from Ur is staggering, but approaching it in defined groupings allows us to progress effectively. We have separate people working on artifacts and archives, but each of these definable groups of data is large and needs dividing into planned workflows. For artifacts, we decided on ceramic objects as an initial group at both museums for several reasons. First of all, pottery is probably the most ubiquitous artifact category found in archaeological excavations and is often at the root of relative dating techniques. This means the potential information value is high. Second, and rather surprisingly, pottery was perhaps the least-well-documented group at Ur. Woolley certainly spent time in creating his typologies, and he produced drawings of many pieces. But considering the vast numbers of pots and potsherds found at the site, there is comparatively little in the paper trail. It is likely that we can reconstruct Woolley’s process from the scattered types written on field note cards, but no direct notes on the typology work overall are available. Moreover, the possibility of more easily comparing his pottery types to others across the ancient Near East will help researchers in understanding the sequence overall.

In a related third point, pottery was so numerous that Woolley collected only a small percentage, and most of these were whole examples. He then gave field numbers only to the best examples of these, meaning that as many as 75% of the pots in our museums now have no field numbers. Many have location designators written on them, however, and this combined with field notes will help us better understand distribution. Woolley and crew, then, wrote down some information onto the artifacts themselves, and they did so in a relatively ephemeral manner. Many of the field numbers and area designators are fading or nearly illegible at this point; documenting peripheral markings on pottery is thus a pressing need.

Finally, the Ur pottery has not been much investigated over the past 90 years. This is possibly because it was not as well documented as other pieces, because publication includes general pottery drawings and these were assumed to be enough, or because groups of largely undecorated pottery on shelves just were not as appealing as more unusual artifacts. The collection of pottery and other artifacts at such an early excavation was also less complete than what would be gathered for study today. Nonetheless, the study of these objects is important and will be much easier to conduct with digital data about them readily available.

 

Along similar lines we are including clay figurines, models, and relief plaques in the early digitization stages. These objects are quite common, and they have been more often investigated than clay vessels due to the more involved symbology they display. Father León Legrain, curator at Penn and epigraphist at Ur for two seasons, created a catalogue of the majority of these pieces. The catalogue was originally intended to be part of the Ur Excavations series but was never published. Instead, some parts of it appeared in other UE volumes. It is now possible to put this unpublished volume online as a scanned document for researchers to utilize along with the digital information on every figurine currently available from Philadelphia and London.

Important information about life at Ur is to be found in the contemporary writing of the people themselves. Thousands of cuneiform tablets were uncovered at Ur, and a good deal of work on them (and many other ancient sites) has already been done. Some of this is available digitally (see, for example, the Digital Library Initiative at cdli.ucla.edu/, and the Open Richly Annotated Cuneiform Corpus at oracc.museum.upenn.edu/). We are continuing this work in an effort to ensure the digital documentation and translation of every tablet excavated at Ur. The British Museum, assisted by many external institutions, has been working hard on this task, to include taking high-quality photographs; work on Penn tablets will begin soon.

The archival team at the British Museum has completed scans of the glass plate negatives of field photos that Woolley produced during the excavation. Many of these photos were published, but by no means all of them, and by scanning the negatives we are acquiring the highest quality images possible. The plates themselves, and all archival materials, will continue to be preserved in the museums, but the digital images will be available to all. At Penn, archival work has centered on Woolley's letters and reports from the field as well as those of Father Legrain. These letters, along with scans of field notes stored in the British Museum, are being placed on a website for volunteers to transcribe into digital text (see UrCrowdsource.org). We began this experiment more than a year ago, and at this time last year we had acquired just over 500 transcriptions. The number has now grown to around 2000. We have also hired a work-study student to assist with editing volunteer input and to manage parts of the site. Our programmer is helping to make the site more efficient, though her primary attention must be directed to the database and web interface for the overall research tool we are creating.

In addition to involving volunteers and work-study students in archival work, we are bringing in a variety of student researchers in an effort not only to accelerate the rate at which we gather data, but also to give these students the opportunity to learn about the site and contribute to the field through their own research. Students are encouraged to develop their interests in conjunction with the needs of the project. In this way they gather important information for our database while learning about research, digital data storage and retrieval techniques, as well as particular aspects of the ancient city of Ur.

The data gathered in student research, archival scans, volunteer transcriptions, artifact examinations, and artifact photography will all be united through the database. It will be connected through metadata and mark-up language, displayed to researchers and the public through Open Context, and to more complex searches through machine-readable code. The project is ambitious, but achievable and being achieved. As we add to our data and hear more from researchers our website will grow, and as data from other archaeological sites are made available in similar open ways, the tool will only increase in applicability and usefulness.

----

Thanks and acknowledgement are due to the entire team at both museums, and to all the volunteers and students who are helping with the project:

The co-directors at the British Museum are Jonathan Tubb, and Irving Finkel leading a team comprising Birger Ekornåsvåg Helgestad, Jon Taylor, Gareth Brereton, Nadia Linder, Alexandra Porter, and Duygu Camurcuoglu. The co-directors at Penn Museum are Richard L. Zettler and Stephen J. Tinney, leading a team comprising William B. Hafford, Sasha Renninger, Tessa de Alarcon, Ryan Placchetti, and Shannon Advincula. Students and volunteers at the Penn Museum are Elena Yandola, Kevin Ennis, Cindy Srnka, and Tom Pedrick, and we must not forget the many tremendous volunteers helping every day at UrCrowdsource.org

-- William B. Hafford

 

 

About this document:

All articles in the CSA Newsletter are reviewed by the staff. All are published with no intention of future change(s) and are maintained at the CSA website. Changes (other than corrections of typos or similar errors) will rarely be made after publication. If any such change is made, it will be made so as to permit both the original text and the change to be determined.

Comments concerning articles are welcome, and comments, questions, concerns, and author responses will be published in separate commentary pages, as noted on the Newsletter home page.