As the Çatalhöyük Archive Report 2007 is online as a large PDF, I’ve copied my report below. In 2007 I’d worked with the team during the off-season (i.e. when people aren’t on site digging) and on-site. Being on site meant providing general IT and network support (and dealing with occasional oddities like a spontaneously combusting monitor) while working on requirements analysis and database development in the lab. I also contributed to the Çatalhöyük blog during the 2007 season.
Clay Databases Development Archive Report 2007 Part 2 – Mia Ridge,
Çatalhöyük Research Project
The clay databases at Çatalhöyük are designed to implement a shared recording
system allowing a unified view of related finds. It provides common value lists for
use when it is supported by the attributes of the fabrics and the artefacts whilst
allowing flexible recording where the fabrics and attributes differ. The extensible
design will enable comparison of artefacts across specialisms on representational as
well as material aspects. One of the goals was to record to a level of detail and on
characteristics supported by the artefact rather than forcing it into a universal system,
while providing enough consistency to ensure accurate counts of each type of find.
An underlying goal is to manage the data entry interfaces and processes and the
integration of previously recorded data to ensure that negative evidence is as reliable
as positive evidence.
The requirements for the shared clay database application were reassessed this season
in light of factors including: changing requirements as the infrastructure is embedded
in practice; the time constraints teams face when recording on site; discussion of the
portability of the ware code and fabric descriptions, and the ability to record technical
material data consistently across teams, labs, languages and specialisms. The work by
Chris Doherty and other teams on understanding the raw materials in the surrounding
area, which has implications for the intentionality represented by the presence of
various inclusions and for the basic understanding of fabrics for different types of
artefacts, is also an on-going influence on the application design.
I reconfigured the recording model accordingly so that it follows a concept of
‘observation, not interpretation’. This conceptual model provides a practical method
for dealing with different levels of technical expertise and for different research
interests across teams. The database schema and forms can be extended as we find out
more about the raw materials around the site and to support different types of
This change particularly affects the recording of colour variations and inclusions or
temper. For example, recording can range from basic observations of the dominant
and variant colours, the probable type of heat exposure and the recorder’s certainty
about their interpretation, to a detailed analysis and interpretation of the reasons for
colour variation. Conjecture about the cause of colour variation, whether a factor of
fabric, manufacture, use, or post-depositional events, can be saved until technical can
be performed. This analysis might also include research about the natural variations
in the fabrics found in the region or experiments with firing and exposure to other
heat to examine how the materials change over time or with exposure to different
heating events. This model also allows the presence of inclusions to be recorded
without requiring speculation about the natural or cultural origin of those inclusions.
This season the mapping the cleaning and normalising of ceramics data, and issues
around recording bulk, grouped or individual sherds, were finalised. I worked closely
with Nurcan Yalman and Duygu Tarkan to devise the best solutions for their
recording requirements and constraints, with the added challenge that other teams will
be using versions of the ceramics databases. Our aim is to have fabrics recorded
consistently across the database, regardless of the recording method (which is often
the result of post-depositional or use events that affect the size and condition of the
sherd) to the extent that the material and physical aspects of the artefact permit.
Work included analysis of the portability requirements for off-line analysis, the use of
images and review of existing structures to remove redundancies and data cleaning.
Some data cleaning included the removal of non-figural objects such as clay balls and
geoshapes so that the recording structures can be tailored for the confirmed figural
objects that remain. The non-figural objects can be exported and integrated into the
appropriate specialist database.
The Figurines database uses HTML forms to provide a cross-platform application for
centralised data entry. It also uses SQL views to provide equivalents to the ‘repeating
fields’ used in FileMaker Pro. These tables are also available for use in user-created
Integration with Geographical Information Systems (GIS)
While on site I explored the requirements for, and created some test data structures to
link excavation and find data into the Geographical Information Systems (GIS)
software that could serve as a ‘proof of concept’ for user-friendly methods of
importing central database records into GIS software.
I experimented with linking X Finds recorded in the central database with views of
the spaces, features and units in Building 51. A comparative experiment attempting to
display the location of skeletal remains according to the age at the time of death,
burial position and alignment, using records from the Human Remains database and
human remains information from the excavation database, showed the benefit of
highly structured data for pre-defined searches based on particular criteria. However,
free form and subjective data is useful in contextualising the finds once a set of
records has been selected according to criteria and displayed in the GIS application.
This experiment showed why it is important to plan and digitise properly; and
demonstrated that team members cannot analyse data points that have not been
recorded individually and consistently.
However, it is impossible to anticipate all research questions so extensibility is also
very important. Trials this season might point to a need for summaries of specialist
data created for use by people from other teams or members of the public who do not
have the same ability to review, interpret and update specialist data into a highly
structured schema suitable for querying in GIS. These summaries could be recorded
in supplementary data structures when initial specialist data is recorded or after a
survey of the collection of objects for interesting possible points of analysis.
The use of geographical information system (GIS) could have exciting implications
for the on-going work on clay artefacts and geological materials around the site. By
using GIS in conjunction with the central database, we might be able to relate
artefacts to the location of their source material, effectively re-populating the
landscape with objects made from local materials.
Documentation and training
This season I ran an informal workshop on ‘database querying’. This was very useful
as those who attended were able to construct their own queries and reports, resolving
some of their on-going questions; it also helped us understand how people query the
database and where we can make improvements.
I also created tables (‘Database_Documentation_Table’, for example) to store
documentation so that it is accessible at the point of use in the AllTables database.
The AllTables database was created in 2005 to allow team members to use the
Microsoft Access visual interface and ‘wizards’ to create their own queries.
It is worth recording the decision to create a pseudo unit to allow with finds with
Mellaart numbers to use the ‘GID’ or common object identifier in use for all finds
recorded in the current system. This will require some additional search functionality,
including an equivalency table to link the new GID with the previous Mellaart
identifier. This table would serve as a central register, handing out new Find Numbers