Dr Nick Ryan
University of Kent at Canterbury,
Databases and information systems have long enjoyed, if that is the right word, a reputation as essential, if rather dull, administrative tools. University students flock to information systems courses more often because they see them as relevant to their future employment prospects than because they expect excitement and intellectual stimulation. Today, businesses and organisations large and small rely on computer based information systems to maintain their lists of personnel and customers, for inventory or stock control and for a variety of financial purposes. These applications are now so well established; they are often regarded as mission critical and so only attract attention when things go wrong or when impending disaster provides journalists with a good story. We are all familiar with the old tales of multi-million pound gas bills, and many will have noticed the recent outbreak of articles predicting chaos as system clocks fail to cope with the change to the new century.
Why then, if this characterisation is valid, have I chosen to talk about information systems at a forum dedicated to arousing and widening interest and, indeed, why talk about them in the context of archaeology? There are, of course, several reasons. Firstly, this image of information systems as essential but uninteresting really applies only to the long-established applications. A number of innovations including Geographical Information Systems (GIS) and the World Wide Web (WWW) have led us away from the traditional commercial and administrative view of information systems towards one which is altogether more wide ranging.
Certainly, archaeologists, whether in academic institutions, excavation units, or local and national government bodies, all need the support of conventional applications for basic record keeping and administration, but their needs go much wider. They have to deal with a remarkable variety of different types of information. Archaeology is above all a multi-disciplinary subject drawing on a wide range of skills and specialisms, from the arts and humanities through to the biological and physical sciences.
From a computer scientists perspective, archaeological applications provide some significant challenges, one of which is to develop information systems that can cope with this variety. In the processes of research, excavation, analysis and publication, each of the many specialisms generates vast quantities of data, much of it of widely differing types, and the challenge is to provide ways in which this can be presented to and used by all who need it.
It is this challenge that I shall be addressing today. I shall briefly review the ways in which information systems have been and are being used in archaeology, before describing some recent innovations and their likely impact on archaeological computing. Many of the results of the past decade and more of database research are only just beginning to have an impact on the commercially available systems. However, I hope to show that they offer an enormous potential in helping us all to use archaeological data. I shall concentrate on the role of computer based information systems in excavation, but much will be equally applicable to other archaeological activities from field survey through to individual specialised research.
Excavation has become a familiar sight to many people living in or visiting our historic towns and cities and to others through television and other media. Together with the museum displays of recovered artefacts, this is really only the tip of the archaeological iceberg. As we shall see, excavation generates large quantities of data but, even before it starts, the modern archaeologist may have sifted through and collated a wealth of material from many different sources.
Information gathering during the planning stage may involve field walking and other surveys, building surveys, consulting historical and other documentary sources, and the sites and monuments records maintained by local and national government bodies.
During excavation detailed information is recorded about every excavation context. These include each layer of soil and features such as post holes, pits and ditches. Each artefact is recorded together with information about its exact find spot. Numerous environmental and other samples are taken for laboratory analysis and the location and purpose of each is carefully recorded. Large numbers of photographs are taken, both general views of the progress of excavation and detailed shots of contexts or finds, and many plan and section diagrams are prepared showing the shape and location of contexts and larger artefacts.
Unlike most other techniques used to investigate our past, excavation is a destructive process. The visitor to a preserved monument is not always aware that some of what they see may have been removed during excavation and later replaced. Even where the remains are still in situ, virtually all of the information about the contemporary environment and human activities on the site has been destroyed by excavation. This fact alone makes it imperative that archaeologists employ the best techniques available to them, and ensure that their recording methods are both accurate and reliable. Unlike many other applications of information systems, it is simply not possible to go back and check at a later date.
The post-excavation stage is where all of this recorded detail must be collated and analysed. The main task is the development of a structural history of the site from a study of the physical relationships between excavation contexts. Working in parallel, finds and environmental specialists contribute both to the central task of interpreting the site, and to the more general aims of extending understanding of past material culture and environment in the area. The eventual outcome of this work is a site archive containing both the original excavation records and their interpretation. This archive is a critical resource: it provides the basis for all future work including publications, museum displays and, in due course, input into future project planning.
My first experience of archaeological excavation was in the early 1970s. At that time most recording was hand-written and in notebooks. Some early experiments were made with direct recording by computer in the field but the techniques, involving mechanical teletypes connected to mainframe computers via telephone lines, were neither affordable or very beneficial. It was in the early 1980s with the increasing availability of microcomputers and database software that computer based recording started to become more popular. Today, it is almost universal.
A significant change which preceded widespread computerisation was a move away from free-form notebooks to pre-printed recording forms. Usually, one form is completed for every excavated layer, or context. Further forms may be used to record finds and samples. This development accompanied a number of other improvements in excavation techniques which were eventually to lead to highly structured and more consistent recording. In due course this eased the move to computer based records. Indeed, computers soon became necessary to manage the volume of data produced in this way; the larger urban excavations might produce ten or twenty thousand of these forms.
Early database systems brought many of the basic benefits of computerisation in that they helped to maintain the consistency and reliability of the data in ways that were all but impossible with paper records. For example, in this form the Category, Interpretation and Period fields might be limited to a range of pre-defined values stored in the system. Similarly, the fields showing physical dimensions might be configured to accept only numeric values. Further constraints might help to ensure that related information about the artefacts found in each layer or environmental samples were always associated with the correct context.
The next developments came about through improvements in the user interfaces of micro computer software. This slide shows a typical excavation context record produced using a popular database package. We cannot ignore the way in which these developments helped to make systems easier to use through their use of colour and direct interaction with controls such as buttons and sliders.
However, it is questionable whether this information system is more archaeologically useful. The changes are often little more than cosmetic, at best ergonomic. Notice that although the appearance is more attractive and it is certainly more pleasant to use, the information provided is exactly the same as with the earlier systems. It is restricted to handling text, numbers and exact dates. In fact, the particular system used here can display images, but it cannot do anything else with them. This is an issue that I shall return to later. In this example, the fields for plans, sections and photographs show only the reference numbers of these objects. They act simply as a index to conventional physical media which are usually stored in map or slide cabinets.
There are, of course, many Computer Aided Design (CAD) and other drawing programs that allow us to create and store important resources such as plans and sections on a computer. Equally, image editing or painting software enables us to store and manipulate many types of images including photographs. Indeed, many archaeologists use such programs for these purposes but, until quite recently, it has not been possible to integrate these with the references to diagrams and photographs stored in the excavation database.
This, then, is the typical form of a information system for excavation records. It is based on a variety of software, often one that has been assembled over many years. Each different tool performs a separate task, largely unrelated to that of the others. A database management system with purposewritten application programs provides the facilities for data entry and the storage of context, finds, environmental and other records. Written reports, both in final form and as interim working documents, are produced with word processing software. Increasingly, the more technologically advanced excavators also store photographs on PhotoCD and use a DTP package for producing their publications.
The main benefit of this approach is that if the software has been chosen wisely each tool is used for the task to which it is best suited. Unfortunately, the lack of any integration between programs prevents automation of searching, or the recording of intelligent links between items. The separation of the different types of data make it difficult to move smoothly between different representations of important information. For example, when viewing a particular context record it will usually be possible to see related finds records if they are also managed by the same database system. Viewing plans showing physical location, or photographs, or a textual description of a group of related contexts in a partiallywritten report is not always so straightforward. Often this can only be achieved by starting a different program and then searching manually for the required information.
One of the major developments in desktop computer software of recent years has been the possibility of communication between programs; programs that talk to each other. This is not a particularly new idea. Indeed it has been commonplace in larger systems for many years.
Many archaeological computing tasks can be performed with off-the-shelf software, although it may need to be adapted to specific requirements. In practice, archaeologists require few tools that cannot be produced in this way. One possible exception is in creating and manipulating stratigraphic diagrams.
The stratigraphic diagram or Harris Matrix is another of the developments of the 1970s referred to earlier. It is in fact a form of directed graph showing the chronological relationships between excavated layers or contexts. In general, if two layers are in contact with each other and one overlies the other, then the upper layer is chronologically later. This, although somewhat simplified, is the basis on which the structural history of a site is founded. The construction of this diagram and its subsequent use in the interpretation of structural phases is central to both the understanding of the site during excavation and to the post-excavation process.
The program illustrated here shows a tool for manipulating stratigraphic diagrams. The user can create and alter the diagram by adding, moving or removing the boxes representing excavated layers. As interpretation progresses, layers can be grouped together to illustrate the components of a building or other contemporary groups.
This particular program stores all of its working data in the main excavation database and can communicate with other programs. Thus it is possible to move between different representations of the same object such as its place on the diagram, an associated record in the database, or a photograph of an appropriate part of the site. In this way the stratigraphic diagram can be used as a road map, a means of navigating around the database.
This approach offers some benefits over earlier systems, but it is not without problems. The programs are still separated and the communication structure is fragile because no one piece of software has overall control. A user might inadvertently change something using one program that might disrupt the network of links - e.g. an image is accidentally deleted or moved to another location. These are amongst the problems that database systems are designed to overcome. Unfortunately, most current database systems simply cannot handle this variety of different types of data.
Much can be achieved with current tools, but most of the software we use today is very complex, often too complex for the needs of any one group of users, and we may often use several programs that largely replicate each others capabilities. One possible scenario envisages a return to a collection of simpler limited-purpose programs. A simple example would be the replacement of the current generation of large and complex word-processors by something very much simpler, but with optional add-on tools to perform tasks like manipulating tables or editing diagrams. A user might then choose to purchase only those tools that were necessary for their work.
This theme of simplification of the programs presented to end users is also reflected in many of the recent developments in database systems. These developments are only slowly finding their way into commercial products, but they promise to make it much easier to develop applications to handle complex and varied data. The two most significant areas of innovation are methods for extending the range of data types that these systems can handle, and the ability to hold both the data and behaviour for these new data types in the database. A simple example will help to explain the possibilities.
Most current database systems are limited to storing text, numbers, dates and very little else. For these types of data they provide a number of basic operations needed to manipulate them, such as converting between upper and lower case for text, arithmetic operations for numbers, and simple date arithmetic for dates. These operations define the behaviour of the data type and ensure that only legitimate operations take place. They ensure that whilst numbers can be added or subtracted, addition and subtraction have no meaning for text. Similarly, text can be converted to upper case, but this operation has no meaning for numbers.
But what of the more complex types of data that are so widespread in archaeology? Some current systems are capable of storing additional data types, such as images, but most do not define any behavioural operations beyond simply displaying the image on the screen. However just as with numbers and arithmetic, there is a set of basic operations that are commonly used in most image manipulation programs. These include windowing - the extraction of a part of an image, adjustment to colours, brightness and contrast, various image enhancing operations such as sharpening of edges, and methods for combining two or more images to produce a composite result. To perform any of these operations on an image stored in such a database, the image must be first be copied to a file, then manipulated by a suitable program, then copied back in to the database. With large volumes of data this would be a tedious and time consuming task.
The more advanced systems are now making it possible for a programmer to define new data types to suit specialised requirements and to provide their appropriate behavioural operations. For example, conventional database systems can only handle exact dates specified as day, month and year. These are of little use in archaeology where dating is usually much less precise and a date may be indicated only by a century or a named period such as late-roman or early-medieval. I have recently developed a collection of data types that can handle these imprecise, or fuzzy, methods of expressing dates, periods or intervals of time. These enable the system to answer queries about the ordering of imprecise events, the overlap of periods, or the lengths of intervals between events.
In some cases, special add-on libraries are available that provide the extensions needed for common data types such as images and spatial data. Once these new types have been defined it is much easier to write programs that access and manipulate the data. Indeed, it becomes possible to develop systems in which most of the operational intelligence is located in the database system rather than in the programs that use it.
For archaeological purposes, the addition of data types suitable for dealing with photographic images, line drawings such as maps, plan and section diagrams, imprecise dates, and so on, opens the possibility of constructing information systems that satisfy most, if not all, of the requirements of modern archaeological research.
This example shows an experimental system that is being used in a field survey project in southern Corsica. Unlike excavation, such fieldwork is non-destructive, but it nevertheless may generate large amounts of data, all of which need to be carefully interrelated and brought together for analysis. The system shown here combines maps, photographs and textual information in a single database and is now being extended to deal with aerial and satellite imagery. The behaviour of each of these different classes of data is built in to the database system. Chronological data is stored using the fuzzy date mechanisms referred to above. The programs that access the data are very simple, they only have to be able to draw lines, or display images or text. The real work is done by the database system and the programs seen here can be written very quickly. Similar programs might have taken many months to develop using conventional programming methods, but these were produced in a matter of days.
What then of the future? The most immediate developments will, I believe, come from steadily improving access to information. Advances in database technology will make it possible to build information systems to support the full range of archaeological activity from planning, through fieldwork or excavation, to archiving and publication of results. Indeed, they should help to make archaeological archives more widely accessible, including the possibility of presenting the site archive in various forms suited to professional and public alike.
In the last couple of years we have seen a huge increase in the use of the World Wide Web as a medium for publishing archaeological information, including recent and current excavation and other research, electronic journals and virtual museums. Much of the material on the web is static; the user merely accesses information that has been prepared like the pages of a book. Dynamic web pages that respond to the users enquiries are becoming more widespread. These act much more like conventional information systems where the user can compose queries and retrieve the required information from a collection of data that is too large or too complex to present in its entirety. Many of these dynamic web pages are generated by database systems and, increasingly, they rely on one of the extensible systems that I have discussed above.
In fieldwork and excavation archaeologists frequently need to refer to information from beyond their own geographical area or chronological period. The Archaeological Data Service (ADS) is a recently established consortium of information specialists from several UK universities and based in York. Its aim is to provide a distributed information resource that will enable researchers to locate and retrieve information irrespective of where the data is actually stored. Initially this is intended as a resource for researchers, but we can envisage similar developments that might use summary information from these sources to produce on-line tourist guides and other material for a wider public.
One of my current projects is concerned with exploiting mobile, hand-held, computers both for recording and accessing information whilst in the field. Such a device might be of benefit to archaeologists and public alike. With a mobile computer the archaeologist can take information into the field and we can envisage tourist information offices offering local guides on plug-in cards. Add a Global Positioning System (GPS) receiver and the system can provide information about ones current location. In effect, a position-sensitive guide book. Then by adding a communication device, such as a mobile phone, it will be possible to obtain up to date information at all times. For the archaeologist this offers the possibility of direct access to the computers back at base and, of course, the direct input of field data as it is collected.
Computer based information systems have steadily transformed the ways in which many archaeologists work. Not everyone has been converted for both costs and conservatism remain serious obstacles. However, I hope that I have succeeded in showing that there is still considerable scope for development and that research in this area offers a stimulating prospect.