The authors have previously written in this Newsletter about the use of XML in archaeology ("XML: The Way Forward," Tyler Bell, XVIII.2 [Fall, 2005] http://csanet.org/newsletter/fall05/nlf0501.html; and "XML, Databases, and Standardized Terminology for Archaeological Data", Harrison Eiteljorg, II, XVIII.1 [Spring, 2005]: http://csanet.org/newsletter/spring05/nls0501.html). Each took a rather extreme position for the sake of making strong positions clear, and each ignored some issues that beg for a common ground. We write now together to acknowledge and explore that common ground.

First, we agree that standard relational databases are the preferred mechanisms for organizing, gathering, and storing archaeological data from a project. We do not necessarily have the same reasons for that choice, but we agree that relational databases are most appropriate for recording archaeological data.¹

Second, we agree that terminological ambiguity is the critical enemy to the use of archaeological data by the archaeological community, broadly defined. That is, unique, bizarre, even ridiculous terminology may be used by members of any project team. Assuming all members of the team use the terms regularly, there is no loss of function or understanding. However, the use of ambiguous terms prevents others from understanding and using the data correctly. Terminology matters enormously.

Third, we agree that access to data from other projects and collections will only get more important over time. Archaeologists must have access to the information from other projects if they are to understand what they find. In addition, it is likely that more and more work will be carried out on data from old projects to bring the information up to date, and doing so will demand checking with as many other data sources as possible.

Fourth, we agree that terminological ambiguity is, to some extent, a given in archaeology. Archaeologists will continue to give lip-service to the importance of controlled vocabularies, but they will also continue to ignore them for their own projects in many instances. That does not mean that seeking terminological clarity is useless or a waste of time. It does mean, though, that the most clear and pressing need for carefully defined terminology lies in the realm of data sharing and transfer. Absent well-defined terminology, data exchange can only be accomplished by downloading full data sets and working through the files locally or by spending enormous sums and enormous amounts of time to build unique, complex access systems for each and every data set.

Fifth, we agree that, even with the aid of well-defined terminology, data exchange is neither trivial nor automatic. Indeed, it will remain fraught with problems in a field such as archaeology because even good, well-defined terms must be mapped to the less-well-defined terms used in individual projects.

Given the foregoing, we believe that it is incumbent upon all archaeologists to work to make possible the effective interchange of data. What does that actually mean? Above all, it means that scholars must agree to involve themselves in terminology projects that are relevant to their specialties so that complete and complex, multi- lingual terminological systems can be developed. In this there should be no illusions. The work, when national boundaries must be crossed; not to mention the disciplinary boundaries separating those working in Medieval Britain, Romanesque Gaul, eastern European Iron Age, Roman Italy, Periclean Athens, Hittite Turkey, or Assyrian Iraq; will be slow and difficult, and there will be changes on a regular basis, even as the various areas of interest are represented by more mature efforts. Nevertheless, the work must begin -- more properly it must continue, since there are existing efforts upon which to build.

But what about XML, the subject with which this debate began. On this we do not yet fully agree.

Eiteljorg sees XML as a useful intermediary (not necessarily the only one) by which to carry queries -- using well-defined terminology -- from one computer to another, a way to make queries adhere to a common form so that they can be well and accurately understood. Importantly, Eiteljorg is concerned that XML confuses matters by making it seem as if XML were the important matter when, in his view, terminological clarity is the first priority. For that, he does not see XML as required.

Eiteljorg also sees XML as a very inefficient way to transmit query results, and he worries that a reliance on XML may encourage people to use inefficient systems when better ones could easily be devised. In the end, however, Eiteljorg certainly sees XML as an acceptable way to apply common terms and to do so usefully.

Bell sees XML as the single greatest tool for archaeologists since the advent of the trowel. XML gives us the means to address so many of the issues that remain extremely problematic for archaeologists in regards to data storage, exchange and preservation. His enthusiasm is checked, however, by the understanding that XML is only a technology and is therefore a means to an end. It is not a panacea that will solve all.

At this moment, XML is not the issue of prime importance. The terminological systems that are required to make data exchange possible are the far more important matters upon which we must focus first. The existing groups that have been working on these issues must not only continue; they must also help form other groups and expand their own membership. In addition archaeological organizations of all kinds must act to assist in the work of terminological clarity. Data interchange depends first and foremost on it.

-- Tyler Bell
   Harrison Eiteljorg, II

To send comments or questions to either author, please see our email contacts page.

---

1. Bell takes the view that the major advantages of relational databases are practical. There are many good products and many archaeologists who know how to use them. Eiteljorg views relational databases as providing the most effective data-organization possibilities for archaeology. Return to text.