-Harrison Eiteljorg, II
The title of this session, “Delivering Archeological Information Electronically,” suggests to me two rather different issues — electronic delivery, and electronic or digital archeological information. I would like to use that dichotomy to discuss what we have heard, and not heard, in this session.
First electronic delivery, the topic that received the lion’s share of our attention — appropriately enough, given the way the session was organized. It seems to me that we have had a superb discussion of the state of the art today, from the very practical approaches of Mary Carroll and David Carlson to the more theoretical approach of Christopher Chippindale.
Christopher Chippindale was the only one who mentioned the possible desirability of hybrid publications — partly on paper and partly in electronic form. I am inclined to agree with him that such publications will be more common than we have thought.
Donald Sanders and Richard Leventhal talked about the process of taking an archeological work and transforming it into something unique to the electronic world, complete with complex and wondrous virtual reality effects. Donald Sanders spoke at some length about the questions of authorship and authority raised by the process, and it seemed to me that he showed well how the problems could be dealt with — not overcome — since the electronic world must involve different realities, but dealt with effectively.
Richard Leventhal showed how many differing data types can be combined in electronic resources to produce a unique document. He pointed out how many different skills — and the people possessing those skills — would be required to complete an electronic publication.
Jim Farley and Peter McCartney showed us a different kind of electronic delivery — in their cases less publication than direct computer access to primary data. In Jim Farley’s case, the data are GIS data sets, and the point is to provide a system that will make direct access to the data themselves possible, with little mediation, for the wider public. As a result, some issues of standardization are very important, and those issues figured in Jim Farley’s presentation. Fortunately for that project, GIS data are very important in the commercial marketplace. As a result manufacturers of GIS programs are being — and will continue to be — driven toward standards for commercial reasons that do not apply for very many other data types. Producers of database management systems and CAD programs do not have equivalent commercial needs to cooperate in seeking agreement on standard data access systems. Nonetheless, the planned access to GIS data provides a good example of the kinds of standards many of us would hope to have, not only for GIS but for many other data types as well.
Peter McCartney (and I when representing the Archaeological Data Archive Project) cannot now provide the kind of access that Jim Farley is preparing for GIS. That may come but, in the meantime, Peter McCartney indicated his concern to preserve primary data so that access can be provided — and possibly easier access like that envisioned by Jim Farley in years to come. Peter McCartney also stressed the importance of metadata, the data that will make indexing possible and, consequently, permit us to find the information we need on the Internet. Metadata also will help future users of the data files by providing authentication, information about file types, migration history and so on.
Terry Childs discussed the work of the Federal government in presenting digital information. She and David Carlson were the only ones to discuss listservs, and that brings up the unstated assumption in this session — that electronic delivery equals the Web. Given the experience we have had thus far with computers and the Internet, I think that it is very dangerous to make that simple equation of electronic delivery and the Web. If there is one thing of which we can be sure, it is that there is no way to predict the nature of the computer or the network very far into the future. Therefore, I think we must watch ourselves and try to think more broadly about the kinds of electronic delivery that may be in use in the future.
Terry Childs emphasized the need to consider the audience when designing information presentation systems. The government has many potential audiences; it must deal with all those constituencies, and the experience gained there can be useful to anyone offering information over the Internet.
David Carlson had a very specific audience in mind — students in undergraduate classes. David’s practical suggestions were valuable and helped to remind us that one must very carefully tailor the message to the audience.
Hugh Jarvis, speaking about the use of electronic publication for journals, made the strongest statements about the coming dominance of electronic dissemination of archeological data. While I am not sure I would agree entirely, the basic point is incontestable. Whether we prepare or not, we will soon see enormous increases in the quantity of material presented electronically.
I have saved my colleague and co-chair of the session, Mary Carroll, to last. Mary Carroll reminds us that we must not only plan carefully but also be clear about our goals. Then, goals in mind, we must measure our progress. We dare not march blindly forward without examining the results of our work. This may be the most important message of all. As important — and enjoyable — as it is to work in these new and fascinating areas, the point is to accomplish specific aims. We need to be sure we understand our aims and can measure our achievements. Then we must honestly do the measuring.
Digital Archeological Information
Now let me turn to the second of our issues — electronic or digital archeological information. It seems to me that we have given this area less attention — less than deserved and required. What kind of information is appropriate for what purpose? In particular, I am concerned that we may be tempted to revel in the possibilities of new and impressive “gee-whiz” technology, whether it is useful and economically justifiable or not. I am also concerned that even appropriate, less exotic technologies may be used in ways that are as potentially harmful as helpful.
I would like to approach this issue from two angles. First, what level of technological wizardry is appropriate for what purpose? Second, when we let computers work for us, are we aware of the potential problems that may accompany the benefits?
We start with the question of appropriate technology, and I begin with a story. Yesterday I went to a west Chicago public school where my daughter teaches. I took slides of an excavation so that I could talk to fourth- and sixth-grade students about archeology and a particular excavation. The promised slide projector turned out to be a non-functioning filmstrip projector instead. I was obliged to talk for better than an hour to each of two groups with only a blackboard as a visual aid. Despite the absence of slides, I was able to keep the children’s attention. Now none of us would choose to make a slideless presentation of an excavation, but my point is that the tools we expect may not be as necessary as we think.
Put that in an electronic context. If we are presenting information, do we need all the bells and whistles? For instance, does a scholar need a virtual reality presentation or only simple tabular presentations of data — or both? Does a student need the tabular data or the virtual reality presentation — or maybe just good renderings of long-gone structures — or all of those presentation types? If we are showing the results of a survey, do we need colorful maps and charts or something much simpler? I suspect each of us can come up with an example of computer overkill. Given the
costs — something almost all of our speakers mentioned — this is an important question.
With each sophisticated technology there is also the question of the appropriate extent of its use. Which bells and whistles are needed at any given moment? While it may be useful to show a house as a virtual world in which a viewer may move, at least metaphorically, the same person viewing the same excavation may not need to see every pot (or any pot) with that same level of realism or to have other objects from the excavation available as virtual objects to be lifted, moved, and rotated. Virtual houses or pots are only stand-ins for the real things; after all, how close to the real items must the stand-ins be? The issue, once again, is partly cost. When does a virtual reality pot serve a purpose so necessary that the expense of its creation is justified?
The use of extremely realistic presentation systems brings an added problem: how clear is the distinction between the real item and its stand-in if the representation is photographic in its realism? Of course, that distinction becomes more and more important as the proportion of restored material increases and the proportion of real, extant remains decreases. Highly realistic presentations of mostly restored artifacts, buildings, or sites can be hard to resist — even if they are not supported by the evidence.
There may also be problems with the use of well- and appropriately presented data. Imagine CAD models, databases, or GIS files as parts of electronic publications. In order to use the data effectively, a user will need some skill and experience with the appropriate software — CAD, database management systems, and GIS programs. Yet how many of those here in this session concerned with technology have experience with all of those program types? Too few professionals — and fewer graduate students — have enough familiarity with different software types to access effectively the many different types of digital data that are available. If that is true, what should we publish or make available in digital repositories and when should we begin to do so? Should we wait for the technological sophistication of users to catch up with the technology, or should we make available extremely sophisticated types of computer data on the assumption that the users will ultimately learn enough to be able to use the data we store today?
As Mary Carroll pointed out, we need to specify our aims and to measure our progress. That is as true of the kinds of data presented as of the forms of presentation.
Now let us turn to the second of my concerns about digital data. I believe there is a problem — or a potential problem — with the way we may obtain and use digital data from computerized repositories. Peter McCartney and Christopher Chippindale talked about primary data and capta — I take the terms to be equivalent (and I like the term capta). The next level of information is those capta with context — something Peter McCartney called information. Information in a wider context may become knowledge, and knowledge may then be summarized. What is probably obvious to all is that the summary from one level of study becomes the capta for the next level, and the summary from that level becomes the capta for the next. The cycle will repeat itself endlessly, and at each step — from capta to information to knowledge to summary to new capta and so on — there are rules, processes and assumptions that guide the transformation. Those rules, processes and assumptions are crucial to the whole enterprise; they determine the outcome of data searches, aggregations and summaries. The potential problem lies in the fact that the nature of the transformation may too often be ignored in the haste to obtain results. However, the nature of any transformation must always be clear, explicit and transparent. That is, the scholar must always be able to know how information was generated, starting with the capta at the very beginning of the process, and including all the transformational processes thereafter — every rule, process and assumption that participates in the data transformation. Most scholarly publication, after all, is an attempt to show how explicit transformational processes allow certain information to be transformed and made meaningful in a new context; the nature of the transformation is often the point at issue.
I worry that computers make it far too easy to assume that the transformational processes are computer-generated and, consequently, above reproach. We cannot permit that. Each and every process in this transformation chain must remain not only explicit but also transparent, and we must train ourselves to pay as much attention to the processes that generate the information as to the information itself. Otherwise, we risk having and using information that has been created by processes that are unacceptable to us or based on rules or assumptions we do not accept – perhaps stemming from old and outmoded models, possibly using inappropriate statistics, maybe involving data aggregation assumptions we believe to be faulty.
Digital data, then, must be treated with great care. The chain of creation must remain with the data, and potential user(s) must always be aware of the way the chain was forged. One bad link will render the remainder useless. Those of us who provide digital data must take great care to be sure that we are able to provide more than the data. We must be able to provide an “audit trail” that connects the first capta to the latest summary so that the archeological information is not only digital, it is also understandable, reliable and useful.
Like archeological information on paper, the archeological information we get in digital form must
be both appropriate and trustworthy. The forms we can obtain must be the right ones for the job — cost effective and usable. At the same time, what we receive — capta, information, knowledge or summary — must be dependable, resulting from transformations that we can understand and evaluate.
In sum, we have seen here many interesting perspectives on electronic delivery of archeological information. It is clear that we will be using electronic forms of delivery more and more in the future; it is clear that we can do many new and exciting things in the process; it is clear that new paradigms must be developed. At the same time, it is equally clear that issues of cost remain to be determined in many areas, that we must resist the temptation to use the technology for its own sake, and that we must consider the skills required of users. Finally, we need to be more aware of the potential problems of providing data that have been manipulated by computers for us — but not in ways permitting and requiring our inspection and examination. In each of these areas the key to moving successfully forward may be found first in having real aims and goals, second in planning appropriate measurements of our results, and third in carrying out the measurements. Some of these issues, however, require more open discussion among interested scholars — particularly the issues surrounding the question of appropriate digital forms for archeological data. Perhaps these issues would be appropriate subjects for another SAA session.