Archaeological Information Management - Idaho State …



Archaeological Information Management

E.S.Lohse

Archaeological Collections: Electronic Databases

Contemporary archaeology must move from simple artifact collection to a charge to collect, organize and distribute archaeological information. Excavations, testing and survey produce artifacts and documentation that constitutes archaeological data. This data must be harvested, salvaged and used for research and education purposes if archaeology is to fulfill its potential of informing us about ourselves in the past and in the present. The traditional practice of carefully storing special artifacts away in secure drawers behind locked doors is no longer viable. Archaeologists are no longer working on shoe-string budgets on individually designed research projects. Archaeology is now funded by nation states and corporations, with contracted work mandated by laws and regulations. Research requires permits and consultation with governmental and advocacy groups. The scope of research is directed by contractual agreements and resulting information is required to be published and distributed within approved venues. Old work must be integrated in new work, and old and new collections must be securely and carefully stored for future generations. This paper addresses the current state of archaeological knowledge handling, and emphasizes that archaeologists must move aggressively to computer information systems as the primary way to curate and distribute archaeological knowledge.

Stewart (1997), in summarizing the course of archaeological data moving to information, identified the “Great Chain of Being” in archaeological computing, moving inexorably in logical stages from data collection, to data management, to data analysis, and on to dissemination. Stewart notes, however, that we can appraise archaeological use of IT and information systems as more of a complex, multi-stranded web than as a linear feature on the computing landscape. Archaeological IT venues now cover quantitative methods, statistics and classification, archaeometry, visualization (including imaging, CAD, multimedia and virtual reality), expert systems, artificial intelligence, and GIS. All depend on effective design of usable digital archives and databases. This assessment should not ignore enhanced education and publication venues as well (for a history of the development of archaeological computing, see Booth 1995; Hansen 1993; Lock 1995; Reilly and Rahtz 1992; Webb 1986).

The term “database” became common in archaeology by the 1980s, supplanting the earlier label “databanks” (Moffett 1984). Wilcox (1978) made one of the strongest early pleas for use of personal computing and microcomputers to aid in site recording and data retrieval, data analysis, and publication. Booth (1982) and Stewart (1980) published some of the first articles on use of microcomputers and relational database management systems. By 1988, Nick Ryan published a bibliography of over seven hundred publications in eleven different categories on computer applications in archaeology (Ryan 1988). By the 1990s, numbers of articles on a wide range of computing venues expanded dramatically, causing Ucko (1992) to speculate on the dramatic impact of computers and IT on archaeology (cf. Cheetham and Haigh 1992; Hansen 1993). Today, computerization is a given, and there has been a decided shift toward use of industry standard software and evolving standards for data recording and retrieval. At the same time, government organizations are investigating moves toward encompassing strategies for digital data management (Booth 1995; Clubb and Startin 1995; Clubb and Lang 1996; Murray 1995). The shift is away from straightforward data collection to increasing public access (Booth 1996).

Stewart (1997) argues that the central issue to emerge in over twenty-five years of archaeological computing is whether there has developed greater physical and intellectual access to archaeology. We still are struggling to bring archaeological collections as digital archives into arenas of ever greater professional and public access. The backlog of information seems daunting, and though successful pilot projects can be cited, we still are unsure of our level of success (cf. Stewart 1995; Niccolucci and D’Anrea 2002).

The Need for Informatics and Data Theory

These and other efforts are attempting to bring information online to support scientific communication across disciplines and across the world. Accessible databases are key, and accessibility is directed by success in defining standards and guidelines for practice. Archaeological databases today cannot be static compilations of text fields and numbers. Databases must strive to bring collections and documentation information to users in dynamic environments. We must also design these databases to be sustainable over the long term.

The use of electronic media to enhance communication is a major shift in the conduct of science. Pure access to information is a boon to scientists as is their ability to handle massive amounts of information. Cross-disciplinary and international collaborations are booming. Keys to nurturing this shift in the computer realm are building adequate metadata, migrating data and controlling access to information. Kling and McKim (1998) express concern over unsupportable risks rising if this transformation in scientific communication occurs in a pure laissez-faire environment. They point out that we cannot assume that “everyone will catch on” to using e-media structures eventually. They also argue that we cannot simply assume that various e-media initiatives reflect a creative period of problem-solving. Perhaps these developments instead reflect runaway agendas and proprietary interests that will eventually retard use of powerful electronic venues. Kling and McKim are informatics analysts, and view a lack of theory guiding this process as potentially hazardous. They note that huge amounts of money, resources, and effort are being committed by government agencies, by private firms and organizations, by academic departments, by publishers, by professional societies, and by individual researchers for the development, maintenance and promotion of variable communications technologies for the sciences. At the same time, scientists and policy-makers have no accepted theory of how scholarly fields should adopt and shape technology. Producers and users instead tend to work within context-free models. Often, the result is ongoing prototyping and fledgling projects with high promise and withered funding (e.g., Eiteljorg 2003). This result wastes funding, denies the efforts of researchers, and highlights the serious plight of data orphaned and dying in marginal, decaying, dead systems and formats.

A powerful response to this lack of an information paradigm is the so-called electronic communication reform movement, which has been recognized in the field of e-publishing, but these concerns extend across all e-media(cf. Kling and Iacono1995; Iacono and Kling 1996). Core instigators include a range of well known scientists (e.g. Paul Ginsparg and Paul Harnard). Harnard is an outspoken advocate of radically decentralized scholarly publishing that may or may not be peer-reviewed (cf. Brent 1995), and is the editor of the electronic journal “Psycholoquy.” He is also the originator of the concept “scholarly skywriting,” where scientific communication is confined to short, discursive, iterative bursts of e-communication (Harnad 1991). Ginsparg was a developer of the Los Alamos National Labs Physics E-Print Server, a working paper server used by high-energy physicists. This venue has found acceptance in the communications system of that field (Odlyzko 1996). Many in the scientific e-communication audience are accepting the admonishments of outside-the-box thinkers like Harnad and Ginsparg, prompting analysts like Morton (1997) to identify a paradigm shift toward electronic communication and away from hard-copy journals and archives, in all forms, whether centralized or decentralized.

There is a shared ideology in this movement. The basic precept is that electronic media is better than traditional media: e-communication will be less expensive, access to e-media will be easier and wider, and systematic use of e-media will dramatically speed up scientific communication (Kling and McKim 1998:2; Brent 1995). Examples of actions subversive to the established publication industry abound in the field of e-communication. The editors and organizers of the “Electronic Transactions on Artificial Intelligence” (ETAI) have created an open article review process (ETAI 1997; Sandewall 1998). ETAI reviews in two phases: after submission, the article is open to public online discussion for a period of three months; after author response, the article is reviewed for acceptance using confidential peer review and journal level quality criteria. Another AI journal, the “Journal of Artificial Intelligence” (JAIR), emphasizes online appendices and discussions of published articles. Scholars are encouraged to cite JAIR’s articles as they would articles in paper journals, but JAIR is distributed without charge on the Internet (Kling and Covi 1995).

The overriding characteristics of the e-media explosion in science communication are differences in structures, roles and uses from one discipline or field to another. Designing electronic applications in this new age will require close attention to the social contexts in which scientists operate. Scientists in different arenas use e-resources very differently in the conduct of their basic research and communication (Kling and Covi 1995; Walsh and Bayma 1996; Finholt and Brooks 1997; Kling and Covi 1997; Walsh and Bayma 1997). Examples abound. The discipline of Particle Physics uses the E-Print server at Los Alamos National Labs, and preprint servers at CERN, DESY and the American Physical Society. Biologists tend to depend on broader access supplied by publication of papers in archival journals, but researchers use digital databases like the Protein Data Bank, Flybase, and Saccharomyces Genome Database, as repositories for genomic sequences published in refereed journals (Letovsky 1998, 1999). The discipline of information systems has created ISWORLD, an extensive, distributed web-based collection of links, papers, course syllabi, tools and resources, sponsored by “MIS Quarterly.”

Not all disciplines have followed the lead of particle physics in opening up papers online. The American Psychological Association (1996) admonishes authors not to place papers online at any stage. The American Chemical Society (1996) has a similar policy regarding papers considered for publication in its journal. Not all scientific societies are as strict with authors. The Association for Computing Machinery (1995) copyright policy states that authors retain rights to their work including unlimited reuse of the work with citation of ACM. The ACM sees its role not as sole provider but as facilitator of information access. Similar policies are held in many computer applications journals and newsletters in archaeology, such as “Internet Archaeology” (Internet Archaeology 2003; Richards, Heyworth and Winters 2000; Winters 2002), and in the major professional journal “Antiquity” (cf. Champion 1997).

All e-forms, including pure e-journals, databases, and preprint servers are highly valuable for communication within and across disciplines. Disciplines have different strictures but access to data is a primary concern (cf. Mark Bide and Associates 2000). For our discussion, it is important to note that friction arises between disciplines emphasizing use of shared, static-but-growing databases like archaeologists and biologists and disciplines that do not like computer science. Commonality is in wanting to access data in shared electronic environments and in rapidly and easily communicating with other scholars. These kinds of questions represent a growing body of research about information technology and social change called Social Informatics (MacKenzie and Wajcman 1985; Silverstone and Hirsch 1992; Williams 1997).

The social shaping of scientific communication systems includes: access to resources including data; speed of work and results-sharing; selection of target audiences for research, allocation of credit for work performed, and allocation of professional status (cf. DiMaggio and Powell 1991). Kling and McKim (1998) acknowledge that trust plays a central role in the articulation of all issues within all disciplines. Scientists will only use a report if they are assured that it is legitimate. Formal peer review is only one traditional process of legitimation. To be willing to share information, scholars must have confidence that dissemination will not harm future access to resources or their career enhancement.

Important issues in determining how actors in social systems work focus on required research project costs, mutual visibility of the ongoing work, the degree of industrial or corporate integration, and the degree of concentration of communication channels. Costs can affect communication structures by requiring greater collaboration. Costs may also increase visibility at the expense of imposition of greater and greater control. Visibility is always a concern. Practitioners in some disciplines like Particle Physics share research results online before publication. In other disciplines, researchers are not aware of colleagues’ work because of sanctions requiring publication in formal peer-reviewed settings. Increasingly transparent distribution systems may cause actors to perceive lower risks correlated with sharing reports and data. Market trends buck this revelation, however, since industrial or corporate support, while a commendable boon to accelerated research and development, typically creates authoritative, owner-driven sanctions on information dissemination. Here, systems are opaque, hidden behind secure doors. Release is carefully controlled, if allowed, and timing is geared to corporate advantage and profit margin. Much of what we now do lies between these two poles: open access and controlled access (cf. Williams 1997).

Database projects have been particularly prone to boom and bust cycles. Development is prompted by high promise and demise is driven by low scholarly acceptance or limited funding accrual. Star and Ruhleder (1996) record the creation of an online “Worm Community System” for molecular biologists, which proved too complicated and technical for the larger community of its users. It was abandoned and recast as the web-based system “A.C. Elegans DataBase (ACEDB),” which has won greater acceptance. Letovsky (1998) reported that many biologists initially invested in the “Genome Database” (The Genome Database 2003) Human Genome Project,” only to see financial support withdrawn because funders did not see adequate value for their constituencies. The demise of the heralded “Archaeological Data Archive Project” (Eiteljorg 2003) is an example for our own discipline.

Archaeological Data: Practical Constructions

Perring and Vince, in an online project outline titled “Liberating Archaeological Data” (1999), set out an online guide for bringing archaeological data out to view. The expressed aim of their proposal is to describe ways of using IT to facilitate presentation and dissemination of complex archaeological data. They admonish that this will unlock research potential for classes of data and will encourage more amibitious approaches to the study of archaeological sites, landscapes and assemblages. They cite Hodder (1998) on how the Internet impacts organization of archaeological knowledge, allowing a shift from hierarchical structures to open networks and flows. Though these impacts have been documented, the majority of the archaeological community seems unaware of the implications of these new approaches. This may be a direct result of the reality that most archaeological research is initiated under national, federal and state mandates that do not favor innovative approaches. Archaeologists would seem to be running an appreciable risk: that establishing a commercial basis for archaeological research will fossilize research practice along traditional lines.

Archaeological typology practices and classification methods are obvious areas where innovations in thinking will have to take place, gearing to producing optimal structures in the organization of archaeological data that will be accessed and distributed in electronic networks. Interpretive structures will have to be devised which permit variable ways of grouping data. For instance, higher order groupings structuring archaeological data (phase compilations and typologies) will have to be supplemented by alternative analytical divisions of the data (e.g., functional classes irrespective of material type; deposition classes; competing spatial and sequential boundaries). Different conceptualizations of data structure will be required to develop flexible and analytical data structures.

Perring and Vince (1999) list obvious shortfalls in traditional database constructions. Results from excavations of complex archaeological sites are notoriously difficult to study, and both published and “grey literature” reports commonly report on only a fraction of the available evidence. Traditional interpretations typically follow a rigid linear framework based on chronological groupings cast from stratigraphic analyses, providing concrete narratives for successful publication, in which structures, sequences and assemblages are built on reconstructed landscapes. Problems inherent in recasting these structures to electronic databases include: post-excavation procedures are limited by intractable data sets; important data remain inaccessible to the research community because revising the structure is simply so expensive; there is a lack of integration allowing descriptions of different data classes published in specialist reports to be linked in the overall data structure; data structures are specific to individual investigations of specific sites and data classes, limiting potential for synthesis with other investigations using different data structures; research potential is effectively curtailed by narrowly defined data structures.

There are new methodologies that can be imposed in resurrecting and organizing effective data structures but without new conceptual organizing approaches these will only compound, not transcend the limitations noted. We cannot simply enter new data in old structures into IT developments (HTML, interrelational databases, and GIS), and expect to have working databases. The theory driven structure of the data must be revised, coupled with advances in IT.

The 2000 Society for American Archaeology session “Digital Data: Preservation and Re-Use” saw many of these issues addressed. Robinson (2000) summarized work on the “Digital Archiving Pilot Project for Excavation Records” (DAPPER). DAPPER was a collaborative venture between the Archaeology Data Service (ADS), English Heritage, the Museum of London Archaeology Service (MOLAS) and the Oxford Archaeological Unit. It focused on how digital project information could be most effectively archived, how best to deliver data over the Internet, what was the best costed model for archiving and delivery, how to assess user reaction to digital project archives, provide examples of best practice in archiving, and explore the close relationship between digital archiving and publication. Robinson points out that a traditional archive documenting an archaeological project would be transferred to a museum, although Swain’s (1999) survey of archaeological archives in England concludes that most museums do not have the technology to store, access and curate archives containing computer files. Condron et al. (1999) reinforces this assessment. DAPPER was a pragmatic approach to the imminent threat of systematic loss of electronic archaeological data.

Two large high profile collections were chosen as the ADS pilot study: the Royal Opera House excavated by the Museum of London Archaeology Service and Eysham Abbey excavated by the Oxford Archaeological Unit. Both projects were at the dissemination stage and had different sorts of digital archives to deposit. The projects were done by two different archaeological units with different working practices.

The Eynsham Abby digital archive contained text files, databases in comma delimited text, JPG images, a 3-D reconstruction of the medieval abbey, and digitized drawings. The Royal Opera House archive consisted of database files and GIS files summarizing context, artifact and ecofact attribute sets (Robinson 2000).

These data sets were designed to be accessed using the Archaeology Data Service ArchSearch Catalog. Site metadata records can now be searched using keywords. Flagged records can then be followed to more detailed catalog records. A Project Archive button in the left hand frame of the ArchSearch window allows the user to list projects with downloadable resources. Both the MOLAS and OAU have hotlinks to connect their own web pages to the DAPPER project archives.

A central issue in resource delivery was the user interface, and whether this should be aesthetically pleasing or simply allow clean access to raw data. Emphasis on usability was abandoned in the development phase because it was deemed very expensive and perhaps would have created problems in the realm of data migration. The compromise was to present data in standard formats with online support documentation to spur data reuse. This was seen as sustainable and more cost-effective.

The Archaeology Data Service receives core funding from the UK Higher Education sector and aims its data resources at the scholarly community. The ADS has sought funding to repurpose the DAPPER raw data into Internet deliverable teaching modules to enhance future use of the digital excavation archives.

Cost was summarized by Robinson (2000) and is an interesting footnote to our discussion of digital archiving. Robinson reports that digital archiving of Eynsham Abbey cost 1.2% of the excavation and post-excavation budgets. Digital archiving of the Royal Opera House collections cost .1% of the total project cost. The cost difference is attributed to differences in analysis during the post-excavation stage on the projects. For example, digital stratigraphic drawings are available for both sites, but the ROH-MOLAS excavation used GIS, creating three deposit files, while Eynsham Abbey, done in CAD, deposited 404 site plan files, 80 structure plans, 15 phase plans, two sections, a trench plan, a composite plan, and a 3-D reconstruction. Robinson notes, however, that it is not a question of whether GIS or CAD is most cost-effective. The GIS archive requires special training to use. The CAD archive is constructed like a digital publication, with separate structure and phase plans. The GIS database is a more powerful research venue while the CAD archive is the better nonspecialist venue.

The Archaeology Data Service has recognized that the Royal Opera House and Eynsham Abbey projects are showy, high profile, well funded excavation projects, and that many archaeological projects will not merit this level of attention and expenditure. ADS-DAPPER defined four different levels of digital archive. The index level archive or minimum digital archive, where the project did not merit further assessment work. This contains an index record for the ADS catalog and a site summary document. The assessment level archive or larger digital archive, which maintains an index record, the project design, the assessment report, specialist level databases, and a site matrix. The research level archive, a large, complex digital archive, warranted when the project is seen to be significant for analysis beyond the assessment performed. This archive is not integrated into the final project report and holds the results of analytical and publication process. It will contain an index record, a project design, the research database with stratigraphic databases, digital plans, site matrix, artifactual and ecofactual databases, the site, and artifactual and ecofactual reports. Finally, the integrated archive, which holds scholarship resulting from ongoing analysis of projects published as traditional archaeological monographs, linking texts seamlessly with the digital site archive. Users should be able to query the range of site data through various interfaces, including searchable relational databases and web-based GIS. The potential is to bring unwieldy technical data out of the excavation report and locate it in a universally accessible digital environment.

DAPPER constitutes the first functional online digital project archive. It created a costed model for the ADS in working with commercial archaeology. It also resulted in reassessment of post-excavation procedures and documentation methods implemented by the Museum of London Archaeological Service and the Oxford Archaeological Unit. DAPPER suggests that digital archiving can proceed for less than 5-1% of the total archaeological project budget. DAPPER has also received high numbers of visitors who have downloaded thousands of digital files. The DAPPER resources have also been used as teaching data sets in the Archaeological Information Systems Masters degree developed by the University of York.

Eitlejorg (2000) admonishes that the real goal of any digital archive project is to ensure reuse of the data. He considers three kind of digital data: databases, CAD models, and GIS data sets, chosen because they require access in a highly interactive digital environment. Eiteljorg in particular is concerned with the skills needed by users to access these materials, identifying four distinct levels: skills needed to use data files if the appropriate software is available; skills required if the files are not in the format required by the user; skills needed to evaluate the quality of the data; and skills needed to aggregate the data.

It is axiomatic that data will be stored in formats devised by other scholars. This data will have been organized according to the needs and perceptions of this original scholar, so the user must know something about how the original data were gathered, organized, entered and stored. For example, in a CAD model the user must know how the model has been segmented and how the data segments have been incorporated into the CAD layer names. For GIS data sets, the user must know the scales used for the map data, and how the data tables were constructed.

Data translations will be tricky for the user of the digital archive. An example cited by Eitlejorg is the DBF format used in dBase. Moving the data tables is straightforward but the user will encounter problems in complex databases with related or linked tables. These relationships are not included in the DBF format and must be specified in accompanying documentation. The user must have access to this documentation.

Data quality evaluations will be a constant concerns for digital archive users. Questions about whether the data is primary or secondary in nature will be obvious. The necessary scholarship framed in careful attributions will be required documentation. Questions concerning whether all potential data were reported on are pertinent. Was collection methodology systematic? How were corrections to the data made? Have these modifications to the original data sets been tracked? Evaluation of the digital data is a time consuming and complex task requiring high levels of professional expertise. Without adequate documentation the information needed for users of the digital archive is lost. As Eiteljorg (2000) notes, data producers and users must instruct the archives, even carefully designed projects like the Archaeology Data Service. Producers and users must provide the parameters for data manipulation.

Metadata that documents the data resource, specifying the information that the user needs to adequately use and explore the digital resource, is the required bridge between data and data use (Cartney and Robertson 2000; Michener et al. 1997).

Metadata

Metadata can be defined as data about data, and provides information essential to use of any database (cf. Quine 2001 on role of data standards). Metadata can refer to an agreed upon set of fields and associated lexicon or it can be a detailed description of measurements systems and rules for application. Metadata are required so that the user can make intelligent decisions in selecting, using, adding to, or translating a database. A library card catalogue holds metadata that allows users to find particular books. Maps include metadata as scales, dates of survey, and dates of publication. For electronic resources, there are an increasingly large number of standards. MARC, a Machine Readable Catalog, is used in library cataloguing (British Library 1980; Library of Congress 1994). The Text Encoding Initiative (TEI) allows standardized description of electronic texts. The Directory Interchange Format (DIF) provides metadata for satellite imagery (GCMD 1996). The U.S. National Spatial Data Infrastructure (NSDI) approaches complex descriptions of spatial data. Content standards have been defined for U.S. geospatial metadata (CSDGM 2003; FGDC 1997) The U.K. National Geospatial Database (Nanson et al. 1995) integrates governmental and nongovernmental spatial data. NASA has produced the Global Change Master Directory, which presents a writer’s guide for DIF - Directory Interchange Format (NASA 2003).

Miller (1997:4) characterizes these kinds of schemes as extremely complex and geared explicitly to creation by experts and interpretation by computers, rather than designed to facilitate information dispersal to as wide a range of users as possible. They operate within a narrowly defined field of work and are not suited to describe wide ranges of resources. Miller applauds eLib projects such ROADS and ADAM. The Arts and Humanities Data Service (AHDS) for archaeology, history, text and the performing arts is seen as particularly productive. Both AHDS and ADAM use the Dublin Core Metadata Element Set (cf. Miller 1996, 1997a, 1997b).

The Dublin Core has been developed to supply metadata descriptions between the crude metadata of search engines and the complex systems developed for MARC and the Federal Geographic Data Committee (FGDC 1994) (Dempsey 1996; Dublin Core 2002). The Dublin Core model can describe resources available on the Internet and can be used to insert a range of file types from simple HyperText Markup Language (HTML) to Postscript files and other image formats (Miller 1996; Weibel 1996). The Dublin Core consists of thirteen core elements, each of which can be extended by use of Scheme and Type qualifiers.

Scheme and Type qualifiers are used to better describe the resource. The Scheme qualifier identifies any recognized coding system used in the description of a specific Dublin Core element, and allows consistency and standardization. Scheme should only refer to an existing coding system such as the Internet Media Type (IMT) or to the International Standards Organization standard on dates (ISO31). The Type qualifier is used to identify, as in a name, email address, or the like.

Designers feel that any necessary extensions of the Dublin Core should be included in a separate framework as in the Warwick Framework (Lagoze et al. 1996). Descriptions stored there may be from a different metadata scheme such as DIF or FGDC, or could be simple extensions for the thirteen Dublin Core elements.

Databases

Metadata associated with a database should act to improve or restrict access to data, facilitate sharing and interoperability, and characterize and index data. Metadata designed to support data quality and longevity is the inherent concern of any database construction.

Rothenberg (1995) offers a succinct overview of the factors to be considered in database design and development, and focuses specifically on inherent problems in maintaining digital records as software becomes .

Rothenberg (1995) admonishes that data are a model of the real world, a description that is arbitrary and biased. Data models incorporate very different data views. For example, the speed of an object through a medium might be measured as slow - medium - fast, as a single numerical value as in 20mph, or as a table of numerical values. The choice of the kind of data to produce is decided prior to the building of the data model.

Rothenberg (1997), in discussing verification, validation and certification of data quality, notes that the assessment cannot be a binary value as in good or bad. It must instead be evaluated in the specific context of use, and seen as a desire to move from evaluation to improved quality. He notes two quality attributes: objective correctness (accuracy and consistency) and appropriateness for the intended purpose. It is axiomatic that data users will not be able to control data quality if data are taken from outside or intermediate sources. All data must be augmented with metadata to record information needed to assess data quality, record the results of assessments, and support process control. Producers must perform explicit verification, validation, and certification on the data, using metadata to direct this activity and to record results. Producers must also establish control over the data processes to improve data transformations, using metadata to support this activity and to record the results.

Rothenberg (1996:6-8) lists contextual categories for data quality that include adequate description and meaning, specification of intended use and range of purposes and constraints, requirements for access and use, description and rationale for structure or design, global relationships to other databases, and update cycle information. Source information includes identification of source and assessment of source credibility, characterization of classification, accessibility and reproducibility, and clear notice of release authority. Rothenberg (1996:6-8) lists other criteria, including data-element and data-value metadata.

Limited media life and rapid obsolescence of software and hardware highlight the need for concern with longevity of electronic data records. Increasing use of graphics, hypertext, linked structures and multimedia only accelerates this projected obsolescence. Data records are becoming increasingly dependent on specific software for continued interpretation. Data files will become increasingly useless without software to interpret the structure and meaning. Traditional hard copy documents and records have linear content that is relatively independent of this concern. For electronic media, record keeping paradigms are essential, and are evolving in direct response to accelerating software paradigm changes (Michelson and Rothenberg 1992).

We should save original records, and importantly, we need to save application software and descriptions of the required hardware environment as prerequisites of constructing emulators in the future. Compression should not be considered an option. Saved bit-streams should be copied verbatim. Annotated metadata must be transparent and act as a “bootstrap standard” (remember that ASCII will change to Unicode to some other standard) (cf. Rothenberg 1996).

Information Standards

Much of the current discussion on information standards in archaeology is headed by the Getty Information Institute and the International Committee for Documentation of the International Council of Museums. These two organizations sponsored an initial meeting on the urgent need for standards in Canterbury, England, in September 1991. Results of this overview were published in a brochure titled “Developments in International Museum and Cultural Heritage Information Standards,” first published in 1993 and updated in July 1995.

Standards are models that organizations, projects, and vendors can use as the basis for creating information systems and guidelines. These are rules for structuring information, enabling data entered into the system to be reliably read, sorted, indexed, retrieved and distributed between systems. The most compelling reason for standards is to ensure that the data will have long term value. It should be acknowledged that the largest investment in building any database is in the cost of assembling the data and the time required to enter them into the system. Further, all computer technology will change routinely and all systems will need to be upgraded and data moved to different hardware and software. Standards ensure that the database is consistent internally and permit data to be formatted and stored for export to other systems.

Standards can vary from strict forms to flexible guidelines developed for specific institutions. Three standards are commonly defined: technical standards are exacting (e.g.,

The ASCII character set of 128 codes defines alphabet, numbers, punctuation, and control codes for text processing and data communication); conventions are more flexible than technical standards (e.g., the MARC formats and Museum Documentation Association Data Standard); guidelines are broad sets of practical criteria against which products are measured (e.g., style manuals).

International and information standards recognized by museums and cultural heritage organizations constitute four main groups: information system standards define the functional components of the information systems used; data standards define the structure, content, and values that collections information comprises; procedural standards define the documentation procedures required for system management; Information interchange standards define the technical framework for exchanging information. (e.g., ISO 8879 or Standard Generalized Markup Language - SGML).

An infrastructure of agreed upon standards will make development of shared text and image databases much easier, will ensure quality, allow reuse of information, and create effective transfers of information. National bodies like the International Organization for Standards (ISO), the American National Standards Institute (ANSI) or the British Standards Institute (BSI) will publish and maintain standards once these are approved. Groups like the Getty Information Institute and CIDOC will focus interest and lobby for standards for particular disciplines and applications.

Selected guidelines for museum records development have been recognized by CIDOC as part of “International Guidelines for Museum Object Information: The CIDOC Information Categories” (CIDOC 1995, 2003a, 2003b). ICOM-CIDOC also maintains lists of museums and cultural heritage information organiations that concentrate on standards, and importantly, successful museum documentation initiatives or projects (ICOM-CIDOC 2002). The list of important international and national projects and organizations is growing routinely (cf. ICOMOS Documentation Centre 2003).

Australia: Australian Museums and Galleries Online (AMOL) is Australia’s online cultural heritage resource. AMOL offers details of museums, art galleries and historical societies across Australia. Over four hundred thousand item records represent thirty-six cultural heritage collections. Users are offered online museum forums, an online peer reviewed journal, and consultations in museum resource areas.

Canada:The Bureau of Canadian Archivists’ Planning Committee on Descriptive Standards has issued a data content standard, “Rules for Archival Description” (RAD) for Canadian archives. RAD is based on the International Standard for Bibliographic Description, and archival descriptive records based on RAD can produce records in MARC format. The Canadian Heritage Information Network (CHIN), a branch of the Arts and Heritage Sector of Communications Canada, maintains the national inventory of Canadian collections and offers services to museums, including automated collections management and advice on documentation standards and new technology. CHIN offers international access to various specialized databases on natural and cultural heritage through partnerships like the Conservation Information Network (CIN). CHIN is a participant in the International Committee for Documentation (CIDOC) of the International Council of Museums (ICOM).

United Kingdom and Europe: The Archaeology Data Service (ADS), a branch of the Arts and Humanities Data Service (AHDS), supports research, learning and teaching through dissemination of high quality digital resources in archaeology, offers technical advice to the research community, and produces guides for good practice (ADS 2003a, 2003b; AHDS 2003; Bergrie and Greenstein 1998; Burnard and Short 1994; Greenstein and Trant 1996; Richards 1996; Richards and Robinson 2000; Wissenburg 1997). The ADS maintains an ARCHSearch Catalog, an ARCHway of journal holdings from twenty-five UK research libraries, the HEIRPORT or Historical Environment Portal with resources from ADS (cf. Baker et al. 2000; Cultural Heritage Consortium 2002; Fernie 2003; HEIRNET 2002), the Royal Commission on the Ancient and Historic Monuments of Scotland (see Mowat 2002 for an overview of SCRAN), and the Portable Antiquities Scheme, and the Society of Antiquaries Library Catalog. ADS projects include “Digital Archiving Pilot Project: Excavation Records”(DAPPER) for English Heritage (1999), “Archaeological Records of Europe Network Access Project” (ARENA) for the European Commission, “Online Access to the Index of Archaeological Investigations” (OASIS) for RSLP and English Heritage, and “Publications and Archives in Teaching with Online Information Sources” (PATOIS) for JISC. Hunter and Ralston (1997) offer a cogent summary of archaeological resource management in the United Kingdom.

The International Committee for Documentation of the International Council of Museums (CIDOC) has over 700 members in sixty-five countries. It supports many working groups concerned with standards issues (Crofts et al. 2003). The Archaeological Sites Working Group collaborates with national sites and monuments organizations and the Council of Europe in developing standards for site documentation (Council of Europe 1995). The Documentation Working Group compares museum data standards and is refining a Data Model incorporating practical standards and reviews of terminology resources. An example project using this standard is the “Network of Art Research Computer Image Systems in Europe (NARCISSE). Other important working groups include the Iconography Working Group examining classification schemes for iconography and the Multimedia Working Group examining standards for application of multimedia technology. The Museum Documentation Association has produced SPECTRUM, the UK museum documentation standard (MDA 1997).

The Inventaire General des Monuments et des Richesses Artistiques de la France, under the French Ministry of Culture, is responsible for recording all cultural property in two databases: I-ARCHI, which inventories architecture and the built environment, and I-OBJET, which holds information on movable objects. The Inventaire has developed terminology authorities for Architecture, Objects Civils Domestiques, Le Mobilier Domestique, La Sculpture, Tapisserie and Vitrail.

The Istituto Centrale per il Catalago e la Documentazione (ICCD), part of the Italian Ministry for Cultural and Environmental Property, maintains a database holding the patrimony of Italy. The ICCD has produced catalog manuals and terminology dictionaries outlining data content and data value standards for documenting Italian heritage collections.

The Royal Commission on the Historical Monuments of England (RCHME) is responsible for surveys and records on the historic environment of England, and provides users with advice and information from publications and database records (RCHME 1993, 1998; RCHME and English Heritage 1995). RCHME has developed the “National Monuments Record” database, which indexes and correlates information on architectural and archaeological sites and archives. RCHME has developed and published information and data standards and guidelines for information levels recording historic buildings and archaeological sites in England. RCHME also publishes thesauri of architectural and archaeological terminology, and works closely with the Council of Europe Division for Cultural Heritage on development and promotion of core data standards for recording architectural and archaeological information in Europe.

United States: The Getty Information Institute seeks to make cultural heritage information more accessible through computer networks in collaboration with domestic and international institutions and organizations. The focus is on policy, standards, and practice. Initiatives have been developed to define issues of access and distribution of cultural information in networked environments. The Institute maintains research databases offering content on effective creation, maintenance and retrieval of information.

The Museum Computer Network (MCN) is a consortium of museums and individuals promoting excellence in museum automated information systems. MCN formed the “Computer Interchange of Museum Information” (CIMI) in 1990 to develop a standards framework for exchange and sharing of data in museum environments (Moen 1998).

The Research Libraries Group (RLG) maintains a project to develop a “Cultural Heritage and Museum Object Information Resource,” an initiative to improve access to information on works of art, architecture, visual culture, and materials culture held in museums, historical societies and other cultural heritage institutions. The goal is a web-accessible database of object records linked to visual images and associated texts. Data is loaded from the “Reach Project, the Getty’s “Provenance Index,” and the “Vision Project.” The “Reach Project” is an effort to create a testbed database of museum object records, where machine-readable data from heterogeneous museum collection management systems can be exported to analyze research value. The “Vision Project” for shared visual resource records, is a joint effort of the RLG, the Visual Resources Association, and the Getty Information Institute. Enhanced searching will use Getty vocabularies: “The Art & Architecture Thesaurus” and the “Union List of Artist Names.” The Council on Library and Information Resources, Library of Congress, has published a national strategy for digital archiving (NDIIP 2002).

The Hammer of Federal and State Regulation: United States

U.S. archaeological resource managers over the past decade have habitually referred to a “curation crisis” wherein many museums and repositories cannot accept new collections because of lack of proper funds and storage space (e.g., Thornbury 2002). The failure of ADAP and its electronic archive is a mirror of collection profiles in general in the U.S. Even securely stored collections of artifacts and documentation often do not meet Federal standards. They may not be stored properly, they may be at immediate risk for deterioration, or they may never have been completely inventoried, studied, or reported on. From our perspective, these collections cannot be easily gathered up as orphans and entered into an electronic database. Careful salvage of these collections is required, object and hard copy documentation needs to be secured and stabilized, metadata written, and the whole transferred to electronic form. This is salvage archaeology or data reclamation in its truest form, and the need for action is paramount.

Legislation pertaining to archaeology in the United States does create a priority in preserving the archaeological record but interpretation of State and Federal laws and regulations does allow considerable ambiguity. The Antiquities Act of 1906 asserts that all objects must be “properly cared for” but contemporary archaeological research recognizes the importance of records and metadata as well (Thompson 2000). Other legislation like the Reservoir Salvage Act of 1960, the National Historic Preservation Act of 1960, and the National Environmental Policy Act of 1969, installed further mandates on the collection and care of archaeological objects and collections, though not necessarily emphasizing maintenance and accessibility of hard copy records. Electronic archives were certainly not highlighted nor even envisioned. King (2000) admonishes that this body of legislation prompted a huge volume of archaeological research as part of large and small scale projects performed to mitigate development activities. At the same time, legislation and accompanying precedents and protocols failed to provide effective procedures for protecting the collected artifacts and the growing documentation.

The Archaeological Data Preservation Act (ADPA) of 1974 approached this problem by urging concerns regarding protection of these collections. ADPA stated that the Secretary of the Interior must consult with groups with the goal of determining ownership of archaeological materials and decisions on deposit of these materials in an appropriate repository. It also called for the Secretary of the Interior to issue regulations re: curation of federal archaeological collections (National Park Service 2000). The following Archaeological Resources Protection Act (ARPA) of 1979 strengthened procedures requiring permits to conduct fieldwork, asserted federal ownership of artifacts removed from federal lands, required collections to be stored in “federally compliant” repositories, emphasized written agreements, and asserted that the Secretary of the Interior should issue regulations on the care and management of archaeological collections (Carnett 1991; Cheek 1991; National Park Service 2000). Even with this emphasis, in 1987 the U.S. Government Accounting Office issued a report with the following findings: many repositories had no collection inventories, repositories had lost or destroyed records, collections had never been inspected for conservation needs, many had catalog backlogs, and there was insufficient storage space to match anticipated future needs (USGAO 1987).

The Code of Federal Regulations Title 36 Part 79 (36 CFR 79), “Curation of Federally-owned and Administrated Archaeological Collections,” 1990, mandated guidelines for preserving and handling archaeological materials and associated documentation Code of Federal Regulations 1990). It specifically charged federal agencies with determining the capabilities of curation facilities. The Society for American Archaeology responded in 1991 with a “Task Force for Curation,” which resulted in the document “Urgent Preservation Needs for the Nation’s Archaeological Collections, Records, and Reports” in 1993 (Childs 1995). These efforts were followed with formation of an Advisory Committee on Curation who publicly presented the curation issue in “Crisis in Curation: Problems and Solutions,” presented at the 65th Annual Meeting of the Society for American Archaeology (Bustard 2000). The Committee highlighted continuing problems in curation in the March 2001 issue of the “Society for American Archaeology Archaeological Record.” Immediate needs were cited: better integrated field collection strategies, enhanced dependable collections funding, improved long term care and maintenance of collections, priority in deaccessioning collections, accreditation of repositories, improved access and use of collections, and emphasis on public outreach and education through better access to collections and “grey literature” (Childs 2001).

Complicating institutional improvements in establishing care of collections and improvements in professional and public access to archaeological data was passage of the “Native American Graves Protection and Repatriation Act” (NAGPRA) in 1990. NAGPRA requires comprehensive inventory and repatriation of Native American human remains and associated funerary and sacred objects and objects of “cultural patrimony.” This legislation affected all museums receiving federal funding, and specified deadlines for compliance and inventory assessment as well as penalties for noncompliance (McManamon 1992). Repatriation issues moved to the fore, and federal and state agencies, universities, museums and other repositories laid aside many curation and conservation issues to immediately deal with this strident political charge. Issues of ownership and access restrictions served to derail much of the mandate established in 36 CFR 79 (cf. National Park Service 1997a,1997b, 2002; US Army Corps of Engineers 1999, 2003).

Thornbury (2002) summarizes State laws following changes in Federal legislation. By 1991, fifteen states had passed laws, regulations or policies regarding management of archaeological collections (Carnett 1991, 1995). By 1999, thirty-seven states cited curatorial issues as paramount, highlighting required improvements in museum accessioning and deaccessioning, imposition of curation fee schedules, and explicit loan policies. Both 36 CFR 79 and NAGPRA have resulted in museums carefully evaluating whether or not to assume responsibility for new accessions. Museums are now raising costs and requirements for basic curation. Collections are becoming more and more expensive to maintain. Unfortunately, dead storage or very limited access has been a popular solution en lieu of adequate funding. Bustard (2000) cites lack of funds as the major problem in the United States, noting that 36 CFR 79 provided standards for curation but did not secure sources of federal funding. Costs have risen in all categories: staffing, record updating, storage of artifacts, processing of loan requests, purchase of archival quality boxes and polyethylene bags and acid-free paper, purchase of computers and software, development of computer databases, and improvements in physical structures and climatic controls.

Access to Collections

Museums in the United States, and elsewhere, are challenged by the prospect of moving collection information online to provide greater access (Dunn 2000). Moving information online is a logical reaction to rising costs of curation, display, and onsite exhibitions. To be made available collections must be placed in management databases, these databases must include item level metadata for public access, and all improvements must aim at enhanced “resource discovery.” In a well designed environment the actual database should be invisible to most users. Research and development issues are many: collection level descriptions, standard terminology, secure but transparent access between organizations, across disciplines, among resources with different content, and for audiences with varying expertise.

Many museums and organizations have now brought collections online, often designing and managing their own websites. Many times these individual sites are linked through principal gateways. These contain item-level descriptions and rich images within distributed networks as in the Archaeological Data Service framework. The emphasis is on data sharing among organizations that are interoperable at local, national and global levels. The ideal is that collection-level descriptions follow a well designed standard, that these databases are automatically, dynamically created to match user requirements, and that they strive to be multi-lingual and provide semantic links between object and class. Collection-level description should provide access for both general and specific requests, regardless of user knowledge level, discipline, data requirements, or language of users.

Significant resources are being produced by the Consortium for Interchange of Museum Information (CIMI), which lists standards for museum resource description and sponsored the CIMI “Dublin Core Testbed Project” in 1998. In this effort, seventeen CIMI member organizations created object-level description using the Dublin Core standard. Primary problems were found in characterizing resources as item-level or collection-level. A “Guide to Best Practice” for museum collections was produced by CIMI in 2000. Examination of resource description frameworks have lead to goals to enable interoperability between applications exchanging metadata, and focused on enhanced resource discovery, cataloging, and collection-level descriptions. In this vein, use of Extensible Markup Language (XML) has been documented to offer significant improvements toward achieving cross-domain interoperability (Dunn 2000; Crescioli, D’Andrea and Niccolucci 2002; Schloen 2001). There is also a continuing emphasis on developing sound knowledge representation tools like thesauri for specific disciplines.

Other fine efforts are being produced by the Canadian Heritage Information Network (CHIN). CHIN is working toward developing terminology in collections-level descriptions that is specific enough to allow users to decide whether they have found an appropriate resource, but general and descriptive enough so that people from a wide range of disciplines and knowledge levels can discover the resource. For over twenty-five years, Canadian museums have been contributing object-level metadata to collective resource management led by CHIN. “Artefacts Canada” holds information on over two million objects accessed through “The Great Canadian Guide,” using specialized and general terms in English and French developed as the “Revised Nomenclature for Museum Cataloging” (Blackaby, Greeno and the Nomenclature Committee 1988). To avoid inconsistent nomenclature, CHIN has intergated Getty’s “Art & Architecture Thesaurus” with “Artefacts Canada.” The AC records are visible to online search engines by providing CHIN web pages with collections-level descriptions from the “Great Canadian Guide.” The Guide serves as an online gateway to over twenty-four hundred Canadian cultural institutions. It allows users to link collections-level descriptions to corresponding object records in “Artefacts Canada.” The linkages are not automatic and still not always successfu because of vagaries of class definitions. Controlled vocabularies like classification rools and thesauri are emphasized. CHIN has also produced “Learning with Museums,” presenting online educational materials using thesauri of subject areas based on Canadian school curricula. CHIN is currently developing an initiative on “Virtual Museums of Canada.”

Preservation of Data: ADS Guidlines

Archaeologists are good are creating data but are not good at arranging and preserving data in ordered, accessibly public archives (cf. Richards 1997). Archaeologists also tend not to be very proficient nor interested in reusing other peoples’ data. The United Kingdom heritage community has been a consistent leader in attempting to preserve collections and information about the past. UK research in archaeology over the past thirty years has been very concerned about the preservation and reuse of archaeological data. This has culminated in the mission of the Archaeology Data Service, a branch of the Arts and Humanities Data Service (AHDS). AHDS-ADS emerged from a 1994 feasibility study conducted by an Information Services Sub-Committee (ISSC) of the Joint Information Systems Committee (JISC) of the UK Higher Education Funding Council (Burrand and Short 1994). This feasibility report recommended creation of a central coordinating group to carry out management and user-support functions for maintenance and operation of a comprehensive heritage database. This organizational structure would support dispersed service providers for particular disciplines or groups of disciplines.

The AHDS was formally established by the JISC in June 1995, based in King’s College, London. Over twelve months, five discipline-based service providers were recognized: Oxford Text Archive (OTA), Oxford University Computer Service; Historical Data Service (HDS), The Data Archive, Essex University; Performing Arts Data Service (PADS), University of Glasgow; Visual Arts Data Service (VADS), Surrey Institute of Art and Design; and the Archaeology Data Service (ADS), University of York. Each of these service providers were designated responsibility for their discipline’s data and were held to develop standards definitions and guides for best practice for particular classes of data. There was an AHDS-wide emphasis on spatial data, including GIS, which was seen as essential for organizing all discipline’s data sets.

Envisioned was a distributed but integrated approach to data access. Metadata needed to address discipline specific needs and problems, that economies of scale required developing shared migration strategies for specific data types, and that a vision needed to be developed for how researchers would conduct cross-disciplinary searches from a desktop (text-objects-images). Reuse of data requires that users can locate the data they require, and that this old data can be accessed in contemporary formats (cf. Miller 1996; Wise and Miller 1997).

The explicit Archaeology Data Service goal was to collect, describe, catalog, preserve, and provide user support for reuse of digital data generated in the course of work by British archaeologists. The focus initially was on Britain but has shifted toward worldwide application, since there were no stated geographic boundaries, and there was an expressed interest in working with organizations in other countries to develop reciprocal archiving policies. Directing development is the dispersed model, with archives linked by a single gateway to provide integrated access to distributed collections.

ADS collections policy operates on two distinct fronts. First, facilitate access to existing archives and second, accession orphan data sets. Orphan data sets were common as residue of centuries of work recorded in national and regional Sites and Monuments Records. Digital archiving on this ambitious scale requires development of a national strategy. Various pilot studies have been developed in AHDS-ADS to work with a range of data types to develop “costed models” for preservation of digital data. These models must distinguish between issues of open public access versus stored and maintained data that has not been released. Key issues include negotiated access, copyright restrictions, and developer funding.

The Archaeology Data Service has moved to meet its mission through a strategy of information sharing embodied in production of its “Guides to Good Practice” series, which currently include guides to excavation and fieldwork, GIS, geophysical surveys, and satellite imagery and aerial photography. Other guides will address databases and sound and video images.

These guides to good practice perforce emphasize creation and preservation of appropriate metadata, enabling the projected user to locate required data and to assess data fitness. Metadata ranges from simply noting names of excavation directors to detailed descriptions of equipment and software used to algorithms run in processing. The overriding goal is to describe data content at a level sufficient to enable the potential users to discover data resources. Indexing terms or keywords frame searches, defining subjects, geographical areas, and chronological ranges. Keywords also help to assess data quality and methods of recording, as well as reflecting changes in intellectual and social context reflecting changes in theoretical frameworks. Recognizing that there is no single set of standards for data description, the AHDS developed and application of the Dublin Core (Weibel and Miller 1997) to enable searching of the AHDS catalog through use of the Internet. Discipline-related fields were created through application of the Warwick Framework (Dempsey and Weibel 1996). The core elements were extended through use of Scheme and Type qualifiers, which faciltiated description through existing standards (e.g., thesauri of object and monument types).

Data standards, in accepted or draft form, are evolving to better handle huge complex data sets and to effectively link archives around the world, however, Eiteljorg (1997) highlights the difficulties inherent in managing electronic archives (cf. Wise and Richards 2001). Collections of objects and hard copy records are often kept reasonably well long term but electronic data presents serious hurdles. The collections were created others, using idiosyncratic methods and forms, and the contents are disparate with files including text, databases, images, CAD files, and GIS files. Eitlejorg notes that storage is easy compared to making the electronic archive accessible over an extended period of time. Data migration is a constant, and if accessibility is to be maintained, electronic files must be routinely upgraded to current software and hardware standards. Data transformation is problematic, though technically straightforward, because of the complexity of archaeological data. Migrating an archaeological database, unlike for instance CAD files, requires intimate knowledge of archaeological method and theory. The data categories are seldom as obvious as creators think and datum points must be routinely specified. Documentation is crucial for effective use and migration strategies.

The social contexts of archaeological research, whether in practical realms of funding and legal agreements or in the shifting ground of accepted method and theory, are important considerations in the development of electronic archives. Eiteljorg’s (1997) article was written as he was managing the Archaeological Data Archive Project (ADAP) at Bryn Mawr University in the United States. From 1997-2003 ADAP was a developing model for electronic archiving. At the World Archaeology Congress, June 21-26, 2003, in Washington, D.C., Eiteljorg reported that ADAP was dead because archaeologists would not submit their data to the archive. ADAP was terminated because it was not economically viable, even though it had initial support through the Archaeological Institute of America, the American Anthropological Association, and the Society for American Archaeology. Eiteljorg framed his discussion by asking whether archival preservation of digital information was a reasonable expectation in the United States. He went on to speculate that a major issue for the demise of ADAP was the lack of computer sophistication and computer training for U.S. scholars. Another issue it would seem would be the lack of secure, ongoing funding directed by State and Federal mandate for preservation of electronic data. Archaeological resource management in the United States is a very different research context than that described for the United Kingdom.

Databases: Research and Communication

Databases are digital collections. These are the computer equivalent of the traditional object collections in museums and the records and provenience information stored in archives. Computer databases, if designed correctly, bring objects and documents and research records into a single, protected but accessible area where users can manipulate authenticated data (cf. Ferguson and Murray 1997; Flecker 2002).

Database management systems required the researcher to apply systematic categories and criteria in consistent ways to research a subject. Consistent categories allow transparent communication and discussion. Data models are used to guide the progression from data capture to data analysis to open dissemination. Creation of these data models require explicit refinement of traditional archaeological theory and method (e.g., Hadizlacos and Stoumbou 1995). Fields and measurements recorded are conditioned by the researchers’ interests as these are guided by the larger perspectives of the discipline. Bader (2000) declares firmly that any database application working with a two- or three-dimensional visualization tool like GIS enables transparent research and the promise for open communication. For Bader and other database designers and users, effective database management is a key to making dramatic strides in archaeological research and in communication and education (cf. Yang 1986).

Today’s computer-driven research is making significance strides but there is a major hurdle evident in the day-to-day work of archaeologists, revolving around the computer sophistication of archaeological producers and users (cf. Andresen and Madsen 1996; Campana and Crabtree 1987; Desse and Chaix 1986; Harland et al. 2003; Johnson 1997; Lohse 1996; Lohse and Sammons 1998; Powesland, Clemence and Lyall 1998; Rulf 1993; Ryan and van Leusen 2002; Winder 1994). Regional or national database collections are usually supported by IT teams but primary data is still produced principally by individual researchers with widely varying approaches, research questions, and expertise. Most of these archaeologists are not programmers nor data designers, they simply are using readily available, easy to use relational database management systems like dBase for Windows, Paradox, or Access. These products are usable in similar ways and data tables can be moved from one system to another through dBase files or ASCII delimited files. There are complications in data transformation and migration but most data producers and users operate above these concerns.

Data models are the singular requirement in developing any database. A researcher who wants to enter data must have designated a field. The fields are used to systematically capture the data. The data captured is directly related to the research question defined. Some fields are standard expectations, the kind of information all archaeologists are expected to gather (e.d., site location, description, collector, level, etc.). Other fields are unique to the methodology employed by the particular researcher pursuing specific research questions. The field chosen will channel and limit and all subsequent analysis. As Bader (1997) notes, it is commonplace still for researchers to neglect defining a data model or identifying appropriate fields before they start gathering data.

Data Transformation

Data analysis requires checking for errors, often using descriptive statistical analyses and visualization techniques. These can also be used to transform raw data into analytical data (cf. Fletcher and Lock 1991). For example, ordinal data can be transformed into interval data using normalization. Interval data can be placed in fewer intervals or to presence: absence categories. Factor analysis and cluster analysis are common approaches to data transformation. Different statistical techniques will also need transformations to produce different data characteristics. All transformations will insert varying levels of bias but cannot be avoided, since data must be prepared for subsequent analyses and for use of multivariate statistics like factor and cluster analysis, Bayesian classification or construction of neural networks.

Commercial database packages are available to run data transformations. DBASE is still a standard choice combined with various statistical packages like SPSS or WinState. Even database programming is becoming simpler as with the Delphi software package that is becoming increasingly popular for computer literate archaeologists.

Data Comparison

Researchers will inevitably want to run comparisons between their data and that constructed by others. Too often these comparisons generate debate without reference to the underlying data models or the use of variable analytical techniques. Use of shared databases enable archaeological discussions to move beyond opinion and into the realm of actual data comparison. The prerequisite is definition of the data model and effective capture of raw data.

Comparing two different relational database management systems will begin with comparison of their field structures and table structures. The comparable data categories can then be arranged in a common database model. Next, criteria lists are developed for each data category, transforming these into a common criteria list with the least raw data loss possible. Common data tables are then transformable into analytical data through solutions developed by the cooperating researchers. One result can be two different analytical data sets as well as a common shared data set. The discussion again settles on proper filtering and transforming of data. Data table comparison can be a tedious exercise but development of common platforms or analytical tools like the Bonn Seriation Package may accelerate easy data comparisons. Great potential is also presented in use of images in databases, which negates much of the tedious discussion centering on accepted descriptors or scales of measurement.

Knowledge Representation: Expert System/AI Databases

Development of optimal knowledge representation is a productive venue for maximizing data use (Adaptive Intelligent Systems 2003). Davis, Shrobe and Szolovits (1993) describe five very important and distinctive roles for knowledge representation. KR is a surrogate for the thing itself. It is a set of special commitments. It is a fragmentary theory of intelligent reasoning in three components: representational thinking; sets of inferences representing different sanctions; and sets of recommended inferences. It is a medium for pragmatic efficient computation, guiding how information is organized to maximize effective inference. It is an actual medium of human expression, constituting a language for describing the world.

These KR roles provide the framework for data characterizations. This framework drives realization that data capture must strive to represent maximum richness rather than reducing complexity to arbitrary, truncated categories lending only sparse, incomplete and inconsequential description. Richer data captures per force will feed stronger logical inferences. Enhanced KR frameworks will require elaborated KR technologies to be developed, including logical rules, frames, and semantic nets.

Representational thinking revolves around a fundamental observation: the intelligent observor who wishes to make sense of the world needs to acknowledge that reasoning is an internal process while the things reasoned about exist only externally. Neustupny (1992) identifies this juxtaposition in describing how archaeologists can define data only within a theoretical framework or knowledge base that then channels how information flows are constructed. Researchers don’t find data, they may not even recover data. They direct data capture based on categories established by their research interest conditioned by accepted theoretical and methodological positions accepted within their research field. To improve capture systems, we must accept the KR charge that we enrich our capture nets to better reflect the complexities of our study domains.

This dichotomy between the researcher’s mind and the world observed is the fundamental rationale and role for knowledge representation. Reasoning is seen as a surrogate for external actions. Questions revolve around the identity of the surrogate or what it is a substitute for and also the fidelity of the surrogate or what it includes and what it omits. Inevitably, we must acknowledge that the only completely accurate representation of any object is the object itself. Any representation short of this ideal is incomplete and must be explicitly defined and documented. By extension, we can argue that the more imperfect the surrogate, the more imperfect the inferences drawn. All surrogates must constitute distortions to some degree. Any description of the archaeological object is a reflection of our own contemporary thinking, with representations functioning only as surrogates for abstract notions of action, process, belief, causality, and categorization. Enriching data capture models insulates our research against an overriding truthful observation: the soundest reasoning will not avoid reaching wrong conclusions about our study domain. All representations, logically, are imperfect, and as such, lead to inevitable error. Good or best representations are those that minimize errors and good data models are those that emphasize as complete and representational a description as is theoretically and methodologically possible.

Knowledge representations are sets of ontological commitments. These reflect the imperfect nature of representations of reality but act to focus attention on selected features, characteristics or general descriptions that are seen to be useful by a particular discipline. Different task domains inevitably build different ontologies. These ontologies are written in different languages and reflect different theoretical and methodological foundations. The ontological commitments made begin at the level of the representation technologies and accrue at higher levels. Additional layers of commitment are added as the technology is applied. This constitutes the application of the constructed knowledge base. Layers and hierarchies are integral. The ontological questions are fundamental to the construction and application of this knowledge base. At each layer the choices made are about representations and not data structures. For example, a semantic net is a representation but a graph is a data structure. One simply implements the other. Every representation must be implemented in the computer by a data structure.

A useful perspective is presented in Minsky’s (1974) description of frame theory. The intelligent actor confronted with a new situation selects the frame from memory. A remembered frame is adapted to fit reality by the selective changing of details. The frame is simply a stereotyped situation that supplied context for resolvable, directed actions. Frames are used by archaeologists as elements of classification, whether types, stages, phases or narrative conventions. These frames are then defined as fields in computer analyses, and guide decision making on the categorization or chopping of data in the capture process.

Designers construct representations that offer a set of ideas about how to organize information in ways that facilitate the inferences. For instance, in frame theory, any frame will have attached to it several types of information. Frames are then used to organize information, mirroring the thinking of the researchers. Stone projectile point types, produced in the computer system as digital images, serve as frames with linked information. These same frames or types can then be constructed as taxonomic hierarchies with accompanying taxonomic reasoning and inference execution.

Obviously, representation and reasoning are inextricably intertwined. Building a comprehensive knowledge base will require use of accepted representations. However, the more creative aspect of database building and of data extraction involves designing systems that will feed innovative and intuitive use of the data. Advances will be made when new, more efficient representations are defined to reflect application of new technological, theoretical and methodological developments.

AI in Archaeology

Van den Dries (1998:13-17) offers one of the few summaries of knowledge-based applications developed for archaeology. He notes that archaeologists have been interested in developing expert systems since the early 1970s (cf. Doran 1974, 1977; Doran and Hodson 1975: 309-316; Lagrange and Vitali 1992). Expert systems have been accepted as amplifying knowledge transmission (e.g., Ennals and Brough 1982) and for instruction (van den Dries 1998). They have also been developed for improving theoretical and methodological research (Francfort 1990; Gardin et al. 1988; Lagrange and Renaud 1985;).

Expert systems in archaeology have been principally a means of formalizing and modeling knowledge molded by theories and methods. They are used to evaluate hypotheses, classify artifacts, predict site locations, standardize analytical frameworks, and run simulations on archaeological problem areas. Van Den Dries (1998) extends their use into teaching and knowledge transmission. It is important to note that experts may construct such systems to develop and apply expertise, but the non-expert may benefit in using them for consultation and communication.

Interpretive Context

Science, in itself, constructs mental maps. A goal of science is to explain the natural world, ranging from simple causality to elaborate chains of explanation. One powerful approach is reductionist, with its focus on internal structure: laws of motion, law of gravity, laws of aerodynamics. The result are general principles that interpret complex phenomena. Reductionism presents a clean cognitive map with meta-features and carefully defined layers of escalating or descending complexity. Another powerful approach is contextualist, where the view shifts from internal to external. For example, what external constraints molded the object under examination. Archaeological classification can be exceedingly reductionist. We draw up types based on morphology for instance, or correlate morphological types with chronological layers. This type of approach negates any explicit tie with prehistoric cognitive maps, and instead, is simply a handy means of heuristic classification. This is comparable to John Searle’s thought experiment of the “Chinese Room,” wherein a person seeking to understand Chinese manipulates huge stacks of paper according to rigid pre-prepared instructions. Questions in Chinese come in from the outside, pieces of paper get moved around, and an answer in Chinese will eventually go back out. Yet, we know the person cannot speak Chinese, so we have to assume that real intelligence cannot be reduced simply to a set of underlying rules. The experiment rests, of course, on a false analogy. A proper inference would be that the entire room, rules, paper-pusher, etc., is analogous to a person who understands Chinese. The person in the room would be analogous to one nerve cell in a Chinese speaker’s brain. According to Stewart and Cohen (1997: 180), since the room can carry on Chinese conversations flawlessly, the whole system must be held to “understand” Chinese (analogue for that system or understanding in the individual). A major flaw in the Searle mind experiment is that the fact that something is possible in principle, is much less informative than the fact that it is impossible in practice (Dennett 1991). For example, the instructions on paper would have to contain contingency plans for every possible Chinese question. They cannot simply be a huge catalog of questions and answers, no room would be large enough and no scheme could be far reaching enough. The rules for moving the paper would have to be devised by a fully fluent master of Chinese.

The “Chinese Room” metaphor warns us that logically developed classification systems, databases, and retrieval systems can establish types of objects and create solid statistical bases for discrimination but these do not inform us about their prehistoric makers. As analysts, we must attempt to place artifacts into their contexts of design, manufacture, use and re-use. The artifact forms should be viewed as constituting mental templates that have embedded meaning related to cultural mores, standards, economic decisions, and adaptive strategies.

Knowledge Elicitation

Creation of any knowledge-based application requires a trajectory of knowledge elicitation, analysis, modeling, and application design. The first and critical step is extraction of expert knowledge (cf. Burge 2003; Kidd 1987; Ford and Sterman 1997). The designer must initiate a full analysis of the problem domain and of the knowledge base required to develop an expert system that can perform a specific task successfully. Once the knowledge base has been constructed, a model can be made for subsequent design.

Designers have found that experts release an idealized view of their work (Payne and McArthur 1990). They implicitly emphasize ideal and clear situations and overlook problem scenarios. This lends a biased view, not fully representative of real world situations.

Van den Dries (1998:44-51) reviews difficulties encountered in eliciting expert knowledge for construction of WAVES-WARP, an expert system for lithic use-wear analysis. van den Dries reports that information was unbalanced and often insufficiently detailed, consisting predominantly of data on wear patterns that were considered diagnostic and under representing deviations from these ideal patterns. This tendency was particularly pronounced if working from published sources rather than in interviews with experts. There were also gaps in different expert’s knowledge and inevitable problems of cross-referencing different approaches and decision-making structures. Experts’ transmission of knowledge is also complicated by formal training (facts and theories) and their own subjected rules-of-thumb gained by experience.

The knowledge engineer (extractor) needs to construct systematic information in the format: “If you observe wear attributes A, B and C then this is an indication with certainty X that this tool was used in activity D or E, and not in activity F, because ....” As van den Dries (1998:45) concludes, lack of complete logical constructions from experts is due to practical factors: these are typically not required in reports of results and because use-wear types or categories are not clear-cut, frequently grading one into the other. This means that for most analysts, interpretation does not always occur within the absolute structure of formal rules but as the result of complex sets of associations.

Knowledge Handling

Once acquisition of knowledge has been accomplished, and summarized in a formal model, a conceptual map is made of what the application is going to look like, how tasks will be carried out, how knowledge will be represented, the nature of inference mechanisms, and what other aspects of transmission will be built in. Different types of programming can be used, including linear programming, rapid prototyping, and incremental programming. In linear programming, the application is implemented and evaluated as a complete product. A limitation is that users are not integral to the full development process and late discovery may reveal significant problems requiring fixes. Prototyping necessitates longer development but systematically and efficiently utilizes user input and patterned fixes (cf. Bratko 1989; Hayes-Roth et al. 1983; Kahn and Brauer 1989; Whipp and Lewis 1989). Both programming approaches are combined in incremental programming, wherein the system is divided into smaller parts, tasks or modules and each of these is implemented in the linear method and evaluated through prototyping. This has advantages: design can be quickly and efficiently updated, users are involved constantly, and the resulting prototype has been significantly debugged (cf. Hollnagel 1989).

Directionality, key to productive analysis, is keyed to problem solving within patterned cultural perceptions. Large complex problems will routinely be broken down into smaller less complex problems more amenable to resolution through use of practiced activity chains or programmed responses. Segal (1994:25-26) asserts that any analysis of problem solving has four components: identify the problem space as that range between the initial and goal states (Newell and Simon 1972); identify intermediate states between the initial and goal states (only trivial problems will allow direct movement from the initial state to the goal state); identify what needs to be done (movements as transformations); identify the resources (knowledge, skills, material, time) needed to execute each move. Directionality is achieved/defined by tracking moves from stage to stage.

Effective knowledge handling is measured as representational adequacy, correctness, modularity and simplicity. Assessment of the user interfaces attached involve measures of graphical possibilities, user friendliness and explanatory facilities. All design decisions will be measured for flexibility, transparency and ease of maintenance. In general, all knowledge handling systems should strive to be fully representative of the extant knowledge base, and they should allow easy maintenance of the constructed knowledge base. A modular approach is often the best solution, dividing knowledge into orderly, coherent chunks. The structure must be transparent with detailed data descriptions to ensure coherent use and ready upgradability. Overriding all design is that the system be made as simple as possible. That the system adhere to correctness is inescapable. It must exhibit all required reasoning facilities to initiate and complete routine tasks.

WAVES and WARP

Van Gijn and Fullgar (1998:203-207) welcome van den Dries’ work on WAVES in an addendum to his ground-breaking monograph (van den Dries 1998). They admonish that the primary thing missing in lithic use-wear analysis has been a definitive key for assessing stone tool function accepted by all analysts. In lieu of this key, analysts have developed different systems for recording and interpreting microscopic observations. Variable importance is then attached to different characteristics: scarring, striations, polish, beveling and rounding of surfaces. Analysts also enter myriad descriptive attributes, ranging from generally accepted standard terms to quite ideosynchratic constructions: melting snowfield and comet tails. Van den Dries offers expert systems as a means of formalizing and structuring the data capture and transmission process.

WAVES (Wear Analysing and Visualising Expert System) was developed to analyze use-wear traces on flint implements. WARP is an associated prototype based in neural networks. Van den Dries (1998) summarizes testing of both prototypes, assessing their relative qualities under comparable constraints. WAVES was tested twice, with evaluation by four analysts, with outcomes compared to other “blind tests.”

WAVES models expert knowledge development. The aim was two-fold: to successfully standardize the knowledge acquisition and application process and to refine the methodology of use-wear analysis. van den Dries (1998:43) reports that collecting and analyzing expert knowledge revealed that training of apprentices involved two levels: basic and advanced. Basic focuses on teaching methodological principles and guiding analysis. Advanced students reach autonomous interpretations and are more interested in verification. Hypothesis verification was seen to be a principle concern. van den Dries assumed that the WAVES application should be divided into two independent components: analysis and hypothesis validation. Of course, boundaries between the two stages are often fuzzy. In each phase of analysis the work from the previous phase is modified and refined and all ensuing steps condition decision-making on the part of the analyst.

WAVES is a rule-driven expert system, depending upon IF-THEN decision rules rather than on propositional logic, semantic nets or frames. van den Dries (1998:52) asserts that rule-driven decision making matches the conduct of lithic use-wear research best, and that it lent to simplicity and modularity of system design. Correctness was insured by confining WAVES to connecting a wear pattern to a specific material or motion is the observed features matched the motion exactly. Non-matching features, not seen previously, are excluded.

WAVES user interface was designed to include graphics and be highly interactive with attractive, intuitive screens. Built in, are facilities to guide information input and provide additional user support. Information on internal procedures was also made transparent. On request, a user is informed how and why certain facilities had to be incorporated, why a conclusion was drawn and why not. Rules active at a particular stage of the reasoning process are shown, and those activated next are indicated. Large collections of use-wear photos are included to support description of traces and to support interpretations resulting from the analytical procedures.

WAVES seems to make significant strides over Grace’s (1989) pioneering expert system in lithic use-wear analysis, FAST (Functional Analysis of Stone Tools). FAST like WAVES was designed to facilitate training students. The two are very different in basic design. FAST was developed for functional analysis of entire stone tools. WAVES was built to focus on interpretation of use-wear traces. van den Dries (1998:77) asserts that because WAVES does not focus on gross morphology, it is more applicable to analysis of a wider range of tool uses in different materials and across cultural periods. Further, FAST produces an interpretation as one final answer, where WAVES introduces all appropriate complexity to the student user, and includes an analytical and a hypothesis validating procedure.

Van den Dries (1998) included development of a neural network prototype called WARP. The neural network can handle more complex bits of data, involving unknowns. Neural networks are increasingly used in industrial, medical and other fields involving many practical and time-consuming redundant tasks like classification, pattern recognition and prediction. Neural nets effectively learn, by means of mathematical models that predict outcomes of complex situations by generalizing from similar, previous situations. The neural agent must be trained by entering lots and lots of information and running autonomous searchers for non-linear relationships between variables. It learns to select important variables and disregard others. Neural networks have a very different architecture from expert systems, use specific knowledge storing and processing methods and are applicable to different study domains (cf. Gibson 1992: 265; 1996). Van den Dries’ (1998) WARP was developed to investigate the possibilities and difficulties of formalizing and modeling expert knowledge involved in use-wear analyses by means of a neural network. The findings were compared with those generated from development of WAVES. It was found that WARP was easier to handle and to understand. Users found interfaces highly intuitive and the necessary user learning curves were much less than for the expert system WAVES. Van den Dries (1998:90-91) also concluded that WARP produced comparable analytical results with the added advantage of dealing with previously unclassified or unknown features. These could be used to guide future research. A profound difficulty in using the neural network centered on the neural agent having difficulty being trained in introduction of complex datasets. Too many contradictory facts led to many stumbling points in successful training. A lack of validation for answers and the autonomous character of the neural agent were also phrased as possible drawbacks.

SIGGI-AACS

SIGGI-AACS represents a significant step beyond WARP-WAVES in use of a neural network to develop both an automated classification system and an authoritative online database (Lohse et al. 2003). Past experiments with AI systems have largely been confined to rule-based forced classifications as in that developed by van den Dries (1998) for teaching use-wear analysts. Use of a sophisticated neural network that is trainable and capable of making novel intelligent decisions is an important approach to improving information sharing and exploring theoretical tenets of archaeological classification and data design.

SIGGI-AACS is a working prototype for online classification of stone projectile points in a neural network. The initial application uses specimens drawn from the North American Pacific Northwest cultural area but the system is extensible. The current database design is not software specific and was initially done in Microsoft Access. The autoclassification system consists of three interrelated products. Product 1 is the classification system, with software that allows users to submit images of artifacts or actual specimens to be digitized. This stage generates projectile point classifications with specimens assigned to recognized types and is a .NET standalone application. Product 2 consists of release of a typological descriptive report to system users, including a full image inventory of submitted and classified specimens with attached statistical probabilities of type assignment. Product 3 is a web-based educational venue for public access and study. SIGGI functions as a virtual analyst, which given some basic rules and concepts, is trained by introduction of new data sets. To improve its accuracy, SIGGI must be continually exposed to new and amplifying data fields. SIGGI is capable of accurately applying extant projectile point typologies; however, SIGGI can also identify outliers or unique data sets and suggest that these represent new types or that previous analysis identifying types needs modification within new explicit data ranges. As with any student, we must be certain that the data we ask SIGGI to analyze has been authenticated, and that we gather samples that are clearly representative of defined research populations. Because SIGGI learns by mimicking expert’s decisions, behaviors, and explicit rules, and then creates new decision frameworks integral to the compilation of new data, SIGGI eventually may generate insights into decisions made by human analysts and by prehistoric makers.

The principal barrier to training SIGGI is to retrieve collections that have fine excavation and analytical context. A primary assumption in archaeological typologies is that the knappers of the stone points were operating within a very well defined cultural model that laid out clear expectations regarding what a particular projectile point form should look like. This working assumption is borne out in the clear temporal and spatial separation of populations of projectile points (e.g., Lohse 1985). Essential for training this virtual analyst are retrieval of sample populations that as nearly as possible represent these real time actors in the past. For example, for training SIGGI needs projectile point samples that were found in large numbers on a single site, within a specific layer, in association with cultural features representing clear prehistoric human activity, and bracketed by good reliable radiocarbon dates. These samples supply the virtual analyst with numerous points made to a prehistoric standard, and reveal expected ranges of statistical variation in basic variables of form. This allows SIGGI to make intelligent decisions on where to draw lines demarcating the distinctive types of projectile points. SIGGI’s ability to explicitly handle multiple variables in a multidimensional statistical environment promises insights into clarification and refinement of chronologies of prehistoric projectile point types, a result of considerable interest to the practicing archaeologist (cf. Lohse et al. 2003).

The current SIGGI-AACS Project has focused on obtaining authenticated data in an effort to produce a “clean” set of data that reproduces exactly the classification published by Lohse (1985). Data collected are stored in an image database with attached descriptive fields. SIGGI is in a sense, Lohse’s virtual brain. Since SIGGI represents the thinking of only one archaeologist, future work intends to extract knowledge from other archaeological typologists. SIGGI can “think” like Lohse, who summarized previous typological work, but the project is expanding and will have SIGGI interact with other researchers’ ideas, i.e. bring SIGGI an education from the larger community. This reflective activity is one of the more important aspects of the project. Obviously, certain kinds of things can be classified in proscribed ways, but project research focuses on identifying WHY things should be classified in certain ways. By watching SIGGI make classifications, these researchers hope to gain a better understanding of why archaeologists make classifications and how these might be continually improved as research methodology improves.

The accuracy of classification provided by SIGGI depends on a large authenticated knowledge base. The SIGGI database performs several functions that influence both the performance of SIGGI and the sharing of data among researchers. At a minimum, SIGGI-AACS is required to (1) store images; (2) store locations, (3) store characteristics, and (4) keep information secure.

SIGGI also allows for improved research collaboration. Using the underlying SIGGI-AACS research database, many researchers can maintain information about many collections on-line. Each researcher is able to maintain complete ownership of access to his own material. This allows research bridges to be built while maintaining the unique identity of each research collection. When fully implemented, the system can maintain collections of many types and provide neural network analytic services where appropriate.

A crucial aspect of the SIGGI-AACS project is how users will access and input information on the Internet. Four major user groups have been identified: government agencies, researchers, Native Americans, and the general public. Tasks have also been identified that will be requested by each of these user groups.

A primary user of SIGGI-AACS should be federal and state land-management agencies (e.g., USDA Forest Service, USDI Bureau of Land Management, National Park Service). Agency users will be interested in the automated typology made possible by SIGGI’s neural network. Three major tasks include analysis of point images submitted, data storage, and summative transfers of information.

For Task 1, the web site contains appropriate interfaces to allow the agency (or a SIGGI-AACS staff member) to enter an image of the projectile point for automatic type assignment. This site allows the user to progress through the different stages of image acquisition, manipulation, and analysis. At each point in this image manipulation process, the user must be able to check the image and authorize its movement to the next stage. Once the image has been successfully entered, SIGGI will analyze it and return the appropriate typological assignment. At the same time (and out of the user’s view), SIGGI will also encode the relevant datapoints (size and shape indicators), and file the datapoints, image, and typological, locational, and temporal information into the SIGGI-AACS database. In Task 2, the agency user is able to access information which is already in the database and for which that agency has statutory responsibility. In this case, the web site provides an interface by which the government agency can request images, specimen or inventory numbers, counts, chronologies, charts, graphs, or maps displaying information about specific projectile points. In Task 3, the web site provides interfaces for the agency to request similar, but summative, information for projectile points not within their statutory responsibility.

Researchers are interested in the automatic typology made possible by SIGGI’s neural network, but also need to access information from the database for specific projectile points, including images but perhaps excluding locational data. For Tasks 1-3, the web interface is the same as that set for government agencies. For Task 4, the interface must also allow the user to request information about specific projectile point(s). For example, a researcher may request images of points from the type site or of a recent find. Providing this type of specific information to archaeologists advances the resolution of study, adding chronology, spatial distribution, functional studies, or symbolic studies. However, locational information would be excepted: academic or contract archaeologists cannot be given exact locations for artifacts without permission from the federal, state, or tribal landowner.

The third profiled user group are Native American tribes and First Nations in the United States and Canada. SIGGI-AACS, with its chronological and spatial data, may have profound implications regarding cultural affinity issues, and tribes may be interested in this information for both heritage and legal reasons. Tribes may be able to use the data stored in SIGGI-AACS to argue for traditional use of land or rivers not currently within the tribes’ legally recognized authority. In addition, tribes and first nations may wish to use the projectile point database as a mechanism for storing cultural heritage information and to provide that information in educational contexts (tribal museums, schools, etc). Security will be a major issue, and United States federal law requires that locations of archaeological sites not be made available to the general public. A fifth tasks is added for these users: the imposition of secure access controls.

The fourth major user group for SIGG-AACS are members of the public. These users will ask such questions as, “I found a point. What type is it?” “Where are these points found?” and “How old is it?” They may also be interested in images of projectiles in the database, perhaps to compare with a point in their own possession or to copy into a school report. The materials provided to these users will be the result of scripted actions and not the result of active database searchers.

Implications for Future Developments

WAVES-WARP and SIGGI-AACS projects, while still in the prototype phase, provide innumerable examples illustrating fundamentals of database design, user interface design, and relational database design. Both operate on multiple levels, from development of an explicit statistically based online classification system with attached database, to use of an electronic or virtual agent to augment archaeological training in classification, to observation of an artificial agent to study the character and effectiveness of archaeological thinking. Anthropologists and archaeologists are beginning to join cognitive psychologists and learning theorists in the use of artificial intelligence systems to explore human thought and behavior (e.g., Baylor 2002; Conte and Castelfranchi 1995; Cumming 1998; Doran various; Epstein and Axtell 1996; Gonzalez and DesJardins 2002; Russell and Norvig 1995; Woolridge and Jennings 1998; Woolridge, Muller and Tambe 1996).

Although others have used neural networks in archaeology, WAVES-WARP and SIGGI-AACS are partially successful prototypes in archaeological information technology. Key now is to expand on these prototypes and authenticate their potential (cf. Stead 2003). Obvious productive spinoffs from this research include: (1) training of an online neural classification system capable of accurately identifying archaeological artifacts (SIGGI in this sense constitutes a highly interactive user interface sitting atop a secure database); (2) creation of new theoretical and methodological frameworks to accelerate effective information design; (WAVES and SIGGI offer advantages in teaching and insights into how we conceive of our study domains); (3) further development of artificial intelligence systems linked to giant heritage databases that are constantly maintained and revised to ensure secure storage, organization and transfer of our archaeological heritage.

Construction of large databases supervised by intelligent agents is a completely attainable, realistic projection not just for archaeology but for all information heavy, data rich disciplines (e.g., Egenhofer 2002; Farenc et al. 2000; Thalmann, Farenc and Boulic 1999; The Semantic Web 2003; Tsukii, Kihara and Ugawa 2001). This is a major break from past practice in archaeology: where laborious searches in libraries and archives for hard-to-find publications and “gray literature” was the norm; where tedious and time-consuming requests are made to overworked archive and collections managers to hand-relate various hard copy finder’s guides in order to find and pull specimens from cabinet drawers and storage boxes (cf. Huggett 1995; Lock 1995; Lock and Brown 2000; Madsen 2001; Pinto et al. 2001; Stewart 1996). The vision that information can be accessed through a central portal and seamlessly indexed and sorted dependent upon researcher interest and creative motivation constitutes a paradigm shift in archaeological information management very like that declared for e-publishing by Morton (1997), fueled by a shared ideology steeped in IT (cf. Ginsparg 1996; Odlyzko 1996). An affirmation of technology has taken place and is driving significant changes in the infrastructure of scientific research (cf. Dreyfus 2001; Dreyfus 2002; Dreyfus and Dreyfus 2002; Hodder 1999; Neustupny 1992). Use of the Internet for delivery of scientific information not only speeds access but forces changes in the social organization of scholarship and the authentication of information (cf. Fulda 2000; Gray and Walford 1999; Holmen et al. 2003; Lamprell et al. 1995; Miller, Dawson and Perkins 2002; van Leusen et al. 1996). New kinds of interfaces will be developed that sit atop these huge heritage databases, ensuring that users have virtually seamless interaction with data, whether through virtual documents or virtual agents (Gruber, Vemuri and Rice 2003; Miller 2000; Ryan 1995). GIS is still the backbone for many archaeological data applications, and innovative uses of maps to transmit information is yet another form of enhanced interface produced in Ian Johnson’s impressive TimeMap Project (e.g., Johnson 1999, 2002a, 2002b, 2003 ( see also Lancaster and Bodenhammer 2002 on the Electronic Cultural Atlas Initiative). Virtual Reality applications also have tremendous potential in transferring information as 3-D data (e.g., Pringle and Moulding 1997; Razdan et al. 2001, 2002).

We are not at our vision yet. Technology, in the form of hardware and software is available, and we can do things today in manipulating huge data sets, that were unthinkable a decade ago. Computers are increasing in power and our students have an underpinning in IT that readies them to join our vision for the future. The principle problem lies in social informatics, in understanding how our discipline has assembled data and conveyed information. Standards are in place that gauge scholarly contribution and data integrity, and some of these will have to fall away to accept the shared vision of IT applications across the board.

References Cited

Adaptive Intelligent Systems

2003 Adaptive Intelligent Systems in the Knowledge Systems Laboratory at Stanford.

AHDS

2003 AHDS Case Studies.

All Party Parliamentary Archaeology Group

2003 Report on the State of British Archaeology - Recommendations. Response from the Council for British Archaeology, September 2003.

All Party Parliamentary Archaeology Group

2003 APPAG Submissions. Revised Summary of Issues from Submissions.

American Chemical Society

1997 Journal of the American Chemical Society: Policy Statement.

American Psychological Association

1996 Posting Information on the Internet.

Australian Museums and Galleries Online

2003 AMOL: Australian Museums & Galleries Online.

Andresen, J. and T. Madsen

1996 IDEA - the integrated database for excavation analysis. In Kamermans, H. And K. Fennema (eds.), Interfacing the Past. Computer Applications and Quantitative Methods in Archaeology CAA95, Vol. 1, pp.2-14. Leiden: Institute of Prehistory.

Archaeology Data Service

2003 Archaeology Date Service: homepage.

Archaeology Data Service

2003 Digital Archiving Pilot Project: Excavation Records.

Association for Computing Machinery

1995 ACM Interim Copyright Policy. Version 2, November 15, 1995.

Australian National University

1997 Archaeology World. .

Bader, Hans-Dieter

1997 Databases - research tools and communication aids.

Baker, D., G. Chitty, J. Richards and D. Robinson

2000 Mapping information resources: a report for HEIRNET. Council for British Archaeology.

Baylor, A.

2002 Agent-based learning environments as a research tool for investigating teaching and learning. Journal for Educational Computing Research 26(3): 227-248.

Beargrie, N. and D. Greenstein

1998 A Strategic Policy Framework for Creating and Preserving Digital Collections. In Guidelines for Digital preservation. Arts and Humanities Data Service.

Blackaby, J.R., P. Greeno and the Nomenclature Committee

1995 The Revised Nomenclature for Museum Cataloging. A Revised and Expanded Version of Robert G. Chenhall’s System for Classifying Man-Made Objects. American Association for State and Local History Series. Walnut Creek: AltaMira Press.

Booth, B.K.W.

1982 Some projects to store and process data from survey and excavation with the aid of a computer. In Computer Applications in Archaeology 1981, pp. 20-25.

Booth, B.K.W.

1995 Has archaeology remained aloof from the information age? In J. Huggett and N. Ryan (eds.), Computer Applications and Quantitative Methods in Archaeology 1994, pp.1-12. BAR International Series 600.

Booth, B.K.W.

1995 Strategic planning issues: Preface. In J. Wilcock and K. Lockyear (eds.), Computer Applications and Quantitative Methods in Archaeology 1993, pp.1-12. BAR International Series 600.

Brent, D.

1995 Steven Harnad’s ‘Subversive Proposal’: Kick-starting Electronic Scholarship - A Summary and Analysis. The Information Society 11(4): 261-273.

Bratko, I.

1989 Fast prototyping of expert systems using Prolog. In G. Guida and C. Tasso (eds.), Topics in Expert System Design: Methodologies and Tools, pp. 69-86. Amsterdam.

Brown, A. And K. Perrin

2000 A Model for the Description of Archaeological Archives. Information Management & Collections, English Heritage, Centre for Archaeology. .

Burge, J.E.

2003 Knowledge Elicitation Tool Classification. Artificial Intelligence Research Group, Worcester Polytechnic Institute.

Burnard, L. And H. Short

1994 An Arts and Humanities Data Service, report of a Feasibility Study commissioned by the Information Systems Sub-Committee fo the Joint Information Systems Committee of the Higher Education Funding Councils.

Bustard, W.

2000 Archaeological curation in the 21st century or, making sure the roof doesn’t blow off. CRM 23(5): 10-15.

Campana, D.V. and P.J. Crabtree

1987 ANIMALS - A C language computer program for the analysis of faunal remains and its use in the study of early Iron Age fauna from Dun Ailinne. Archaeozoologia 1: 57-68.

Canadian Heritage Information Network

1997 Canadian Heritage Information Network.

Carnett, C.L.

1995 A Survey of State Statutes Protecting Archaeological Resources. Archaeological Assistance Study No.3, Archaeological Assistance Division. Washington, D.C.: National Park Service.

Carnett, C.L.

1991 Legal Background of Archaeological Resources Protection. Technical Brief No. 11, Archaeological Assistance Division. Washington, D.C.: National Park Service.

Champion, S.

1997 Archaeology on the World Wide Web: A user’s field-guide.

Cheek, A.

1991 Protection of Archaeological Resources on Public Lands: History of the Archaeological Resources Protection Act. In Protecting the Past. Boca Raton: CRC Press.

Cheetham, P.N. and J.G.B. Haigh

1992 The archaeological database - new relations? In G. Lock and J. Moffett (eds.), Computer Applications and Quantitative Methods in Archaeology, pp.7-14. BAR International Series S 577.

Childs, S.T.

2001 SAA committee on curation: who we are, our goals, and our issues. The SAA Archaeological Record 1(2): 10-11, 37.

Childs, S.T.

1999 Symposium: Managing archaeological heritage through collections, documentation, and datbases. Archaeological Collections: Managing an Emerging Frontier. World Archaeological Congress 4, University of Cape Town, 10th-14th January 1999. Archaeology and Ethnography Program, National Park Service.

Childs, S.T.

1995 The curation crisis. Federal Archaeology 7(4): 11-15.

CIDOC

1995 Draft International Core Data Standard for Archaeological Sites and Monuments. Paris: International Council of Museums.

CIDOC

2003 CIDOC Conceptual Reference Model. Version 3.2.1. Committee Draft ISO/CD21127.

CIDOC

2003 “Sharing the Knowledge” International CIDOC CRM Symposium, March 26-27, 2003, Washington, D.C.

CIMI

2003 CIMI Consortium. Museum Intelligence.

Clubb, N.D. and W.D. Startin

1995 Information management systems strategies in national organisations and the identification, legal protection and management of the most important sites in England. In J. Wilcock and K. Lockyear (eds.), Computer Applications and Quantitative Methods in Archaeology 1993, pp.67-73. BAR International Series 598.

Clubb, N.D. and N.R. Lang

1996 A strategic appraisal of information systems for archaeology and architecture in England - past, present and future. In H. Kamermans and K. Fennema (eds.), Interfacing the past. Computer Applications and Quantitative Methods in Archaeology, pp.51-71. CAA95.

Code of Federal Regulations

1990 Code of Federal Regulations Title 36, chapter I, part 79: Curation of federally owned and administered archaeological collections. Federal Register 55(177).

Comte, R. and C. Castelfranchi

1995 Understanding the functioning of norms in social groups through simulation. In N. Gilbert and R. Conte (eds.), Artificial Societies, pp. 252-267. London: UCL Press.

Content Standard for Digital Geospatial Metadata

2003 Content Standards for Digital Geospatial Metadata (CSDGM).

Council of Europe

1995 Core data index to historic buildings and monuments of the architectural heritage. Council of Europe.

Crescioli, M., D’Andrea, A. and F. Niccolucci

2002 XML Encoding of Archaeological Unstructured Data. In G. Burenholt (ed.), Archaeological Informatics: Pushing the Envelope CAA 2001. Computer Applications and Quantitative Methods in Archaeology, Proceedings of the 29th Conference, Gotland, April 2001, pp. 267-275. BAR International Series 1016, Oxford.

Crofts, N., I. Dionissidou, M. Doert and M. Stiff (eds.)

2003 CIDOC Conceptual Reference Model. Version 3.2.1. Committee Draft ISO/CD21127.

Cultural Heritage Consortium

2002 Users and their uses of HEIRs’, HEIRNET.

Cumming, G.

1998 Artificial Intelligence in Education: An Exploration. Journal of Computer Assisted Learning14(4): 251-259.

Davis, R., H. Shrobe and P. Szolovits

1993 What is a Knowledge Representation? AI Magazine 14(1): 17-33.

Dempsey, L.

1996 ROADS to Desire: Some UK and Other European Metadata and Resource Discovery Projects. D-Lib Magazine July/August.

Dempsey, L. And S.L. Weibel

1996 The Warwick metadata Workshop: a framework for the deployment of resource description. D-Lib Magazine July/August.

Dennett, Daniel C.

1991 Consciousness explained. Boston: Little, Brown & Co.

Desse, J. and L. Chaix

1986 OSTEO: AN operational osteometric database. In L.H. van Wijngaarden-Bakker (ed.), Database Management and Zooarchaeology, pp. 93-98. PACT 14. Strasbourg.

DiMaggio, P.J. and W.W. Powell

1991 Introduction. In W.W. Powell and P.J. DiMaggio (eds.), The New Institutionalism in Organizational Analysis, pp. 1-38. Chicago: University of Chicago press.

Doran, J.E.

1974 What computers can’t do for archaeologists. In J.D. Wilcock and S. Laflin (eds.), Computer Applications in Archaeology, Conference Proceedings 1974, pp. 69-70. Birmingham.

Doran, J.E.

1977 Automatic Generation and Evaluation of Exploratory Hythpotheses. In M. Borillo, W. Frenandez dela Vega and A. Guenoche (eds.), Raisonnement et methodes mathematiques en archeologie, pp. 172-181. Paris.

Doran, J.E.

1997 Artificial Societies and Cognitive Archaeology. Archeologia e Calcolatori 7: 1231-1245.

Doran, J.E.

1997 Distributed Intelligence and Emergent Social Complexity. InS.E. van de Leeuw and J. McGlade (eds.), Time, Process and Structured Transformation in Archaeology, pp. 283-297. London: Routledge.

Doran, J.E.

1998 Simulating Collective Misbelief. Journal of Artificial Societies and Social Simulation 1(1).

Doran, J.E.

2000 Hard Problems in the Use of Agent-based Modelling. In Proceedings of the Fifth International Conference on Social Science Methodology (ISA-RC33): Social Science Methodology in the New Millenium, Cologne, October 3-6, 2000. CD-ROM.

Doran, J.E.

2000 Prospects for Agent-based Modelling in Archaeology. Archeologia e Calcolatori 10: 33-44.

Doran, J.E.

2000 The Archaeology of Artificial Societies. CPM Report No. 00-61, 2 May 2000, “Starting from Society” Symposium, ASIB ‘2000 convention, 4/16-19, Birmingham University.

Doran, J.E. and F.R. Hodson

1975 Mathematics and Computers in Archaeology. Edinburgh.

Doran, J.E. and M. Palmer

1995 The EOS Project: integrating two models of Paleolithic social change. In N. Gilbert and R. Conte (eds.), Artificial Societies: The Computer Simulation of Social Life, pp. 103-125. London: UCL Press.

Dreyfus, H.L.

2002 Highway Bridges and Feasts: Heidegger and Borgman on How to Affirm Technology. Version of a presentation delivered at the 1996 Bugbee Lecture, University of Montana.

Dreyfus, H.L.

2001 On the Internet. London: Routledge.

Dreyfus, H.L. and S.E. Dreyfus

2002 From Socrates to Expert Systems: The Limits and Dangers of Calculative Rationality.

Dublin Core

2002 Dublin Core Metadata Initiative.

Egenhofer, M.J.

2002 Toward the Semantic Geospatial Web. Proceedings of the Tenth ACM International Symposium on Advances in Geographic Information Systems, pp. 1-4. New York: ACM Press.

Eitlejorg, H.

1997 Archaeological Data Archive project. .

Eiteljorg, H.

1997 CSA/ADAP Archaeological Database Discussions.

Eiteljorg, H.

2001 Archiving archaeological data.

Eiteljorg, H.

2000 If we preserve the files, who will use them? .

Ennals, R. And D. Brough

1982 Representing the knowledge of the expert archaeologist. In S. Laflin (ed.), Computer Applications in Archaeology, Conference Proceedings 1982, pp. 56-62. Birmingham.

Epstein, J.M. and R.L. Axtell

1996 Growing Artificial Societies: Social Science from the Bottom Up. Washington, D.C.: The Brookings Institute.

Association for Computing Machinery

2003 Association for Computing Machinery. The First Society in Computing.

ETAI

1997 Call for Papers for the First Issues of the ETAIL - Electronic Transactions on Artificial Intelligence.

Farenc, N., F. Sidler, A. Ferrando and D. Thalmann

2000 Database and Modelling Strategy: A Compliant Way for Display optimization. Paper presented at the First International Conference on Web Information Systems Engineering (WISE ‘00), June 19-20, Hong Kong.

Federal Geographic Data Committee

1997 Content standards for digital geospatial metadata. April 1997 draft. .

Ferguson, L.M. and D.M. Murray

1997 Archaeological Documentary Archives: Preparation, Curation and Storage. Institute of Field Archaeology Papers 1.

Fernie, K.

2003 Getting it together on-line: HEIRNET and Internet-based resource discovery tools for the Historic Environment. Internet Archaeology 13.

Finholt, T. and J.M. Brooks

1997 Analysis of JSTOR: The impact on scholarly practice of access to on-line journal archives. Paper presented at the Scholarly Communication and Technology Conference, Emory University, April 24-25.

Fink. E.E.

1999 The Getty Information Institute. D-Lib Magazine 5(3).

Flecker, D.

2002 Preserving Digital Periodicals. In National Digital Information Infrastructure and Preservation Program, Building a National Strategy for Digital preservation: Issues in Digital Media Archiving. Washington, D.C.: Council on Library and Information Resources and Library of Congress.

Fletcher, M. and G.R. Lock

1991 Digging Numbers. Oxford University Committee for Archaeology Monograph 33.

Ford, D.N. and J.D. Sterman

1997 Expert Knowledge Elicitation to Improve Mental and Formal Models.

Francfort, H.-P.

1990 Modelling Interpretativ Reasoning in Archaeology with the aid of Expert Systems: Consequences of a Critique of the foundations of Inferences. In R. Ennals and J.-C. Gardin (eds.), Interpretation in the Humanities: Perspectives from Artificial Intelligence, pp. 101-129. LIR Report 71. The British Library.

Fulda, J.S.

2000 The Internet as an Engien of Scholarship. Computers and Society, March 2000, pp. 17-27.

Gardin, J.-C., O. Quillaume, P. Herman, A. Hesnard, M.-S. Lagrange, M. Renaud and E. Zadora-Rio

1988 Artificial Intelligence and Expert Systems: Case Studies in the Knowledge Domain of Archaeology. Chichester.

GCMD

1996 Directory Interchange Format (DIF) Writer’s Guide. Version 5.

Getty Research Institute

2003 Getty Research Institute. Research at the Getty.

The Genome Database

2003 The Genome Database. An international collaboration in support of the Human Genome Project.

Gibson, P.M.

1992 The potentials of hybrid neural network models for archaeofaunal ageing and interpretation. In J. Andresen, T. Madsen and I. Scollar (eds.), Computing the Past: Computer Applications and Quantitative Methods in Archaeology 1992, pp. 263-271. Aarhus.

Gibson, P.M.

1996 An archaeofaunal ageing comparative study into the performance of human analysis versus hybrid neural network analysis. In H. Kamermans and K. Fennema (eds.), Interfacing the Past: Computer Simulations and Quantitative Methods in Archaeology 1995, pp. 229-233. Leiden.

Ginsparg, P.

1996 Winners and Losers in the Global Research Village. Paper presented at the Joint ICSU Press/UNESCO Expert Conference on Electronic Publishing in Science.

Gonzalez, J. And S. Desjardins

2002 Artificial Neural Networks: A New Approach to Predicting Application Behavior. Research in Higher Education43(2): 235-258.

Grace, R.

1989 Interpreting the Function of Stone Tools. The Quantification and Computerisation of Microwear Analysis. BAR International Series 474. Oxford.

Gray, J. And K. Walford

1999 One good site deserves another: electronic publishing in field archaeology. Internet Archaeology 7.

Greenstein, D. and J. Trant

1996 The Arts and Humanities Data Service: Collecting digital research data: building a framework for digital resource preservation and interchange. Ariadne 4. .

Gruber, T., S. Vemuri and J. Rice

2003 Virtual documents that explain How Things Work: Dynamically generated question-answering documents. Knowledge Systems Laboratory, Stanford University.

Hadizlacos, T. And P.M. Stoumbou

1995 Conceptual Data Modelling for Prehistoric Excavation Documentation. In H. Kamermans and K. Fennema (eds.), Interfacnig the Past. Computer Applications and Quantitative Methods in Archaeology CAA 95, pp. 21-30. Analecta Praehistorica Leidensia 1. Leiden: Institute of Prehistory.

Hansen, H.J.

1993 European archaeological databases: problems and prospects. In J. Andresen, T. Madsen and I. Scollar (eds.), Computing the past. Computer Applications and Quantitative Methods in Archaeology CAA92, pp.229-237.

Harland, J.F., J.H. Barrett, J. Carrott, K.Dobney and D. Jaques

2003 The York System: AN integrated zooarchaeological database for research and teaching. Internet Archaeology 13.

Harnad, S.

1991 Post-Gutenberg Galaxy: The Fourth Revolution in the Means of Production of Knowledge. Public Access Computer Systems Review 2(1): 39-53.

Hayes-Roth, F., D.A. Waterman and D.B. Lenat (eds.)

1983 Building Expert Systems. Reading.

HEIRNET

2002 HEIRNET. Historic Environment Information Resources Network.

Hodder, I.

1999 Archaeology and global information systems. Internet Archaeology 6.

Holmen, J., C.-E. Ore, O. Eide and E. Jordal

2003 Documenting Two Histories At Once: Doing Archaeology On Archaeological Texts. Paper presented at the Fifth World Archaeological Congress, “Archaeology in the Digital Age: Digital Archives and Access,” July - , Catholic University, Washington, D.C.

Hollnagel, E.

1989 Evaluation of expert systems. In G. Guida and C. Tasso (eds.), Topics in Expert System Design: Methodologies and Tools, pp. 377-416. Amsterdam.

Huggett, J.

1995 Democracy, data and archaeological knowledge. In J. Huggett and N. Ryan (eds.), Computer Applications and Quantitative Methods in Archaeology 1994, pp.23-26.

Hunter, J. And Ralston, I.

1997 Archaeological Resource Management in the UK: An Introduction. Sutton Publishing Ltd.

Iacono, S. And R. Kling

1996 Computerization Movements and Tales of Technological Utopianism. In R. Kling (ed.), Computerization and Controversy: Value Conflicts and Social Choices, pp. 85-105. 2nd edition. San Diego: Academic Press.

ICCD

2003 Instituto Centrale per il Catalago e la Documentatzione.

ICOM-CIDOC

2001 Developments in museum and cultural heritage information standards. A joint project of the Getty Information Institute (formerly the Getty Art History Information Program) and the International Committee for the Documentation of the International Council of Museums. Http;//wilpowerinfo.myby.co.uk/cidoc/standI.html

ICOM-CDOC

1995 ICOM/CDIC Data Model Working Group Graphic Data Model. International Documentation Committee of the International Council of Museums. Http:

ICOM-CIDOC

2002 Museum documentation initiatives: projects.

ICOM-CIDOC

2003 The International Committee for Documentation of the International Council of Museums (ICOM-CIDOC).

ICOMOS Documentation Dentre

2003 Centre de Documentation UNESCO. Bibliographic Database.

Internet Archaeology

2003 Internet Archaeology.

Johnson, I.

2003 Integrating Databases with Maps: The Delivery of Cultural Data Through TimeMap. Papers. Museums and the Web 2003.

Johnson, I.

2002 Contextualising Historical Information Through Time-Enabled Maps. Internet Archaeology 12.

Johnson, I.

2002 IDEA. Integrated Database for Excavation Analysis. Version 1.01.

Johnson, I.

2002 From metadata to animation: Web-based searching and mapping of cultural heritage information. Proceedings of UNESCO World Heritage Centre 30th Anniversary Virtual Congress.

Johnson, I.

1999 Mapping the fourth dimension: the TimeMap project. In L. Dingwall, S. Exon, V. Gaffney, S. Laflin and M. Van Leusen (eds.), Archaeology in the Age of the Internet, 21pp. British Archaeological Reports International Series 750. CD-ROM.

Johnson, I.

1997 The Essential IDEA. AN Introduction to the Intergated Database for Excavation Analysis. Draft 1/1/97. Archaeological Computing Laboratory, Sydney.

Kahn, G.S. and M. Bauer

1989 Prototyping: Tools and Motivations. In G. Guida and C. Tasso (eds.), Topics in Expert System Design: Methodologies and Tools, pp. 47-68. Amsterdam.

Kidd, A.L. (Ed.)

1987 Knowledge Acquisition for Expert Systems: A Practical Handbook. New York.

King, T.

2000 Federal Planning and Historic Places: The Section 106 Process. New York: AltaMira Press.

Kling, R.

1999 What is Social Informatics and Why Does it Matter? D-Lib Magazine, January 1999.

Kling, R. And L. Covi

1995 Electronic Journals and Legitimate Media in the Systems of Scholarly Communication. The Information Society 11(4): 261-271.

Kling, R., H. Crawford, H. Rosenbaum, S. Sawyer and S. Weisband

1999 Information Technologies in Human Contexts: Learning from Organizational and Social Informatics. Bloomington: Center for Social Informatics.

Kling, R. and S. Iacono

1995 Computerization Movements and the Mobilization of Support for Computerization. In S.L. Star (ed.), Ecologies of Knowledge: Work and Politics in Science and Technology, pp. 119-153. Albany: State University of New York Press.

Kling, R. and G. McKim

2000 Not Just a Matter of Time: Field Differences in the Shaping of Electronic Media in Supporting Scientific Communication. Journal of the American Society for Information Science 51(14): 1306-1320.

Kling, R. and G. McKim

1998 The shaping of electronic media in supporting scientific communication: The contribution of social informatics.

Lagoze, C., C.A. Lynch and R. Daniel

1996 The Warwick Framework: a container architecture for aggregating sets of metadata. .

Lagrange, M. and M. Renaud

1985 Intelligent knowledge-based systems in archaeology: A computerised simulation of reasoning by means of an expert system. Computers and the Humanities 19: 37-49.

Lagrange, M. and V. Vitali

1992 VANDAL: an expert system dealing with the provenance of archaeological ceramics, based on chemical, mineralogical and data analysis information. In P. Reilly and S. Rahtz (eds.), Archaeology and the Information Age, pp. 276-287. London.

Lahanier, C.

2003 Network of Art Research Computer Image Systems in Europe. Laboratoire de recherche des musees de France.

Lamprell, A., A. Salisbury, A. Chambers and D. Stoddart

1995 An Integrated Information System for Archaeology Evidence. In H. Kamermans and K. Fennema (eds.), Interfacnig the Past. Computer Applications and Quantitative Methods in Archaeology CAA 95, pp. 37-42. Analecta Praehistorica Leidensia 1. Leiden: Institute of Prehistory

Lancaster, L.R. and D.J. Bodenhammer

2002 The Electronic Cultural Atlas Initiative and the North American Religion Atlas. In A.K. Knowles (ed.), Past Time, Past Place: GIS for History, pp. 163-177. Redlands: ESRI Press.

Letovsky, S.

1998 Termination of GDB Project.

Letovsky, S.

1999 Bioinformatics: Databases & Systems.

Library of Congress

1997 MARC Standards.

Lock, G.

1995 Archaeological computing, archaeological theory, and moves towards contextualism. In J. Huggett and N. Ryan (eds.), Computer Applications and Quantitative Methods in Archaeology 1994, pp.13-18.

Lock, G. And K. Brown (eds.)

2000 On the Theory and Practice of Archaeological Computing. Oxford: Oxbow Books.

Lohse, E.S.

1985 Rufus Woods Lake Projectile Point Chronology. In S. Campbell (ed.), Summary of Results: Chief Joseph Dam Cultural Resources Project, Washington, pp. 317-364. Report to the U.S. Army Corps of Engineers, Office of Public Archaeology, University of Washington, Seattle.

Lohse, E.S.

1996 A computerized descriptive system for functional analysis of stone tools. Tebiwa 26(1): 3-66.

Lohse, E.S. and D. Sammons

1998 A computerised data base for lithic use-wear analysis. In V. Gaffney (ed.), Proceedings of the 1997 Computer Applications in Archaeology Conference. BAR International Series S750. CD-ROM.

Lohse, E.S., C. Schou, A. Strickland, D. Sammons, and R. Schlader

2003 Automated Classification of Stone Projectile Points in a Neural Network. Paper presented at the Computer Applications in Archaeology Conference: Enter the Past: The E-way into the four dimensions of cultural heritage, Vienna, Austria, April 8-12. To be published in the BAR series, Oxford.

Mackenzie, D. And J. Wacjman (eds.)

1985 The Social Shaping of Technology: How the Refrigerator Got its Hum. Philadelphia: Open University Press.

Madsen, T.

2001 Transforming Diversity into Uniformity - Experiments with Meta-structures for Database Recording. In Z. Stancic and T. Veljanovski (eds.), Computing Archaeology for Understanding the Past. CAA 2000. Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 28th Conference, Ljublana, April 2000, pp. 101-105. BAR International Series 931. Oxford.

Mark Bide and Associates

2000 Standards for Electronic Publishing: An Overview. standards.pdf

Markus, M.L. and M. Keil

1994 If we build it, they will come: Designing information systems that people want to use. Sloan Management Review 35(4): 11-25.

McCartney, P., I. Robertson and G.L. Cowgill

2000 Using metadata to address problems of data preservation and delivery: examples from the Teotihuacan Data Archiving Project. .

McManamon, F.

1992 Managing Repatriation: Implementing the Native American Graves Protection and Repatriation Act. CRM 15(5): 9-12.

Michener, W.K., J.W. Brunt, J.J. Helly, T.B. Kirchner and S.G. Stafford

1997 Non-geospatial metadata for the ecological sciences. Ecological Applications 7(1):330-342.

Miller, A.P.

1996 Metadata for the masses. Ariadne 5.

Miller, A.P.

1997 The importance of metadata to archaeology: One view from within the Archaeology Data Service. .

Miller, A.P.

1997 Unifying resource discovery metadata for the Humanities: an application based upon the Dublin Core. In P. Miller and D. Greenstein (eds.), Discovering Online Resources across the Humanities: A Practical Implementation of the Dublin Core, pp. 34-55. Bath: United Kingdom Office for Library and Information Networking.

Miller, A.P.

2000 Interoperability: what is it and why should I want it? Ariadne 24.

Miller, A.P., D. Dawson and P. Perkins

2002 Towards a digital cultural content forum. Cultivate Interactive 7.

Minsky, M.A.

1975 A framework for representing knowledge. In P.H. Winston 9ed.), The Psychology of Computer Vision, Chapter 6. New York: McGraw-Hill. Revised version of Memo 306, MIT Lab, June 1974.

Moen, W.E.

1998 The CIMI profile: A Z39.50 profile for Cultural Heritage Information.

Moffett, J.

1986 Archaeology, databases and microcomputers. In E. Webb (ed.), Computer Applications in Archaeology CAA 1985, pp.109-113.

Mowat, E.

2002 SCRAN, Archaeology and Education: Realising the potential of digital resources. Internet Archaeology 12.

Museum Computer Network

2003 Museum Computer Network. Standards & Controlled Vocabulary.

Museum Documentation Association

1997 SPECTRUM: The UK museum documentation standard. 2nd edition. Cambridge.

NASA: Global Change Master Directory

2003 Directory Interchange Format (DIF). Writer’s Guide, Version 8.

National Park Service

2003 National Park Service Archaeology and Ethnography Program. NPS National Center for Cultural Remains.

National Park Service

2002 Cultural resource laws into the 1970s.

National Park Service and CAST

1997 National Archaeology Database. .

National Preservation Office

1997 Preservation and digital information. .

Newman, W. and M. Lamming

1995 Interactive System Design. Wokingham: Addison-Wesley.

NDIIP

2002 Building a National Strategy for Preservation: Issues in Digital Media Archiving. Commissioned for and sponsored by the National Digital Information Infrastructure and Preservation Program, Library of Congress. Washington, D.C.: Council on Library and Information Resources and the Library of Congress.

Nepstad-Thornberry, T., C. Nepstad-Thornberry, M. Stoltz, K. de Dufour and R. Wilshusen

2002 Addressing the Curation Crisis in Colorado. An Assessment for the Executive Committee of the Colorado Council of Professional Archaeologists. Http://

Niccolucci, N. and A. D’Andrea

2002 Archaeological informatics: is there such a thing? Abstract, Computer Applications in Archaeology, Conference 2002.

Odlyzko, A.

1996 Tragic Loss or Good Riddance? The Impending Demise of Traditional Scholarly Journals. In R. Peek and G. Newby (eds.), Scholarly Publishing: The Electronic Frontier, pp. 91-102. Cambridge: MIT Press.

Payne, E.C. and R.C. McArthur

1990 Developing Expert Systems: A Knowledge Engineer’s Handbook for Rules and Objects. New York.

Peebles, C.S.

1992 Rooting Out Latent Behaviorism in Prehistory. In J.-C. Gardin and C.S. Peebles (eds.), Representations in Archaeology, pp. 357-384. Bloomington: Indiana University Press.

Perring, D. And A. Vince

1999 Liberating Archaeological Data. Project Outline. Version 3.2. Department of Archaeology, University of York.

Pinto, F., N. Ryan, T. Austin and J. Richards

2001 An Interoperable Portal for the Historic Environment. In E. Neuhold and L. Kalinchenko (eds.), Proceedings of the Third Delos Workshop: Interoperability and Mediation in Heterogeneous Digital Libraries, DELOS Network of Excellence on Digital Libraries Workshop Series.

Powesland, D., H. Clemence and J. Lyall

1998 West Heslerton: WEB-CD - The application of HTML and WEB tools for creating a distributed excavation archive in the form of a WEB-CD. Internet Archaeology 5.

Pringle, M.J. and M.R. Moulding

1997 Applications for Virtual Reality, and associated Information Technology, in the illustration of archaeological material. Graphic Archaeology 1: 22-33.

Quine, G.

2001 The role of data standards in digital access and interchange. .

Razdan, A., J. Rowe, M. Tocheri and W. Sweitzer

2002 Adding Semantics to 3D Digital Libraries.

Razdan, A., D. Liu, M. Bae,M. Zhu, G. Farin, A. Simon and M. Henderson

2001 Using Geometric Modeling for Archiving and Searching 3D Archaeological Vessels. CISST 2001, June 25-28, Las Vegas.

RCHME

1993 Recording England’s past: a data standard for the extended national archaeological record. London: Royal Commission on the Historical Monuments of England.

RCHME

1998 MIDAS: A manual and data standard for monument inventories. London: Royal Commission on the Historical Monuments of England.

RCHME and EH

1995 Thesaurus of Monument Types: A Standard for use in Archaeological and Architectural Records. Swindon: Royal Commission on the Historical Monuments of England.

Reilly, P. and S. Rahtz

1992 Introduction: archaeology and the information age. In P. Reilly and S. Rahtz (eds.), Archaeology and the Information Age: a global perspective, pp.1-28. London:Routledge.

Richards, J.

1996 The Archaeology Data Service. Archaeological Computing Newsletter 46: 19-23.

Richards, J. and D. Robinson

2000 Digital Archives from Excavation and Fieldwork: A Guide to Good Practice. Archaeology Data Service. Oxford: Oxbow Books.

Richards, J., M.P. Heyworth and J. Winters

2000 Internet Archaeology: Where next?

Robey, D.

1997 The Paradox of Transformation: Using Contradictory Logic to Manage the Organizational Consequences of Information Technology. In C. Sauer and P. Yetton (eds.), Steps to the Future: Fresh Thinking on the Dynamics of Organizational Transformation. San Francisco: Jossey-Bass Inc.

Robinson, D.

2000 Digital Archiving Pilot Project for Excavation Records (DAPPER). .

Rothenberg, J.

1998 Avoiding Technological Quicksand: Finding a Viable Technical Foundation for Digital preservation. Washington, D.C.: Council on Library and Information Resources.

Rothenberg, J.

A Discussion of Data Quality for Verification, Validation, and Certification

(VV&C) of Data to be Used in Modeling. The RAND Corporation. .

Rothenberg, J.

1996 Metadata to Support Data Quality and Longevity. Paper rpesented at the First IEEE Metadata Conference, April 16-18, Silver Spring, MD.

Rothenberg, J.

1995 Ensuring the Longevity of Digital Documents. Scientific American 272(1): 24-29.

Rothenberg, J., W, Stanley, G. Hanna and M. Ralston

1997 Data Verification, Validation, and Certification (VV&C), Guidelines for Modeling and Simulation. The RAND Corporation, PM-710-DMSO, September, 1997.

Rulf, J.

1993 Pre-processing of archaeological data. In Andresen, J., T. Madsen and I. Scollar (eds.), Computing the Past: CAA92, pp.329-332. Aarhus.

Russell, S. And P. Norvig

1995 Artificial Intelligence: A Modern Approach. New York: Prentice Hall.

Ryan, N.S.

1988 A bibliography of computer applications and quantitative methods in archaeology. In S.P.Q. Rahtz (ed.), Computer and Quantitative Methods in Archaeology 1988, pp.1-27. BAR International Series 446.

Ryan, N.S.

1995 The excavation archive as hyperdocument? In J. Huggett and N. Ryan (eds.), Computer Applications and Quantitative Methods in Archaeology 1994, pp. 211-220. BAR International Series 600. Oxford.

Ryan, N. And M. Van Leusen

2002 Educating the Digital Fieldwork Assistant. In G. Burenhult (ed.), Archaeological Informatics: Pushing the Envelope CAA 2001, pp. 401-412. BAR International Series 1016. Oxford.

Sandewall, E.

1998 Scientific Communication on the Internet. The ETAI Experience.

Schloen, D.

2001 Archaeological Data Models and Web Publication Using XML. Computers and the Humanities 35: 123-152.

Star, S.L. and K. Ruhleder

1996 Steps toward an ecology of infrastructure: Design and access for large-scale collaborative systems. Information Systems Research 7: 111-138.

Stead, Stephen

2003 Out and About with the CIDOC CRM: The Lessons and Results of the CRM Dissemination and Training Programs. Paper presented at “Sharing the Knowledge: International CIDOC Crm Symposium, Washington, D.C., March 26-27, 2003.

Stewart, Ian and Jack Cohen

1997 Figments of reality: The evolution of the curious mind. Cambridge: Cambridge University Press.

Stewart, J.

1995 Transforming the management and access of museum collections with IT. In T. Stevens and V. James, The Future for Europe’s Past, pp.37-40. Proceedings of COMETT Workshops 3. Swansea: Swansea Institute of Higher Education on behalf of the EU COMETT Programme.

Stewart, J.

1996 Where is the knowledge we have lost in information? In A. Fahy and W. Sudbury (eds.), Information: the hidden resource, museums and the Internet, pp.419-423. Proceedings of the Seventh International Conference of the MDA. Cambridge: MDA.

Stewart, J.

1997 Data into information: Twenty-five years of data collection and curation.

Text Encoding Initiative

2003 Text Encoding Initiative. Welcome to the TEI Website.

Thalmann, D., N. Farenc and R. Boulic

1999 Virtual Human Life Simulation and Database: Why and How. DANTE 99, pp. 63-71.

The Semantic Web

2003 First International Workshop on Semantic Web and Databases, Humboldt-Universitat, Berlin, September 7-8, 2003.

Thompson, R.H.

2000 The crisis in archaeological collection management. CRM 23(5): 4-6.

Tsukii, Y., A. Kihara and Y. Ugawa

2001 Pubic Domain Image Databases for Taxonomic Research and Education: A Case Study, Protist Image Database.

Ucko, P.

1992 Forward. In P. Reilly and S. Rahtz (eds.), Archaeology and the Information Age: a global perspective, pp.ii-vi. London: Routledge.

US Army Corps of Engineers

1999 Department of Defense and U.S. Army Corps of Engineers Curation Options Project, Western and Mid-Atlantic States Executive Summary. ... n/Legacy/Curations/execsummary.html

US Army Corps of Engineers

2003 MCX-CMAC. Mandatory Center of Expertise for the Curation and Management of Archaeological Collections.

US General Accounting Office

1987 Cultural resources: problems protecting and preserving federal archaeological resources. GAO/RCED-88-3. Washington, D.C.: United States General Accounting Office.

van den Dries, M.H.

1998 Archaeology and the application of Artificial Intelligence. Archaeological Studies Leiden University 1.

van Leusen, M., S. Champion, J. Lizee and T. Plunkett

1996 Towards a European Archaeological Heritage Web. In H. Kamermans and K. Fennema (eds.), Interfacing the Past: Computer Applications and Quantitative Methods in Archaeology CAA 95, pp. 507-514.

Virtual Museum of Canada

2003 Virtual Museum of Canada homepage.

Walsh, J.P. and T. Bayma

1996 The virtual college: computer-mediated communication and scientific work. The Information Society 12: 343-363.

Webb, E.

1986 Introduction. In E. Webb (ed.), Computer Applications in Archaeology 1985, p. iii.

Weibel, S.

1996 A proposed convention for entering metadata in HTML.

Weibel, S., R. and E. Miller

1997 Dubline Core Metadata Element Set: Reference Description. .

Whipp, P. And T. Lewis

1989 Model based reasoning and simulation. In S. Vadera (ed.), Expert System Applications, pp. 97-113. Winslow.

Wilcox, J.D.

1978 The automated archaeologist. A review of new personal computing and office automation methods of relevance to site recording, retrieval, analysis and publication. In S. Laflin (ed.), Computer Applications in Archaeology 1978, pp. 48-52.

Williams, R.

1997 The Social Shaping of Information and Communications Technologies. In H. Kubicek, W.H. Dutton and R. Williams (eds.), The Social Shaping of Information Superhighways: European and American Roads to the Information Society, pp. 299-338. New York: St. Martin’s Press.

Winder, N.

1994 Designing a database for animal boens. In R. Luff and P. Rowley-Conwy (eds.), Whither Environmental Archaeology, pp. 33-40. Oxford.

Winters, J.

2002 Editorial - Towards sustainability in electronic publishing. Internet Archaeology 11, pp. 1-4.

Wise, A. and P. Miller

1997 Why metadata matters in archaeology. Internet Archaeology 2. .

Wise, A. And J. Richards

2001 Digital preservation in archaeology. .

Wissenburg, A.

1997 Guids to Good Practice of the Arts and Humanities Data Service. .

Woolridge, M.J. and N.R. Jennings

1998 Formalizing the Cooperative Problem Solving Process. In M.N. Huhns and M.P. Singh (eds), Readings in Agents, pp. 430-440. San Francisco: Morgan Kaufmann.

Woolridge, M.J., J.P. Muller and M. Tambe

1996 Intelligent Agents II. LNAI 1037. Berlin: Springer.

Yang, C.

1986 Relational databases. New Jersey.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download