|an excavation methodology. The single context method seems to suit the best the contemporary requirements for precision and integrity of recovered archaeological data. At the same time, the excavation sequence implied by the method, when each archaeological context is excavated, examined, and recorded individually in a successive order, fits the demands of a 3-D digital data record in a computer. While the combination of the introduced recording technique and the single context method is the most efficient for the future excavations in Russia, there is a vital need for utilization and re-interpretation of previously collected excavation data. Therefore, the capabilities of digitized data for 3-D excavation record reconstruction are also to be examined.
In order to fulfill both tasks it was proposed to test direct 3-D recording technique against traditional excavation strategy in a sample excavation. Thus, along with the in-the-field elaboration of the computer recording methods the potential of the traditionally recorded evidence was evaluated.
3. Digital excavation record. Theory and practice
3.1. Software environment
The excavation record is a complex of spatial (geometrical), text, and photo and video data. Digital data procession unleashes rich potential for its integration, unification, and analysis. Thereby, evidently the most powerful solution for handling digital excavation record is a GIS-based approach. At the same time, while GIS software systems have proved their efficiency with 2D spatial data archives, their 3-D modeling and analysis capabilities are insufficient.
Spatial data is either the most important excavation record or the most difficult to obtain and process digitally. Taking all these points into consideration, AutoCAD 2000 was used as a core modeling environment for the project. Its powerful COGO (coordinate geometry) engine satisfies strict requirements for accurate and fast direct recording of excavation record on-the-fly in situ, and paper-form data digitizing and procession with minimum of conjectures and approximations. Both vector and raster data support along with integrated database engine for all common formats and internet/intranet support make AutoCAD 2000 the preferable solution for mixed-type and mixed-format data integration.
3.2. Experimental excavations: extending theory in the field
The huge complex of archaeological sites, situated on the upper of the Dnepr river, – Gnezdovo – consists of the central fortified settlement, the hill-fort or “gorodishe”, surrounding rural settlement with area totalling c. 16 hectares, and four groups of mounds, which number little less then 4000 burials (figs 4-5). Belonging to the Viking period, Gnezdovo is considered the major principal proto-urban trade center in the Russian inland on the way from Scandinavia to Byzantium. It has been strongly believed, that the archaeological complex is limited by the first riverbed terrace until mid-1990-s, when the traces of the settlement in the flood-lands of Dnepr was discovered. A series of prospecting excavations has revealed wet cultural layers rich of organic material of an excellent preservation.
The experimental excavation ground was started on a slope on a terrace edge covering both terrace and a flood plain on an area of 40 m?. Thus, the stratification link between flood plain and terrace was intended to be examined. The excavation record has been kept either traditionally, with paper-form sketch plans and drawings, or directly modeled in AutoCAD; for the potential of both recording strategies could be explored and compared, and the integrity of record mutually checked one against another. At the same time the excavation technique, based on the principle of the described above planar/square system, was modified, where it was possible, so that each major stratification unit was excavated individually in a strict consequence of their deposition (fig. 6). The digital excavation record was held with a Notebook computer in three main parallel directions:
3.2.1. Stratification record
Stratification record was the most difficult to handle digitally in 3-D modeled space. Geometrically each stratification unit is a solid of a complex irregular shape. Two main techniques were used to measure its surfaces. In case the “upper” surface of a stratification unit was completely and clearly revealed, the measurements were taken by a 40x40 cm elevation grid, while