The article examines the most well-known works of Russian archaeologists on the creation of databases of archaeological sources, and assesses the current state of this problem. The author's version of a theoretically and information-based database management system (DBMS) for monuments of the Pre-Scythian period on the lower Don is proposed. Each stage of application design is analyzed in detail: description of the subject area, selection of information objects and information analysis, determination of relationships between information objects, and construction of an information-logical model of the subject area under consideration. The author gives a general idea of databases, information models and other technologies, the correct use of which is necessary for a modern specialist-archaeologist.

Keywords: archeology, burial mounds, burial rites, pre-Scythian period, early Scythians, information technologies, databases.

Introduction

The most representative work on creating a database of archaeological sources today is the joint research project "Statistical processing of funerary monuments of Asian Sarmatia" of the IA RAS and the Italian Institute of the Middle and Far East on the topic "Funerary monuments of early nomads of the Eurasian steppes: experience in computer processing of archaeological materials", carried out under the leadership of M. G. Moshkova [Statistical processing..., 1994, 1997, 2002]. This is the first project of its kind, successfully implemented, widely known and accessible to most researchers. The appearance of these publications caused a lively discussion (and even critical opponents recognized them as an undoubted step forward in archaeology), gave an impetus to the development of archaeological informatics in general, and prompted a number of researchers (including us) to create what J. K. Garden defined as " logically unrelated local contributions to the source study apparatus of archaeology, moreover, although without planned interrelations with each other, they nevertheless make up a logical unity " [1983, pp. 82-83].

Recognizing the absolute advantages of this study, it is necessary to note its inherent shortcomings (as well as many other innovations). We consider the most significant lack of the database itself accessible to a wide range of researchers (the publication does not include any electronic media containing at least an introductory version of the product, and the authors obviously did not consider it possible to place it on the IA RAS website). So, we can't estimate how correct it is.

page 92
"the system allows you to enter/edit and view data in a structured form, extract data in a form suitable for statistical processing", as well as a number of other interface characteristics [Lazarev and Barbarunova, 1994, p. 42], although the authors position their database not as focused on "one-time statistical processing", but "one-time statistical processing".in a sense, universal", where "the maximum display of the source information should be provided" [Ibid., p. 39].

Another equally serious flaw is the insufficiently correct, from the point of view of computer science theory, justification of the project. For example, the authors use a relational DBMS (obviously, dBase III+, but this is not explicitly mentioned anywhere in the work), which is focused, respectively, on a relational data model with special characteristics. But we will not find the concept of "relational model" anywhere in this work, although the authors ' idea of the data model corresponds in principle to the above: "a data set (database file) is actually a table with rows ("records"). they describe a single object, and columns ("fields") contain the values of a specific variable. The fields can be of different types (i.e. contain values of different quality - numbers, words, etc.) and of different lengths, but the set of these fields and their characteristics are the same for each object. In other words, each data set corresponds to a kind of form in which information about each of the objects should be entered" [Ibid.]. This leads to the fact that the authors do not comply with the requirements of normalization of relations - due to the lack of understanding about it. As a result, the table "graves" [Ibid., p. 43] turns out to be non-normalized and does not correspond to the relational model, which inevitably leads to duplication of descriptive data and problems associated with maintaining their integrity. In general, we can agree with the information-logical database model, if we consider as such chapter 5 of the considered work "Database organization" [Statistical processing..., 1994, p. 39-53], but we have to state serious deviations from the canonical model. Also, the authors do not use the methods of developing software systems (SADT, DATARUN) and the technologies that implement them (CASE tools).

Another disadvantage, in our opinion, is a certain discrepancy between the created database and the stated requirements in terms of maximum display of source information. The logical structure of the database presented in this paper [Ibid., p. 43] does not include graphic, photo information, or auxiliary data about burial complexes (for example, authorship of excavations, addressing information sources, information about publications, etc.). Although adding such information (for example, using the OLE, MEMO, hyperlinks, or related fields files) does not require the exorbitant capabilities of machine memory.

Some researchers also noted the disadvantages associated with the authors ' use of the principle of encoding the values of funeral rite features (for more efficient use of machine memory and easier statistical processing). Indeed, the presentation of data in the form of one or another value from the standard set provided in advance and, accordingly, input/editing is carried out by a person, which obviously entails all the difficulties associated with the so-called human factor - subjectivism, errors, etc. The presence in the database of fields with additional information (OLE, MEMO). it would eliminate some of these difficulties. But a much more serious problem is the very identification of features and, accordingly, the definition of a number of their categories-values. The literature has already considered the current situation in Russian science associated with the lack of a classification paradigm and archaeological typology. Unfortunately, this issue is still being debated and a unified approach has not been developed. Therefore, any attempt to create a system for determining the characteristics of funeral rites and their meanings can be criticized. However, this does not detract in any way from the advantages inherent in a particular formalization of funeral rite data [Kamenetsky, 1983; Gening et al., 1990], including those developed by a team led by M. G. Moshkova.

In our work, we use the experience gained by scientists of the IA RAS in the process of creating databases on the Sarmatian culture [Statistical processing..., 1994, 1997, 2002] precisely in terms of the proposed formalization option. Creating our own encoding system for archaeological material is not the goal of the proposed research and is beyond our competence.

Designing an application

When starting to design a user application (which is a combination of a certain set of domain tasks), including its database, it is necessary, following the theory and practice of building information systems, to first prepare a description of the domain. Using an analytical approach, the foundations of which were developed by J. Martin [1984], we should identify "a set of data and various information about objects and processes that characterize this area" [Bekarevich, Pushkina, 2003, p. 68]. Further

page 93
it is necessary to formulate "general requirements for the application, functional restrictions, a description of the technology for working with data and documents, requirements for the user interface when working with the database, for entering and correcting reference information, for preparing and entering documents of operational and accounting information, as well as viewing previously prepared and saved documents, and requirements for reports from the database data" [Ibid., p. 656]. Our subject area can be characterized as the funerary rites of the nomadic population of the Lower Don in the Pre-Scythian and Early Scythian periods (according to archaeological data). The created application should be limited to management accounting of available (and newly received) data on archaeological sites of this region and period in order to generate information about the specified group of burial complexes for subsequent processing - traditional analytical or using mathematical statistics methods. The object of automation is a burial complex with its inherent information lines. The application functions defined by its target component include input, editing, ordering, various structuring operations, data extraction in a form suitable for statistical processing, data analysis, report generation, and so on. Reference and operational accounting documents of the application contain reliable information about objects "that the application operates on the basis of identifiers (object codes)" [Ibid., p. 657]. In our case, these are field research materials of Russian archaeologists - reports on the work of expeditions and individual researchers, publications of various levels. Information is presented as a set of details that define the object of the subject area (reference documents also include a system for encoding features-funeral rite details). The technology of working with the data of the proposed application (in contrast to business processes, for which the methods used here were mainly developed) is determined by their nature - the needs of operational processing "in real time" are not relevant here (this is also noted by V. V. Lazarev and Z. A. Barbarunova [1994, p. 41]).

At the next stage of application design, you need to select information objects and perform an information analysis. To do this, we need "based on the domain description... identify source documents and their details to be stored in the database" [Bekarevich and Pushkina, 2003, p. 70]. If in business projects the basis for highlighting information objects is usually such documents as, for example, invoices, invoices, payment orders, etc., where "the form of an out-of-machine document already displays the data structure, since any document combines logically interrelated details" [Ibid., p. 69], then in our case in this case, everything is much more complicated. The process of information analysis is based on the idea that "the structure of non - vehicle information is reflected in its representation by separate structural units-details and their placement in source documents" [Ibid., p. 68]. Sources of archaeological data , such as excavation reports and publications, are usually not clearly structured documents, including due to the lack of common approaches to classifying the material, as mentioned above. The first stage of information analysis, i.e. structuring information in the subject area, requires either creating your own system for determining the characteristics-details of funeral rites and their meanings, or using existing ones. Due to the above reasons, we considered it possible to implement the formalization option proposed by the team led by M. G. Moshkova (the creators intended its subsequent expansion and addition). Thus, the source document of the application being created is a report (or publication) about the excavation, and the details are the signs of the funeral rite highlighted in the specified work (with some changes and additions made due to the specifics of its own material) [Statistical processing..., 1994, 1997, 2002].

In business projects, information is usually contained in a number of different reference and accounting documents, thus facilitating the selection of information objects. In our case, there is only one type of source document-either an excavation report or a publication, which somewhat complicates the task. According to the rules for selecting information objects [Bekarevich and Pushkina, 2003, p. 70], it is necessary to determine functional dependencies between document details and then group those that are equally dependent on key details (the latter are understood as identifiers of all document information). These groups make up information objects. To do this, "the role of banking details in the structure of the object's information is analyzed", "one or several banking details that serve as a general identifier of all document information" are identified, i.e. key, and then descriptive, which are "uniquely determined by the established document identifier" [Ibid., p. 70]. For each descriptive item, a key item is defined and a link is drawn with an arrow pointing in the direction of the dependent item (Table 1).

From the analysis of the document, it is obvious that the Local Zone Name (LZ) attribute is an opi-

page 94
Table 1. Functional dependencies of banking details (signs of a funeral rite)

page 95
End of Table 1

negative and depends only on the key information Region Number (HP). The time of construction of the mound( VRSOOR), the size of the mound (VELNAS), the composition of the mound (SOSNAS), the number of burials of one period in the mound (KPOPER), the position of the burial on the plan (POLPOGR) are descriptive details, and each of them depends only on the key detail Mound number (NC). You should pay attention to the relationship between the Region Number (RR) and Mound Number (NC) banking details . One value of the NR key corresponds to one value of the RR banking details, since the mound is located in a specific region. Similarly, the relationship between the Mound Number and Burial Number details is determined : one value of the Mound key corresponds to one value of the NK detail. since the burial is located in a specific mound. Thus, the NC attribute is key in one connection and descriptive in another-the so-called transitive dependency (but "special actions to split this dependency are not required when using the above rules" [Ibid., p. 74]). Further analysis of the document shows that the key details of the NPOGR, in turn, are descriptive for the details of the Number of finds in the mound (NNAHNAS), the number of structures inside the grave pit (NCONSTR), the number of bones (NCOS), the number of the buried (NSKEL), the number of the vessel (NS) , the number of the item of weapons (NV), The number of the decoration (NU), the number of the inventory item( NI), the number of the ritual substance (NR), and the Number of the ritual item (NR). Indeed, even here one value of the listed keys NNAHNAS, NCONSTR, NCOS, NSKEL, NS, NV, NU, NI, NRV, NRP corresponds to one value of the NPOGR props, since finds in the mound, structures in the grave, animal bones, skeletons, dishes, weapons, jewelry and other inventory relate to a specific burial. Accordingly, there are still a number of transitive dependencies, but quite surmountable, as already mentioned above. Now, according to the rules for selecting information objects, we can "group banking details that depend on the same key details and combine them with key details into one information object" [Ibid., p. 74] (Table 2).

Thus, based on the analysis of our data, the following information objects are distinguished: Region, with its name; Mound, with general information about the mound; Burial, with general information-

page 96
Table 2. Grouping banking details within an information object

Information object details*

Flag for a key or index

Name of the information object

Semantics of an information object

1

2

3

4

NR

Unique

Region

Territorial affiliation of the mound

NLZ

nk

»

Kurgan

General information about the mound

NR VRSOOR VELNAS SOSNAS KPOPER POLPOGR

NPOOGR

»

Burial

General information about the burial

NC NAME MPOGR NADSOOR FORM VELMOG RITE KOLSKEL SPSOHR ZERK

NNAHNAS NPOOGR

Unique, composite

Nakhodki

Data on burial-related finds in the mound of the mound

NAKHNAS

NCONSTR NPOOGR

The same thing

Structures

Data on structures inside the grave pit

CONSTR

NCOS NPOGR

»

Dice

Presence of animal bones in the burial

VIDK MESQUITE

NRHPOGR

»

Substances

Presence of ritual substances in the burial

VRITV

NRP NPOOGR

»

Items

Presence of ritual objects in the burial

VRITPR

NSKEL NPOGR

»

Buried

General information about the buried person

POLSKEL ORISKEL OSKEL POSSKEL POLRUK POLNOG

NS NPOOGR

»

Tableware

Data on the ceramic inventory in the burial site

CS KATS MATS TEHS MPRS PLACES

NV NPOOGR

»

Armament

Information about weapons in the burial

KV CATV MATV METV HARV

page 97
End of Table 2

1

2

3

4

WELL NPOGR

Unique, composite

Decorating

Presence of decorations in the burial

KU KATU STILU MATU MESTO

AND NIOGR

The same thing

Other

Data on other inventory categories (tools, reins, etc.)

KI KATI MATI

* See Table 1.

information about the burial; Finds, with data on burial-related finds in the mound of the mound; Structures, with data on structures inside the grave pit; Bones, with information about the presence of animal bones in the burial; Substances, with information about the presence of ritual substances in the burial; Objects, with information about the presence of ritual objects in the burial; Buried, with general information about the buried person; Dishes, with data on ceramic equipment in the burial; Weapons, with data on items of weapons in the burial; Jewelry, with information about the presence of jewelry in the burial; Other, with data on other categories of inventory (tools, bridle, etc.).

At the next stage of application development, it is necessary to identify the relationships of information objects, which are usually determined by "the nature of real objects, processes, or phenomena displayed by these information objects" [Ibid., p. 78]. In database theory, three types of information object relationships are used: one-to-one, one-to-multi-valued, and multi-to-multi-valued. The former occur when each instance of one object corresponds to only one instance of another, and vice versa. For one-or multi-value relationships, each instance of one object can correspond to several instances of another, but not vice versa. At the same time, they say that the second object is hierarchically subordinate to the first. Multi-valued relationships (which are not implemented in relational databases) represent a similar variant, where the converse statement is also true.

The relationships between Region → Mound items are one-or multi-valued, because the region includes many mounds, and the mound is located in one specific area. Communication between them is carried out by a unique key of the main Region object - the Region number, which is a descriptive attribute in the subordinate Kurgan object. Relationships between Kurgan →objects Burials are also single-valued, since the mound includes many burials, and the burial is included in one mound. Communication between them is carried out by the unique key of the main object Kurgan-the Number of the mound, which in the subordinate object Burial is descriptive props. The Burial object and objects Found, Bones, Structures, Substances, Objects, Buried Person, Dishes, Weapons, Jewelry, etc. are also in a one-to-many relationship, but here the Burial is the main object, and all the others are subordinate (the argument is the same: for example, a burial may include several vessels, but a specific vessel is only included in one burial). Communication is carried out using the unique key of the main object - the Burial Number, which is included in the composite keys of subordinate objects.

Based on the identified information objects and relationships between them, we can build an information-logical model of the subject area under consideration. In Figure 1, it is shown in canonical form and objects are placed on levels, where zero corresponds to objects that are not subordinate to any other.

Conclusion

Based on the created infological model, we can build a datalog model in a specific DBMS. The logical structure of the relational database will be an adequate representation of the resulting information-logical model. Each information object of the model will be displayed as a relational table, where each column (field) corresponds to one of the object's details, and rows (records) correspond to instances of the object. To manage our database, we chose the Microsoft Access relational DBMS as one of the most convenient and affordable tools. In addition, it can work as

page 98

Figure 1. Information and logical model of the considered subject area in canonical form.

Figure 2. Access data schema.

page 99

Figure 3. Final Access data schema.

page 100
an SQL server client, such as, for example, Microsoft SQL Server, reaching a higher level of collective use of information.

An Access data schema that clearly shows the structure of our database is shown in Fig. 2. Due to the fact that we use a special encoding system for attribute values, most of the table fields contain duplicate names. Using the Table Analyzer Wizard, each table was divided into several new linked tables (the converted main table and reference tables by the number of fields containing duplicate data).

Thus, the normalization procedure is finally completed and the problem of representing archaeological data in letter codes for use in statistical calculations is solved. In addition, OLE object fields containing the original graphic and text information (JPEG files) have been added to the tables. In our opinion, the created database on the funeral rites of the nomadic population of the Lower Don in the Pre-Scythian and Early Scythian periods should provide opportunities for both simplified use of data for statistical calculations and obtaining the most complete information. The changes we made to the Access data schema are shown in Figure 3.

List of literature

Bekarevich Yu. B., Pushkina N. V. Microsoft Access Tutorial. - St. Petersburg: Bhv-Petersburg, 2003. - 720 p.

Garden J. K. Teoreticheskaya arkheologiya [Theoretical Archeology], Moscow: Progress Publ., 1983, 296 p.

Gening V. F., Bunyatyan E. P., Pustovalov S. Zh., Rynkov N. A. Formalized statistical methods in archeology: (Analysis of funeral monuments). Kiev: Nauk, Dumka Publ., 1990, 304 p. (in Russian)

Kamenetsky I. S. Kod dlya opisanie pogrebalnogo obryada [Code for describing the funeral rite]. Drevnosti Dona, Moscow: Nauka Publ., 1983, pp. 221-250.

Lazarev V. V., Barbarunova Z. A. Organization of the database / / Statistical processing of funeral monuments of Asian Sarmatia / ed. by M. G. Moshkov. - Moscow: ONTI PNC RAS, 1994. - Issue. 1: Sauromatian Epoch, pp. 39-53.

Martin J. Smith Planning the development of automated systems, Moscow: Finance and Statistics, 1984, 196 p.

Statistical processing of funerary monuments of Asian Sarmatia / ed. by M. G. Moshkov, Moscow: ONTI PNC RAS, 1994. 1: The Sauromatic era. - 224 p.; 1997. - Issue. 2: Early Sarmatian Culture (IV-I centuries BC), 278 p. (in Russian). lit., 2002. - Vol. 3: Central Sarmatian culture. - 143 p.

The article was submitted to the Editorial Board on 15.05.09.

page 101

Permanent link to this publication: