TRANSLATE ENGLISH to Français, Deutsch, Italiano, Português, Español! ^Explanation

AUTOMATION AND MULTI-PART DESCRIPTION
Jan Smits, Koninklijke Bibliotheek, Den Haag, The Netherlands

© LIBER and author

© images: Koninklijke Bibliotheek

Abstract
In this paper I will try to give a definition and examples of multi-part cartographic materials (CM) as may be discerned by certain groups of users, producers, cataloguers etc.
To my opinion a good organization of the cartographic materials to be described into similar groups, and the use of 'copy'-facilities in automated cataloguing systems will increase the amount of cartographic materials to be processed to hitherto unknown quantities and unknown quality.
Furthermore I will try to show that a one-time input of a multi-part CM-description into an automated system can produce a multitude of output forms.
Well designed automated cataloguing systems tailor-made for CM will give a better insight into links between CM mutually and between CM and other carriers of information.

Automation
The role of automation in the realm of accessing cartographic materials is to be better able to manipulate the data of which cartographic materials consist.
Though automation may make processing of data speedier than it was in the old times of manually written or typed catalogue-cards this cannot be the basic drive. Mechanization, which sometimes seems to be an euphemism for automation, for a long time seemed to be the goal of developing-departments of libraries.
But automation not only can speed up processing, but also creates new possibilities for information-queries, especially in an area as cartography, where document-retrieval counts for a minor part of the demand. The major part of demand consists of information-retrieval. To be able to satisfy this demand for information we have to cut up the initial information in the smallest possible parts to make cross-reference c.q. boolean questions possible.
However, not only do we have to cut up the formal information, as used in fields 1 to 6 and 8 of the ISBD, but also that information contained in the cartographic and related data which constitutes the cartographic document. This information should be processed in the fields for coded information in the divers MARC's and verbalized as much as possible in field 7 (notes) of the ISBD (1). If by cross-reference or boolean query one finds the coded information one can read in the note of that record (or the records) what has been coded exactly.

Multi-part cartographic materials
But automation gives also a great advantage when relations or links between cartographic materials must be made transparent. These relations or links can be within a record or between records. This paper concerns itself with the relations between records and the impact it may have on the processing of records.
To define which cartographic materials are meant I use the term 'multi-part cartographic material'. The definition which goes with is mainly based on the definition of 'map series' in Cartographic Materials, Appendix D, p. 176, and with some adjustments reads as follows:

A multi-part cartographic material consists of a number of related but bibliographically distinct cartograp- hic units intended by the producer(s) or issuing body(ies) to form a single group, or seen as such by mapcurators or map-users. For bibliographic treatment, the group is collectively identified by any commonly occuring unifying characteristic or combination of characteristics including a common designation; part identification system; scale; publisher; cartographic specification; uniform format; collection; etc.

I find the use of the word 'series' (2) misleading as this bears (also in the AACR2) the connotation that they must be linked by bibliographical elements which appear on the items of which a series consists. Besides I do not limit this definition to physically separate items, as the parts may be bound together in a book or atlas. Also I specifically want to include the point of view a mapcurator or map-user may have, which may designate a collection of cartographic materials as a multi-part item.
This last demand excludes the equation of the definition with 'component parts' (3) as these are specifically designated for "... parts that are issued with, in, or as part of a host item, and are dependent on the host item for purposes of bibliographic identification or access". The Guidelines for Component Parts are implicit for bibliographically dependent items and primarily exclude bibliographically indepent items. Furthermore they are mainly meant for 'analytic description' and not for retrieval- or catalogue-structuring.

Example
Before looking at a specific example I would like to ride one of my hobby-horses, that is: which cartographic materials should be accessed? As said and written before (4) I think there are two groups of cartographic materials which can be/are recognized by mapcurators as potential materials to be accessed, being those which are independently published and those which are dependently published. In the first group we find the traditional materials which always have, to the best possibility, been accessed in one way or another. The other group, mainly illustrations or appendices to books and journals, have been accessed sporadically (5).
Before validating the importance of the one or the other category our profession should ask itself again the overriding question: why do we access cartographic materials? In my opinion the unquestionable answer should be: because we want to make the whole of cartographic information transparant in support of society in general, especially a society in which spatial information becomes more and more important to support or instigate social, economic, ecological etc. processes and developments. Though policy- makers and scientists can draw on vast amounts of alphanumerical information there are but few means to (mentally) visualize the links between and the complexity of this information as a tool to evaluate or take decisions for future developments (6).
The fact that there are so many cartographic materials produced every year (7) shows some of its importance. However this fund of cartographic information is spread over such an, almost unlimited, amount of sources that it is difficult to find the right item, if one is already conscious of its existence! But if we want to stress the importance of accessing these vast amounts of cartographic information to the decision-makers who decide about means and personnel we should seek support from the users (Geographical departments, economists, statisticians, political representatives, governement agencies etc.). Of course this will be a long and arduous road, but it may free us of the stigma of the appendix (which always carries the danger of removal).

For practical and obvious reasons (lack of resources and lack of personnel) we have busied ourselves mainly with the independent materials.
(large image: 135 Kb)

Photo 1. Variety of maps in the Scientific atlas of The Netherlands, 2nd ed.

BUT, the cartographic information stored in the dependent materials is not less valuable. I would even say that, as these concern mainly thematic cartography, they will become even more and more important. The question arises if the (scientific) community misses the non- accessed data or that they can be satisfied with the existing accessing of traditional materials. As I do not know of any literature concerning this I have to formulate my own answer. I think part of it is ignorance and part of it is the inability to visualize how this stupendous task can be met, a question you as mapcurators will pose yourselve also I think.

But before trying to indicate a solution I want to show you again with an enlightening example why I am so obsessed with this question.
(large image: 60 Kb)

Photo 2. Scientific atlas of The Netherlands, 2nd ed.

This example concerns the 2nd edition of the Scientific Atlas of The Netherlands. It is divided into 20 parts of 24 pages each and according to its editors it contains 1,000 [!] different thematic maps concerning The Netherlands or its parts and its relations with the world at large (8). The contents of this atlas concerns many regions, many themes, many cartographic methods to depict the information, many scales, many dates of contents, many dates of publication etc. etc.. In my opinion all too much to put into one record if one doesn't want to fall into too great generalities.

(large image: 258 Kb)

Photo 3. All 20 volumes of the Scientific atlas of The Netherlands, 2nd ed.

Description
When describing this conglomerate of cartographic materials one can create a simple record to have it over and done with, or one goes deeper than that. If we decide for the simple record to access this document, then the only way in which we can give some more comprehensive information than just the fact that it concerns the 'Scientific Atlas of The Netherlands', is to write out the parts in the notes:

Atlas van Nederland / [samengesteld door de] Stichting Wetenschappelijke Atlas van Nederland. -2nd. ed. - Various scales (E 003°15'-E 007°20'/N 053°35'-N 050°45'). - 's-Gravenhage : SDU, 1984-1990. - 20 parts of 23 p. : in col. ; 36 cm, in box 39 x 32 cm.
Notes.
Consists of: Deel 1: Bevolking, Deel 3: Steden, Deel 8: Werken etc.

As you can see from this description this can be done only when the whole of the publication is published. If not one has to modify the record each time a part is published.

(large image: 257 Kb)

(large image: 248 Kb)

(large image: 255 Kb)

Photo 4-6. Examples of the variety of areas and themes in the Scientific atlas of The Netherlands, 2nd ed.

If we want to go deeper than this the problem seems to become one of resources and personnel!
As an intemediate stage we can write out the parts, as they all are concerned with a specific theme or cluster of themes. Either by extending the contents of the note 'Consists of: ...' or by creating a multi-level description. However it is the experience that extensive notes at a certain moment may hit the barrier of the limited size of a record. I know that there are systems which work with subsequent records, but I guess that there will be problems when a part must be interjected (9). I myself prefer a multi-level description:

Atlas van Nederland / [samengesteld door de] Stichting Wetenschappelijke Atlas van Nederland. -2nd. ed. - Various scales (E 003°15'-E 007°20'/N 053°35'-N 050°45'). - 's-Gravenhage : SDU, 1984-1990. - 20 parts of 23 p. : in col. ; 36 cm, in box 39 x 32 cm.
Notes.

    Deel 1, Bevolking / ...

       Map 1, title ...

       Map 2, title ...

    Deel 3, Steden / ...

       Map 1, title ...

    Deel 8, Werken / ...

       Map 1, title ...

Though resources and personnel may be stretched in this way, the problems still can be solved. But what if we'll describe all cartographic materials??? (10)

Copying-facilities and multi-level recording
Let me tell you that my colleague Jos de Vries described all 1,000 cartographic materials and 20 parts very extensively in less than 8 weeks, give or take a few days. And he didn't even have to hurry! But there were some prerequisites: He is a full-time map-describer with 10 years of experience and he uses a well-researched and tested automated access-system.
One of the features of this system is the copying-facility for on-line (11) and mother-daughter facility for off-line processing. Using a copying-facility means that one copies an existing record from the database, modifies it according to the new map to be described, and writes it back to the database as a new record. The off-line mother-daughter method is based on the same technique, only that the copying of the existing record takes place when the off-line created record is added to the database. These methods, especially when similar or multi-part cartographic materials have to be described, can add 30 to 40% to normal production. One then even can think about creating separate records for sheet-maps, like the topographic series (12), except that one needs also facilities for multi-level recording. However the one who authorizes the description before it is stored in the database has to check very carefully for faults, as because of these methods they multiply easily.
The multi-level recording must also be a special feature of the system used. The method asks for the linking of the records in such a way that multi-level output is possible. The linkage can be by referring to a recordnumber. At the same time one must supply a sorting-criterion (e.g. an alphanumerical designator) to be able to sort the underlying records. Depending on the system one can create 2 to 9 levels (13).
To be able to use the database-records for various kinds of output (e.g. multi-level, independent, one-level), it is preferable to create complete records, even when the prime drive is to be able to create multi-level output. This seems to be a waste of precious processing-time, but here the copy-facility gives its greatest advantage.
But it is not only necessary to be able to manipulate the database for local output, but we must also have in mind the cooperation of two or more collections in one system and the exchangeability of records to other systems. As local collections, catalogues, lists, retrieval systems etc. differ one from the other it is necessary that they can gain as much as possible from already existing databases (14).

Back to GdC Homepage
Back to Jan Smits' Homepage