Table of Contents
Index
Glossary
Table of Contents
Index
Glossary

   
4.2.5 The characterisation by a set of specified bibliographic data

The fourth type of characterisation used in the modular model is a characterisation by a specified set of bibliographic data. The bibliographic data do not actually characterise the information contained within a document, but rather give information about that document: they are metadata. The function of this type of characterisation, as well as the characterisation by the range, is not to delineate concepts, but to distinguish between different uses of the presentations of the concepts. By the range, a distinction can be made between information that has been presented for inclusion in a single article or in more than one article. By the bibliographic data, a distinction can be made between different articles about the same concepts.

Like the characterisation of the information by its scientific content, this type of characterisation is traditionally used for indexing. Bibliographic data are used to label each module, and in particular the article as a whole^4.17, because most of the bibliographic data apply to the article as a whole as well as to all of its constituent modules. The bibliographic characterisation of the article is made explicit in a meta-module m1a Bibliographic information that lists all bibliographic data. This module will be presented in section 4.2.7. We do not address in detail the question as to which types of bibliographic data are most suitable to characterise the modules. We assume that a suitable bibliographic component for the typologies will be available. Examples of existing categories of bibliographic data are given in the bibliographic components of e.g. the highly detailed SGML DTD used by the publisher Elsevier Science [Elsevier, 1999], and the more general resource description record known as the Dublin Core [Dempsey and Weibel, 1996].

In the modular model, a rudimentary bibliographic characterisation takes into account at least the following types of bibliographic data, which are also included in the original corpus articles:

1.

Article title;

2.

Author names;

3.

Affiliations of the authors;

4.

Name of the  journal the article is published in;

5.

Date of publication;

6.

A unique identification.

A module can then be characterised by its location in the `bibliographic space'. This bibliographic space is at least a product space spanned by dimensions corresponding to a) titles, b) author names, c) affiliations, d) journal names, and e) dates. There can be more than one author, with more than one affiliation, the number of authors determining the number of dimensions. The geometry of the subspaces with journal names, author names and affiliations can be determined in terms of `maps' by scientometrics, which provide a sense of distance between the work of scientists (for example, [Van Raan, 1988]).^4.18

The user can  search along the axes of the bibliographic space for a conjunction of these bibliographic data. We do not treat the unique identification number that we assign to each module on the same footing as the other types of bibliographic data, because it is redundant to use this type of bibliographic data in a search in conjunction with any other characteristic. Hence, the identification number is not associated with a character dimension in the `bibliographic space'.