Document repositories

An SDX-2 server hosts applications that gather the architecture and all resources of a document system. But, the primary objective of SDX is to be searching tool so that the core resource of an application is the document repository, which is the virtual location where XML document are indexed annd can be retrieved.

Definition

The document repository defines a group of documents which can be retrieved. This group shares common features, among which several fields and their properties. A document repository is therefore defined by a set of fields and by the values of those fields, produced when documents are indexed.

A search unit

An SDX application may contain several document repositories, but it must contain at least one (even an empty one). Moreover, it is very easy to search various repositories simultaneously, whether defined in a single application or in different applications. This flexibility leads to set the same choice as for the difination of an application : how to choose between the definition of a single document repository or distinct document repositories ?

list of search fields: A document repository is defined by a list of search fields. If two collections have to be indexed by similar search fields, then it may be relevant to make a single repository. On the contrary, if search fields are rather different, two repositories may be better, both to make indexing easier and to improve performance.
definition of document sets to be searched: SDX includes mechanisms which allow to consider only parts of repositories when searching documents: filters. If it is necessary to use filters in all searches and if those filters are not often modified, it may be easier to create distinct repositories in order to separate document collections which bring specific search contexts. Since various repositories can be searches, it is always possible to carry a global search, even if repositories have been separated.

Conclusion

We would like to underline that, even though the concept of document repository may be associated to the concept of document collection or to the concept of document type, SDX does not set any restriction on documents to be indexed in the repository. For example, XML documents can be indexed according to various schemas or DTD within a single repository.

Auteurs : Martin Sévigny (AJLSM) ; Frédéric Glorieux (AJLSM) - 2003-06-03

	Home
Copyright © 2000, 2001, 2002, 2003 Ministère de la culture et de la communication, AJLSM.