ETD Guide/Technical Issues/Database and IR

The next step after identification of the items of the digital library, the ETDs, is to address storage of the cataloguing attributes and the action of searching and retrieving. Remember that the quality of retrieval is dependent both on the programming of the search and retrieve functions but, as important as this, on the quality of the information used to catalog the items of the collection.

Databases are common and suitable tools to store, search and retrieve information. Besides this, they can also be very helpful in the process of capturing the attributes since they have the general function of managing information.

Before implementing the database, the database model must be created. This will happen only after the metadata model has been defined and related to other existing identification procedures such as traditional cataloguing on an automated library system.

If the traditional OPAC is to be maintained during the ETD program, it is desirable to avoid duplicated information. Thus, the attributes that are present in the OPAC should not be repeated and a link between the OPAC record and the ETD metadata is be created.

No matter where information is stored, the user should be able to perform the types of search that are standard in library systems: author, title, keywords, subjects, ISBN, etc.

As mentioned in the section Metadata models for ETDs, some information that is used to identify the items are language dependent (title, keywords, subjects, etc.). If the database holds only one language per record, search procedures are to be performed using the arguments in this language. If a multilingual database is modelled, it is recommended that search be language independent, i.e., that the argument be checked against all languages. In this last situation, after a record is found as the result of a search, all its language instances should be displayed to the user so that he/she can choose the language for retrieval.

Database extenders (text, image) may be considered to increase the number of points of access by performing the searches in the ETD's not only on the cataloguing attributes.

The relation with the legacy systems databases must be examined since information concerning the TDs may be stored on them. Some examples are the graduate program, the mentor, the examining committee, etc.

The next sections address specific aspects of this topic.

Next Section: Packaged solutions